Grout injection of masonry,

scientific approach and modeling

Doctoraatsthesis
promotor: Prof. dr. ir. D. Van Gemert
assesoren: Prof. dr. ir. J. Berlamont
Prof. dr. ir. K. Van Balen

Filip Van Rickstal

Table of contents
Dankwoord . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . I
Nederlandstalig abstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . III
Nederlandstalige samenvatting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . V
List of symbols . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . XXX
Table of contents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . XXXIII

Englisch abstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . -1Chapter 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . -3Chapter 2. Masonry and its need for restoration . . . . . . . . . . . . . . . . . . . . . -52.1. General description of masonry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . -5-

2.2.

2.3.
2.4.

2.1.1. Masonry components . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . -6Bricks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . -6Natural stone . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . -8Mortar . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . -8Pointing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . -11Causes of damage to masonry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . -112.2.1. Physical and physico-chemical mechanisms . . . . . . . . . . . . . . . . . . . . . . -112.2.2. Mechanical damage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . -132.2.3. Biological damage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . -15The loading of masonry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . -162.3.1. Probability of failure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . -162.3.2. Grout injection increases the reliability of masonry . . . . . . . . . . . . . . . -21Conclusions of Chapter 2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . -23-

Chapter 3. Injection as a consolidation technique for masonry . . . . . . . -253.1. Historical review . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . -253.2. Injection Technology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . -27-

3.3.

3.2.1. Modern injection installation, ideal situation . . . . . . . . . . . . . . . . . . . . .

3.2.2. Reality about injection equipment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Realization of an injection work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.3.1. General Requirements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Table of contents

-27-29-30-30-

XXXIII

3.4.

3.5.
3.6.
3.7.

3.3.2. Diagnosis of the masonry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Destructive techniques . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Non-destructive techniques . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.3.3. Preparation of the masonry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.3.4. Injection pressure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.3.5. Execution of the injection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.3.6. Control of quality . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Types of binding agents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.4.1. Polymers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.4.2. Cement grouts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.4.3. Lime based grouts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Grouting improves the load bearing capacity of the masonry . . . .
Masonry grouting, code of practice . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Conclusions of chapter 3 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

-30-31-32-36-37-39-40-42-42-44-45-46-49-52-

Chapter 4. Problems faced during injection - possible solutions . . . . . . -554.1. Wrong materials, chemical, physical and structural incompatibility

4.2.
4.3.

4.4.

..............................................................
Incomplete filling of the voids . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.2.1. Rheology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.2.2. Stability of grouts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Solutions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.3.1. Improving the injectability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.3.2. Improving the stability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.3.3. Injection holes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.3.4. Chemical and mechanical compatibility . . . . . . . . . . . . . . . . . . . . . . . . .
Subject of research . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

-55-57-58-59-61-61-63-63-65-66-

Chapter 5. Experimental program . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . -695.1. Aim of the tests . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . -695.2. Testing the grouts properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . -695.2.1.
5.2.2.
5.2.3.
5.2.4.
5.2.5.
5.2.6.
5.2.7.

5.3.

5.4.

XXXIV

Mixing procedure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Rheological properties of the grout . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Dynamic viscosity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Thixotropy, non linear behavior and time dependent properties . . . . . . .
Stability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Flow time measurements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
General observations about testing the grout properties . . . . . . . . . . . . .
Testing the masonrys properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5.3.1. Diagnosis of the masonry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5.3.2. Reproducible masonry samples: physical model . . . . . . . . . . . . . . . . . .
5.3.3. Permeability of samples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5.3.4. Permeability of masonry structure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Laboratory injection tests . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

-69-70-71-76-77-83-86-87-87-88-89-92-93-

Grout injection of masonry, scientific approach and modeling

5.4.1.
5.4.2.
5.4.3.
5.4.4.

5.5.

5.6.

Description of the tests . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . -93Flow charts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . -94Simplified mathematical model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . -102Mechanical properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . -106Important findings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . -1115.5.1. Blocking mechanisms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . -111Granularity of the cement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . -111Stability of the grout . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . -111Water absorption out of the grout . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . -111Pressure losses, thixotropy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . -1125.5.2. Observations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . -113Conclusions from the experimental program . . . . . . . . . . . . . . . . . . -114-

Chapter
6.1.
6.2.
6.3.

6. Rheology of grouts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Introduction to the rheology of dispersions . . . . . . . . . . . . . . . . . . . .
Non-Newtonian behavior of aqueous dispersions . . . . . . . . . . . . . .
Flow of a dispersion in a cylindrical tube . . . . . . . . . . . . . . . . . . . . . .

6.4.

6.3.1. General equation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

6.3.2. Newtonian fluid . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
6.3.3. Bingham fluid . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
6.3.4. Discussion on the Reiner-Buckingham formula . . . . . . . . . . . . . . . . . . .
6.3.5. Numerical simulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

-115-115-118-119-119-121-122-123-124-126-

Chapter 7. Flow of fluids through porous media . . . . . . . . . . . . . . . . . . . . -1277.1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . -1277.2. Structure and properties of porous materials . . . . . . . . . . . . . . . . . . . -127-

7.3.

7.4.

7.2.1. Definitions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . -1277.2.2. Methods for porosity measurement . . . . . . . . . . . . . . . . . . . . . . . . . . . . -1287.2.3. Permeability, Darcys law . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . -129Equations governing the flow of fluid through porous materials -1327.3.1. Differential form of Darcys Law . . . . . . . . . . . . . . . . . . . . . . . . . . . . . -1327.3.2. The differential equations of fluid flow through porous materials . . . . -1337.3.3. Boundary conditions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . -1347.3.4. Measurement of the permeability using compressible fluids . . . . . . . . . -1367.3.5. Radial flow between concentric cylinders . . . . . . . . . . . . . . . . . . . . . . -137Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . -140-

Chapter 8. Modeling grout flow in masonry . . . . . . . . . . . . . . . . . . . . . . . . -1418.1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . -1418.2. Discrete model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . -1418.2.1. Justification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . -1418.2.2. A network of discrete flow channels . . . . . . . . . . . . . . . . . . . . . . . . . . -1428.2.3. Boundary Conditions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . -144-

Table of contents

XXXV

8.3.

8.2.4. Dealing with water absorption . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

8.2.5. Special features of the program . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Structure of the program . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
8.3.1. Menu items . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
8.3.2. Calculation algorithm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

-144-146-146-147-153-

Chapter 9. Validating and using the model . . . . . . . . . . . . . . . . . . . . . . . . . -1599.1. Validation of the model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . -159-

9.2.
9.3.

9.4.
9.5.

9.1.1. Flow of Newtonian fluid through one dimensional cylindrical pipe . . .

Cylindrical pipe with constant diameter . . . . . . . . . . . . . . . . . . . . . . . . .
Cylindrical pipe with varying diameter . . . . . . . . . . . . . . . . . . . . . . . . .
Flow of Bingham fluid through one dimensional pipe . . . . . . . . . .
9.2.1. Conceptual validation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Using the model for parameter study . . . . . . . . . . . . . . . . . . . . . . . . . .
9.3.1. Grout parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Critical shear stress t c . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Viscosity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
9.3.2. Process parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Injection Pressure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Injection holes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
9.3.3. Radial network: sealing of a leakage . . . . . . . . . . . . . . . . . . . . . . . . . .
Using the model as an engineering tool . . . . . . . . . . . . . . . . . . . . . . . .
Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

-159-159-160-162-163-164-164-165-167-169-169-169-172-173-175-

Chapter 10. General conclusions and future research . . . . . . . . . . . . . . . -17710.1.

General conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . -17710.2.
Future Research . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . -17810.2.1. Information about the masonry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . -17810.2.2. Final goal, online controlled consolidation injection . . . . . . . . . . . . . -17910.2.3. Using the model for other purposes . . . . . . . . . . . . . . . . . . . . . . . . . . . -179-

XXXVI

Grout injection of masonry, scientific approach and modeling

Chapter 1.

Introduction

Imbued with a message from the past, the historic monuments of generations of people
remain to the present day as living witnesses of their age-old traditions. People are
becoming more and more conscious of the unity of human values and regard ancient
monuments as a common heritage. The common responsibility to safeguard them for
future generations is recognized. It is our duty to hand them on in the full richness of
their authenticity. [The Venice Charter, May 1964]
The concerns to maintain ancient valuable monuments are only to a little extend driven by the
desire to keep on using these buildings, but mainly by the desire to preserve cultural heritage as
a testimony from the past for the next generations. Very often this maintenance is not possible
without actions of restoration. The double interest implies that two aspects have to be fulfilled
during restoration: mechanical consolidation has to be combined with a method that preserves
as many original aspects as possible. These considerations are written down in the
internationally accepted Charter of Venice of may 1964 about the criteria for conservation and
restoration of monuments and sites. This chapter expresses the right attitude towards
preservation of valuable monuments. The treatment, i.e. consolidation, must guarantee that
probable external actions produce only repairable damages and no fatal artistic damage. The
authenticity of the historic monument, concerning both its structural and architectural values,
must be safeguarded. One has to realize that without structural safety the architectural value
will be lost for ever. Using the argument of maintaining authenticity of the building to refuse
any structural consolidation could result in the final collapse of the historic monument.
However, the safeguarding of the monument can be improved by using modern materials and
techniques, which preserve its authenticity.
In Belgium, just as in most other European countries, bricks and stones are the common building
materials. Though we often do not realize that our patrimony mainly consists of millions of
cubic meters of masonry. Without any doubt one could state that maintaining important part of
patrimony is maintaining brickwork or masonry. Masonry, just as most other physical entities,
is not for ever. Physical and mechanical actions lead to the decay of building materials. In
course of time, masonry buildings get physically deteriorated, either by inadequate design,
human intervention or natural causes. Frequently occurring natural causes for damage are
weathering, freeze thaw action, erosion of the mortar by rain water flow, overload by
earthquake actions. A frequent harmful human intervention is inappropriate loading which is a
direct action but also the lowering of the groundwater table causes soil settlement and damage
to the surrounding buildings. As these damage processes continue, repair or restoration and
consolidation are required. Among others, grout injection is a powerful consolidation
technique to overcome structural decay. The introduction of a binding agent in liquid form into
the masonry fills the holes, voids and cracks. After the hardening of this binding agent, the

Chapter 1. Introduction

masonry will regain its monolithical behavior and the overall mechanical resistance will
improve. As such, grouting restores the missing or deteriorated binding element of masonry.
Many techniques for the restoration of masonry structures are available. All these techniques
imply a certain infringement on the authenticity of the monument. In the ideal situation, the
intervention may not cause any damage to the structural and architectural authenticity of the
building and should be, as far as possible, reversible. In this complex and multi-disciplinary
domain of restoration, grout injection has found its place as a consolidation technique for
ancient masonry just because the technique is able to mechanically strengthen the historic
masonry monument without changing its outlook and integrity.
The grout is introduced into the internal, non visible part of the masonry and, because of this,
does not damage the aesthetical outlook of the building. However, grouting is not reversible,
but when materials are used that are compatible with the original materials, it certainly is a
justified technique that fulfills the requirements of safeguarding the monuments authenticity.
The Building Materials Division of the Civil Engineering Department of the Katholieke
Universiteit of Leuven has been involved in several practical injection jobs as consultant and
decided to investigate restoration of masonry more thoroughly.
The present thesis discusses in general grouting as a consolidation technique for masonry. In
the first chapter the structural aspects of masonry and the physical deteriorating mechanisms are
listed and a probabilistic method is presented to judge the need for consolidation. It is
theoretically shown how the uniform filling of the voids and holes overcomes the splitting
forces that are present around these holes and how the reliability of the masonry structure
increases after injection. Different binding agents are listed and the advantages and
disadvantages are indicated. A technological part gives an overview of injection methods and
indicates how an ideal injection installation should look like. This book also presents the
experimental program. The results of the experiments provided an enhanced physical
understanding of the injection process. Most important are the findings with regard to the
injectability of grouts, rheological properties of grouts, characterization of masonry for
injection purposes. For all three of these aspects new or existing tests have been developed or
adapted to the peculiarities of grouting. But the originality of this work lies in the modelling of
the flow of the grout through the masonry. The masonry is simulated by a network of discrete
channels representing the big channels through which, starting from the injection hole, the grout
penetrates the masonry. The penetration of the grout into the areas with finer void structure is
built in by means of capacitive elements. The water absorption out of the grout by the dry
masonry is also incorporated. This way, the model is a useful design tool to determine the
injection parameters: grout composition; viscosity, shear stress and hardening evolution,
injection holes pattern and injection pressure.

Chapter 2.

Masonry and its need for restoration

Grout injection of masonry, scientific approach and modeling

2.1.

General description of masonry

Here we intend to give an overview of the most important developments about masonry in
course of time, especially since some developments have their impact on nowadays restoration
projects.
Masonry is a handmade construction element, made by assembling natural and artificial
materials, eventually using a binding agent (Figure 2.1). The natural materials are the stones,
as they were found or as they were prepared into the requested dimensions. The more reliable
the format and the size of the stones, the thinner the layer of binding mortar can be. Masonry
exists for many centuries. It is a gradual evolution of placing big stones one on top of each
other to the agglomeration of fine prefabricated stones and mortar that is used today. With
regard to restoration, there exists a big difference between the massive monuments of the
Egyptians and the Medieval churches of the European regions.
The ancient massive monuments are very impressive, they mainly consist of big walls and
pillars. However, they do not or hardly create a free space: the distance between the supports
is mostly very short. The use of arches, vaults and domes in the more recent monuments allows
to deviate the gravity forces towards the foundations [J.H. Acland, 1972][J. Fitchen, 1985].
An overwhelming feeling of space is created. The latter buildings are more slender, more
elegant, but also more vulnerable to mechanical damage for which consolidation injection can
bring a solution.
Because of that, this chapter describes the different component of masonry, limited to the
European masonry types, being a agglomeration of small stones and bricks. Later in this
chapter possible damage phenomena are discussed. Only part of these phenomena are
repairable using consolidation injections.

Chapter 2. Masonry and its need for restoration

Figure 2.1: The Paterskerk in Tienen, a valuable Medieval monument that was
succesfully safegarded using grout injection

2.1.1. Masonry components

Bricks
Bricks are artificial stones made by baking clay at high temperature. Bricks were originally an
ersatz for the natural stones for those regions where no excavation of stone was possible Figure
10.1. The baking process has been improved a lot since the first field ovens. Nowadays, the
technology has developed and bricks having a constant composition and quality can be
produced. Though it is important to have in mind the ancient way of producing bricks in order
to be aware of some possible problems. It is important to be aware of the historical production
process. To give an example: the bricks that were positioned at the outer side were of a better
quality than the bricks inside. During some restorations this fact was disregarded resulting in
mixing all the bricks. When a soft brick is used to rebuild the facade it will soon be eroded by
rainwater and freeze thaw actions (Figure 2.2).

Grout injection of masonry, scientific approach and modeling

Figure 2.2: During the restoration of the Beguinage of Leuven, soft bricks were placed in
the facade leading to accelerated damage phenomena, such phenomenon can not be
repaired using injections

In our countries, the brick can be seen as a religious stone. Brickwork has been the monopoly
of the monks for many centuries [G. Peirs, 1979]. They have developed the technology of
masonry and transported the knowledge all over Europe. Bricks were a necessity to build
houses of God that would last longer in those regions where no natural stone was available.
The brick has numerous appearances. The dimensions of the brick, the color and the chemical
composition reveal the origin of the bricks. The size, the composition and the color can help to
date the building. These properties and the big aesthetical and technical difference between the
old original bricks and the new bricks make it sometimes hard to replace weathered bricks by
new ones. In any case the authenticity of the building will be damaged. The technical
difference between the modern so called hand made bricks and the old bricks is caused by the
following facts:
C
the clay, used to make the old bricks, contained more organic material than nowadays.
C
there were hardly any possibilities to enhance the mineral composition and the
granularity of the clay. Clay was suitable for the production of bricks if a fist of clay
did not fall apart after drying.
C
a higher water content to provide the necessary plasticity to the clay that had to be
mixed using manpower, implies a higher porosity and a higher shrinkage.
C
the limited pressure applied to the clay when introduced in the mold

Chapter 2. Masonry and its need for restoration

C
C

a slow drying process in open air compared to the drying process in dry ovens
nowadays
uncomplete sintering because of the lower baking temperature or because of less
material, present in the clay, that has a lower sintering temperature.

The higher porosity dos not necessarily mean a negative property for the brick. The high
porosity can have a good impact on the moisture household, the frost resistance, the density and
acoustic and thermal insulation.
The actual production process provides more homogeneous bricks, better baked and containing
less impurities.
Natural stone
Bricks have, in course of time, reduced the use of natural stone for masonry. Though,
brickwork is often combined with masonry of natural stone, especially in valuable monument
since natural stones gave the building an image of wealth and were known to be very durable.
The replacement of natural stones can be problematic if the quarries are no longer exploited.
Besides, many natural stones undergo an accelerated weathering because of air pollution. The
original stones and the replacing stones have a different structure and because of small
differences in the weathering resistance, the homogeneous outlook of the facade is disturbed.
The fabric of natural stones in the old artisanal way is expensive. The knowledge, the tools and
the workers are not readily available.
Most buildings were erected using the stones available in the neighborhood. Four our regions
this means relatively soft stones such as marl, limestone, sandstone or iron sand stone. Those
natural stones withstand poorly the effect of acid rain.
When the replacement of natural stones is necessary, one should try to find a similar stone, with
the same composition, the same porosity providing similar properties with regard to water
transport and the same frost resistance.
Mortar
Mortar consisting primarily of lime and sand, has been used as an integral part of masonry
structures for thousands of years. Until about the middle of the 19th century, lime was delivered
to construction sites, where it had to be slaked, or combined with water. Mixing with water
caused it to boil and resulted in a wet lime putty that was left to mature in a pit or wooden box
for several weeks, up to a year. Traditional mortar was made from lime putty, or slaked lime,
combined with local sand, generally in a ratio of 1 part lime putty to 3 parts sand by volume.
Often other ingredients, such as crushed sea shells (another source of lime), brick dust, clay,
natural cements, pigments, and even animal hair were also added to mortar, but the basic
formulation for lime putty and sand mortar remained unchanged for centuries until the advent of
portland cement or its forerunner, Roman cement, a natural, hydraulic cement. In the 1930s
more new mortar products, intended to accelerate and simplify masons' work, were introduced.

Grout injection of masonry, scientific approach and modeling

These included masonry cement, a premixed, bagged mortar which is a combination of portland
cement and ground limestone, and hydrated lime, machine-slaked lime that eliminated the
necessity of slaking quicklime into putty at the site.
CaCO3 + heat ( 900EC)

CaO + CO2

(Eq 2.1)

CaO + H2O

Ca(OH) 2

(Eq 2.2)

Ca(OH) 2 + CO2

CaCO3 + H2O

(Eq 2.3)

The chemical reactions (Eq 2.1) and (Eq 2.2) take place during the production of hydrated
lime. The calcium carbonate is provided by lime stone or by shells. This calcium carbonate is
dissociated at high temperature. The Calcium oxide is then extinguished using water in the
exact dosage (providing white powder) or in over dosage (providing the white putty as
mentioned above). The lime provided by burning limestone has no hydraulic properties. The
lime made by burning the coquilles has, due to some clay impurities, some hydraulic features.
The hydraulic properties gave these mortars a good early strength development. Reaction (Eq
2.3) is using CO2 and hence requires the presence of air. For a thick wall, this might be a
problem. The transport of air towards the fresh mortar was ensured by leaving some of the
pointing open. This proves the understanding of the hardening mechanisms of lime mortar.
The preference for a mortar showing slightly hydraulic properties, was already mentioned. To
donate some hydraulic properties to lime mortars, mineral admixtures were used. Very well
known in our countries is Trass. Trass has, just as the volcanic earth from the Vesuvius used by
the Romans or the Santorrini earth in Greece, hydraulic properties. Other hydraulic admixtures
existed. Actually portland cement was originally used as hydraulic admixture. Only for the
application of making concrete, portland cement was used as a pure binding agent. Portland
cement was patented in Great Britain in 1824. It was named after the stone from Portland in
Dorset which it resembled when hard. This is a fast-curing, hydraulic cement which hardens
under water. Until the turn of the century, portland cement was considered primarily an
additive, or "minor ingredient" to help accelerate mortar set. By the 1930s, however, most
masons used a mix of equal parts portland cement and lime putty. Thus, the mortar found in
masonry structures built between 1873 and 1930 can range from pure lime and sand mixes to a
wide variety of lime, portland cement, and sand combinations. The mixing proportions of lime
and sand have changed a lot during the 19th century. In the beginning of the century no or very
few sand was added to the lime. In 1833 to 1850 some sand is added, but still the composition
possesses more lime than sand. A manual of 1874 mentions a ratio of 1 part of lime on one part
of sand. By the end of the century a lime:sand ratio of 1:3 is usually used.

Chapter 2. Masonry and its need for restoration

The use of the relatively soft, air hardening lime mortar gave ancient masonry a good capacity
to recover from settlements. The lime mortar has a slow strength development. For historical
buildings that were constructed slowly, this was not a big disadvantage. On the contrary,
deformations during the construction distributed and moderated the stresses. Furthermore, the
mortar remained less strong than the stones. Occurring cracks were located in the mortar joints,
where they could easily be hidden by repointing. Generally, a lime mortar is more elastic and
tougher. This provides an additional safety with regard to differential settlements. Lime mortar
contains no or hardly any sulfate or alkaline. This reduces the risk for salt efflorescence.
The above implies that there is a double reason to use mortars that are compatible with the
original mortars. First of all there are technical reasons: using a modern mortar would result
in introducing a component that is harder than the old mortar and in most cases also harder than
the stones that were used. Settlements become hard to follow. A hard nucleon, created by the
hardening of the cement grout causes tensile forces to occur in the masonry. The new hard
mortar splits the masonry just above the hard zone since the zone on the left and right hand side
of the hard nucleon are softer and more deformable. One gets a kind of bending/splitting action
caused by the hard part of the laying mortar and the mass of the masonry above (Figure 2.3).
Furthermore, a cement mortar has a different porosity causing a different action with regard to
water transport.

Figure 2.3: Bending/ splitting caused by hard cement grout among soft lime mortar
Second valuable reason is the aim not to introduce materials that were not used in the original
building as it is mentioned in the Venice Charter. General rule of thumb is to use mortars that
imitate the original mortar and that are as hard as the original one or even somewhat softer.
Grout composition will be further discussed in Chapter 3 and Chapter 4.

10

Grout injection of masonry, scientific approach and modeling

Pointing
Until the 18th century the pointing was done immediately after placing the stones, with the same
mortar, resulting in a solid unit [G. Peirs, 1979]. Later, in the 18th century when the masonry art
aimed a narrow layer between the bricks, the pointing was done afterwards. Depending on the
fashion of each period and the architects vision, a flat pointing or a drawing back pointing has
been applied.
Repointing is often an important action in the restoration of a facade. Most of the time only part
of the facade is damaged. However, a global repointing is mostly done, using the argument of
aesthetical harmony. Thus, the remaining pointing are cut out. This should be done cautiously
since the stones and the undamaged pointing form a unit that is often damaged by this action.
The new pointing becomes wider than the original one, completely changing the outlook of the
historical building. Likewise, other undesirable effects take place. The new mortar used for
repointing is, when disregarding the original composition, almost impermeable to water in
comparison with the original situation. Therefore, the water concentrates just behind the
repointed layer. Salt crystallization or frost can then easily push the new layer outwards
causing even more damage since the adhesion of the repointed mortar to the existing bricks is
very good. A general advise is only valuable when applicable, but one should limit the
repointing to the damaged parts and use a mortar composition that corresponds as well as
possible with the mortar used for the original pointing. Very often, a consolidation injection is
combined with a partial or general repointing.

2.2.

Causes of damage to masonry

2.2.1. Physical and physico-chemical mechanisms

Physical mechanisms mostly need a long period of time to cause visible damage. Many
physical mechanisms are related to the presence of water inside the structure. Porosity,
capillarity and permeability regulate the transport of moisture. Masonry is a highly porous
building material. The rain water is absorbed by capillary action. The presence of water
inside the stones and the mortar means a real danger for weathering mechanisms. Most
important weathering mechanism is the frost damage. Frost damage can be recognized by the
typical fractures along the frost front. The frost resistance of building materials can be checked
either by a direct test applying frost - thaw cycles [NBN B05-203] or by an indirect test. In the
indirect test [NBN B05-201] the frost resistance is judged by analyzing the capillary water
absorption. From this analysis, the GC-factor can be calculated. The GC-factor gives a good
indication about the frost resistance of the bricks. A low GC-factor (e.g. lower than -2,5)
indicates a very good resistance to frost damage. When the GC-factor is higher, additional
direct testing of the frost resistance is advisable.

Chapter 2. Masonry and its need for restoration

11

Figure 2.4: An example of typical frost damage to masonry, which can not be repaired by
injection

A hydrofobic treatment can reduce the absorption of rain water. Nevertheless, very often the
water inside the masonry is not only due to rain water absorption. Other sources exist. The
capillary suction and rise of groundwater can not be avoided by a surface treatment with hydrofobic agents. When groundwater is absorbed by the masonry, one should not apply a
hydrofobic treatment. The outer layer of the masonry becomes impervious to water. Only
water vapor is able to be transported. This means a much slower process of loss of water then
when the water is able to proceed to the surface itself. This way two phenomena take place:
salts, soluble in water, are left in the transition zone were further water transport occurs by
vapor. The remaining hygroscopic salts cause big internal pressure when crystallizing, resulting
in the spalling off of the treated outer layer of masonry. Since the transport mechanism of vapor
is slower then the transport of water, a concentration of water will be present in the transition
zone. When temperature drops, this water will freeze and the expansion at freezing will push
away the outer layer of masonry.
Furthermore, moisture movements give rise to the dissolution and the corrosion of the binder.
When the mortar is leached out, the internal cohesion decreases. Grout injection is very
suitable to repair this kind of damage. The more uniform the grout fills the voids caused by the
erosion of the binder, the better the final consolidation. Moisture movement also causes the

12

Grout injection of masonry, scientific approach and modeling

transport of soluble salts [D. Van Gemert, 1988(1)]. These salts can crystallize in a different
place and might push off the outer masonry layer.
Lime stone suffers a lot from acid rain. Sulphur dioxide enters the lime stone and dissolves the
CaCO3 and forms calcium sulfate from calcium sulfate and oxygen. The calcium sulfate reacts
with water and forms gypsum. This causes no problems, except for some efflorescence that can
easily be washed with water. When a cement grout is injected, C3A is imported in the masonry.
In combination with gypsum and water, C3A enables the formation of ettringite (C3A . 3CaSO4 .
31 H2O), a very expansive mineral due to its high water binding capacity: 31 H2O. Calcium dinitrate, present in many fertilizers is supplied by absorbed groundwater and crystallizes
expansively.
Thermic cycles cause cyclic stresses inside the material. They cause cracks, situated in the
outer layer of masonry because of the large tensions that occur.
The mortar may be eroded by rain water. A good maintenance of the building is the best
prevention for damage. Lack of maintenance speeds up most of the aforementioned erosion
phenomena. This necessitates a repointing, but if not discovered in time, the structure can be
damaged to such an extend that a consolidation becomes necessary.

2.2.2. Mechanical damage

Mistakes in the original design or concept of a building can cause mechanical damage soon
after or even during the construction of the building. A poor dimensioning of structural parts or
an unexpected settlement of the soil are frequently occurring reasons for early damage.
Modifications by man of the original structure often cause additional damage to the structure.
The construction of a higher tower on the church means a load for which the church and its
foundations were not dimensioned. A different use of the building, for instance a library
function or a dancing room often means a load much higher than the design load. Restoration
actions can disturb the original distribution of forces: the placement of anchors, the placement
of stiff units or enlarging the openings for the windows are just a few illustrations of possible
mistreatment by man. Lowering the groundwater table for construction works in the
neighborhood changes the load bearing capacity and especially the deformational behavior of
the soil. This can cause large settlements and subsequent fissuration to the historical buildings.
These fissures mostly pass through the entire wall. They can be filled by grout injection.
Vibrations induced by man, machinery, traffic or construction works may cause deformations
and smaller or larger forms of distress in buildings, their structural members and nonstructural
elements [H. Bachmann, 1987]. Forms of distress are:
C
cracking of walls and slabs
C
aggravation of existing cracking in structural members and nonstructural elements
C
falling down of equipment or cladding thereby endangering occupants

Chapter 2. Masonry and its need for restoration

13

Continuous vibration, however, can also lead to problems of fatigue and overstress in principal
load-bearing members. The degree of damage depends upon the quality of the building
material, the type of construction, the properties of the building foundation, the main dimensions
of the principal load-bearing members, the age of the building, the duration and the
characterization of the vibrations. The vibration velocities can be measured using
accelerometers. Part 3 of the German standard DIN 4150 treats effects on buildings and
structural members due to an internal or external source of vibration (Figure 2.5).

Figure 2.5: Maximum allowed peak values for vibrations, depending

on the category of building [DIN 4150, part 3]
The Association of Swiss Highway Engineers distinguishes in their Standard SN 640312 four
different categories of buildings mainly according to the type of construction. The acceptance
criteria are again peak vibration velocities. The most severe acceptance criteria are valid for
buildings that are very vulnerable for vibration or that are worth protecting. Historical
monuments belong to this group. The allowed peak vibration velocity depends upon the kind of
source and the frequency of the vibrations.

14

Grout injection of masonry, scientific approach and modeling

Table 2.1: Structural categories according to SN 640312

Structural
Category

Definition

reinforced-concrete and steel structures( without plaster) such as industrial

buildings, bridges, masts retaining walls, unburied pipelines
underground structures such as caverns, tunnels galleries lined and unlined.

II

buildings with concrete floors and basement walls, above-grade walls of

concrete, brick or ashlar masonry; ashlar retaining walls, buried pipelines
uderground structures such as caverns, tunnels galleries, with masonry lining

III

buildings with concrete basement floors and walls, above grade masonry
walls, trimber joist floors

IV

buildings which are particularly vulnerable or worth protecting

Table 2.2: Acceptance criteria of SN 640312 for historic buildings worth protecting
Structural
category
IV: vulnerable
building/ worth
protecting

Source M

Source S

f [Hz]

vmax [mm/s]

f [Hz]

vmax [mm/s]

10 to 30

10 to 60

30 to 60

3 to 5

60 to 90

8 to 12

Source M: Machinery, traffic or construction works

Source S: Blasting operations

If the occurring vibrations cause velocities that are higher than the above mentioned maximum
peak values, they might cause structural damage.
Mechanical damage is not only due to human actions. Heavy wind and rain or storms can cause
severe damage, as well as seismic action [M. Tomazevic, 1982]. The corrosion of steel, for
instance from anchors, is an expansive process. The corrosion spot is only an aesthetical
consequence, but the mechanical damage is often more important.
Mechanical damage is very suitable to be repaired by grout injection.

2.2.3. Biological damage

Just to be complete, biological damage mechanisms are mentioned. The formation of algae is a
more aesthetical problem. Still they produce, in combination with organic material such as
rotten leafs or pigeons excrements, acids that deteriorate limestone and mortar. The fine roots
of plants and trees can enter the openings inside the masonry, spalling it apart. Micro

Chapter 2. Masonry and its need for restoration

15

organisms bring along bacteria that produce nitrates and sulfates as residue of their metabolism.
These are the causes of biogenic erosion phenomena.

2.3. The loading of masonry

2.3.1. Probability of failure
The above description of masonry reveals that masonry is a composite, heterogenous building
material. These facts make it impossible to accurately predict its behavior, by the simple
knowledge of the mechanical properties of its constituents. This is true for plain masonry
structures (Figure 2.6). For three leaf rubble core masonry the situation is even more complex
(Figure 2.7).

Figure 2.6: Plain masonry structure: a collection of heterogenities

16

Grout injection of masonry, scientific approach and modeling

Figure 2.7: Additional heterogeneities in case of double leaf masonry

with rubble core

The general design rules avoid tensile stresses inside the material. The use of arches, vaults
and domes allows to deviate the gravity the foundations [J.H. Acland, 1972][J. Fitchen, 1985].
The combination of a good resistance to compression and the ability to absorb large
deformations provide a great capacity to absorb deformation energy. The dead weight of
masonry helps to protect the masonry from tensile stresses that could be caused by bending,
eccentric loading or horizontal forces such as wind load (Figure 2.8).

Figure 2.8: The dead weight of masonry may prevent tensile stresses due to bending
Chapter 2. Masonry and its need for restoration

17

In the ideal situation, as shown in Figure 2.9, theory of elasticity learns that around circular
holes, not only compressive stresses up to three times the average stress arise, but also tensile
stresses up to the average compressive stress. In the non ideal case of non circular holes the
situation is worse and the stress concentration levels are even higher.

Figure 2.9: A hole in a massive masonry part induces both higher compressive and
tensile stresses. [D. Van Gemert, 1984]

The difference in module of elasticity and the difference in the Poisson ratio also causes tensile
stresses in the bricks as can be seen from Figure 2.10

Figure 2.10: The compressive loading causes tensile stresses where the brick is in
contact with the mortar layer due to a different value of the Poisson ratio.

18

Grout injection of masonry, scientific approach and modeling

To characterize the behavior of masonry under vertical and horizontal loads, the two
fundamental mechanical properties are its compressive strength and its shear strength. The
compressive strength is a function of the quality of the binder and the strength of the stones or
bricks. Empirical formulas found in literature enable a first estimation of this strength as a
function of the strength of binder and stone or brick and according to the quality of the masonry.
Appropriate design methods are based on a reliability analysis of the building [Eurocode 1] [L.
Schueremans, 1996] [L. Schueremans, 1999]. The idea behind a reliability analysis or behind
the evaluation of the reliability is relatively simple to explain using the basic reliability
problem. Both the load on a wall (S) and the strength of that wall (R) are stochastic variables.
Since their exact value is unknown, they are represented by the probability functions, fS(s) and
fR (r) respectively. The loading S and the resistance of the masonry R are both a function of
time. The load has the tendency to increase whereas the load bearing capacity of the structure
has a decreasing trend due to all kind of degradation processes. Many mechanisms of
deterioration are discussed above. The edge of safety will be passed at a certain moment t,
where
R(t) & S(t) < 0

(Eq 2.4)

The probability that this happens is the probability of failure pf. For mathematical reasons the
time dependent probability functions are transposed to time invariant functions. The reliability
calculation is then made for a determined period of time, a reference period. The probability
of failure can then be expressed by
pf ' P[ R & S < 0] for the reference period tL

(Eq 2.5)

The above expression can be generalized to

p f ' P[ g(R,S) < 0] where g(R,S) is the limite state

(Eq 2.6)

The probability of failure is calculated using a FORM-algorithm (a First Order Reliability

Method). This FORM-algorithm calculates the reliability index that is related to the
probability of failure pf by the standard cumulative probability function. The relation between
both parameters is given in Table 2.3. The smaller the probability of failure, the higher the
reliability index. A reliability index of 3,7 corresponds to a probability of failure of 1/10.000
within the design working life, which is generally seen as a minimum safety level.

Chapter 2. Masonry and its need for restoration

19

Table 2.3: The relation between pf and [Eurocode 1]

Pf

10-1

10-2

10-3

10-4

10-5

10-6

10-7

1,3

2,3

3,1

3,7

4,2

4,7

5,2

Looking at the method of reliability analysis, one can discover the parameter that influences the
probability of failure. First possibility is to increase the separation between the strength
function R and the loading function L. This can be done by either adapting the loading, for
instance be reducing the loading of the floors, or by adapting the strength. The latter is
presented in Figure 2.11. The original situation is represented by R1. The overlap area
between the stresses caused by the load on the structure (L) represents the probability of
failure. Increasing the average strength as is the case for situation 2 (Figure 2.11, f R2 > f R1)
makes the strength probability distribution function to shift to the right. Therefore, the
reliability index increases (Figure 2.11, R1 R2)

0.6

R2

R1

4
3
1

0.1

0
7

11
2

Mean value [N/mm ]

R2

0.3
0.2

target value : beta = 3.7

fS
fR1
fR2

0.4
reliability index

0.5

Reliability index

13

10

15
2

stress (S) - strength (R) [N/mm ]

Figure 2.11: Influence of increasing the strength of the wall on the reliability index.

A second possibility is to reduce the variance on the strength. The average strength remains the
same, but the extreme values differ less from the average value. Therefore the uncertainty on
the strength decreases. The strength distribution function f R3 is much narrower than the
distribution f R1. As can be seen, the overlap area (Figure 2.12) decreases significantly
although the average strength does not increase. The reliability index for situation 3 is
significantly higher than for situation 1.

20

Grout injection of masonry, scientific approach and modeling

0.6

6
4

reliability index

0.5

target value : beta = 3.7

0.4

R3

fS
fR1
fR3

0.3

Reliability index

R3

0.2
2

0.1

R1

0
15

25
35
45
Coefficient of variation [%]

55

10

15
2

stress (S) - strength (R) [N/mm ]

Figure 2.12: Influence of the variance on the strength of the wall on the reliability index

2.3.2. Grout injection increases the reliability of masonry

The increase of the average strength of the masonry or the reduction of the variance of the
strength or both are the intention of a consolidation injection as stated by L. Schueremans [L.
Schueremans, 1996]. To have an idea of the influence of a consolidation injection on the
reliability, one needs an expression that indicates how the strength function R changes in
function of the injection: the used grout and the injected volume. It is supposed that the average
strength depends upon the average strength of the original masonry and the average strength of
the injected grout. Both materials contribute to the global strength in relation with their relative
volume. This brings us to the expression for the average strength of injected masonry [P.
Tassios,1995]:
( f wc,i ) '

where

fwc,i

Fwc,o

Fic

Vo
Vi
Vtot

:
:
:

Vtot& V0
Vtot

. ( f wc, o) % (

Vi
Vtot

) . ( fi c)

(Eq 2.7)

the strength of the masonry after injection, which is on its turn a

stochastic variable fwc,i .LN ( (fwc,i),(s (fwc,i))2)
the original strength of masonry without holes, also a stochastic
variable: fwc,i .LN ( (fwc,o),(s (fwc,o))2)
The compressive strength of the injected grout being a stochastic
variable: fic . LN ( (fic),(s (fic))2)
volume of holes in the masonry
injected volume
total external volume of the masonry structure

It is obvious from the above expression that

C
the higher the injected volume, the more the averages strength of the injected material
will increase for constant average compressive strength of the original masonry and of
the injection grout. This is a very important finding and it proves that a uniform filling

Chapter 2. Masonry and its need for restoration

21

of all voids is very important. This fact imposes high rheological demands on the
composition of the grout. They will be discussed later.
the higher the mechanical strength of the grout, the higher the resulting average strength
of the injected masonry on the condition that the injected volume remains the same. It
has to be mentioned that the injection of a very strong grout increases the heterogeneity
of the masonry.
Vtot & V o
Vi
Vtot & Vo % Vi
%
'
< 1 or only the voids can be injected.
Vtot
Vtot
Vtot

It can be concluded from the above discussion that injecting a grout provides a higher average
strength of the masonry and hence reduces the probability of failure.
In the mean time, grout injection also reduces the uncertainty of the strength of the masonry.
Filling the voids reduces the variance on the masonry strength by making the material more
homogeneous. Compression tests on cylinders, cored from ancient masonry, show a variance
of about 40 %. The decrease of this variance on the compressive strength is due to:
C
a uniform filling of the masonry by the grout.
C
a better internal cohesion of the masonry
The final variance depends on the degree of filling, but also on the variance of the strength of
the grout.
It can be concluded that grout injection transforms the masonry into a more homogeneous
material and it provides a higher average strength and that the uncertainty about the strength
becomes smaller, even if the average strength would remain the same.
A possible expression to calculate the variance on the strength is given by
Vi % VL

cov ( fwc,i ) ' Vtot & d12

where

cov(...)
d12
VL

Vtot & V L

cov (fxc,o) % d12

Vi
Vtot & VL

cov ( fic)

:
:

the variance
kronecker delta,

d12 = 1 if Vi 0
volume of remaining voids after injection

(Eq 2.8)

d12 = 0 if Vi = 0

Equation (Eq 2.8) is the result of the following reflections:

C

22

the better the injection, the smaller the amount of voids that are remaining VL.. This
leads to a lower variance on the strength

Grout injection of masonry, scientific approach and modeling

the better the injection, the higher the injected volume (Vi). This increases the relative
importance of the variance on the strength of the injected material in equation (Eq 2.8).
If the variance of the injected material is smaller then the variance on the original
masonry, this effect is enhanced.

2.4. Conclusions of Chapter 2

Above, a general description of masonry as a heterogeneous building material is given.
Possible causes of damage are listed and discussed. From literature study it is obvious that
part of these damaging phenomena give rise to structural problems. In those cases, grout
injection can be a solution to retrofit the building [P. Shing, 1994] [G. Penelis, 1989]. As is
proved by the presentation of the reliability method to judge the safety condition of an ancient
masonry, the increase of the average strength and the reduction of the strength variance give rise
to an improved safety and a lower risk of collapse. The reduction of the variance of the
strength mainly depends upon the degree of uniform filling of the voids and cracks. An
injection of a piece of masonry that only partially filled the masonry, might increase the average
strength but will certainly also increase the variance on the strength. So, a good and uniform
penetration of the grout inside the masonry is crucial to achieve a successful consolidation.
The study, presented in this book, aimed to investigate which conditions give rise to the
required penetration and the uniform filling of internal voids in masonry.
The experimental program enabled to understand the physical phenomena that happen during an
injection. It has shown which conditions cause a poor penetration and which conditions
provide uniform filling. The test injection provided information about the conditions leading to
a good penetration of the grout. The model facilitates to find out which parameters of the
grouting process can successfully be adapted to guarantee the most uniform filling of the voids.
This way the experimental results, as well as the use of the model, can help to realize better
consolidations of monuments.

Chapter 2. Masonry and its need for restoration

23

24

Grout injection of masonry, scientific approach and modeling

Chapter 3.
masonry

Injection as a consolidation technique for

3.1. Historical review

Humanity has always tried to maintain its patrimony. Not in the first place to continue to make
use of it, but mainly to preserve cultural heritage as a testimony from the past for the next
generations. The double interest implies that two aspects have to be fulfilled during restoration:
mechanical consolidation has to be coupled to a method that preserves as many original aspects
as possible. Therefore, engineers need help from art historians to know about the historical
materials, the historical way of building. The materials used for restoration should be
compatible to the historical materials, physically as well as chemically. The intervention may
not cause any damage to the structural and architectural authenticity of the building and should
be, as far as possible, reversible. In this complex and multi disciplinary domain of restoration,
grout injection has found its place as a consolidation technique for ancient masonry to
overcome structural decay.
Grouting consists of introducing a binding agent in liquid form into the masonry to fill the holes,
voids and cracks. The binding agent will cure and increase the internal cohesion. After
hardening, the masonry will regain its monolithic aspect and show an increased load bearing
capacity. The grout is introduced into the internal, non visible part of the masonry and does not
damage the aesthetical outlook of the building.
However, grouting is not reversible but when materials are used that comply with the original
materials, it is a justified technique, complying with the Charter of Venice.

Figure 3.2:
Schematical
representation
gravitational
usedinjection
nowadays in Mexico
3.1: The
grout filling
the rubble of
core
can be seeninjection,
as a gravity

Chapter 3 - Injection as a consolidation technique for masonry

25

One could enumerate different injection techniques depending on the driving force that is
forcing the grout to penetrate inside the masonry. The filling of the rubble core masonry could
be seen as the first injection what so ever. This technique consists in the pouring of a fluid
mortar inside the space between the two leafs of the structure Figure 3.1. This fluid mortar
partly fills the openings of the rubble and provides some adhesion of the middle part to both
leafs. Since gravity was used to make the mortar flow downwards it can be seen as an
injection by gravity. Also for the first consolidation injection of a masonry building, performed
in 1802 under the supervision of engineer C. Brigny, hydrostatical pressure was used to force
the grout to penetrate inside the walls [C. Besson, 1989]. The hydrostatical pressure is created
by a grout column. The constant height of the column keeps the injection pressure constant,
which is a main advantage of gravity grouting. The pressure will be limited because it is
practically unfeasible to create a very high grout column. This procedure guarantees a constant
limited injection pressure, which is a big advantage to prevent further damage. The density of a
cement grout is about 1600 kg/m3. Hence, the height of the grout column to apply a pressure of
one bar equals 6.25 m.
In 1871, the royal building inspector Daser injected cement grout to stabilize tunnels. He was
the first to use a mechanical pump to inject the grout. Already in 1888 the first industrial
pressure grouting machines are produced in England and Germany enabling higher injection
pressure. Nowadays, modern pumps can build up very high pressures. So the grout can be
pumped to elevated injection holes up to 100 meters or higher. Control systems prevent the
pressure to exceed the desired injection pressure. In case of important consolidation works,
pumped grouting is actually the only option due to its superior efficiency and versatility and due
to its ability of injecting large volumes of grout.
For smaller jobs, for instance the consolidation of a gate, the pedestal of a statue or the
foundations of a small building, manually pumped grout can be considered.

3.2. Injection Technology

3.2.1. Modern injection installation, ideal situation
A modern injection installation, suitable for important injection works, consists of a complete
set of devices to achieve an optimal result. Hereafter, the necessary components of a
professional injection equipment for mineral grouts, as represented on Figure 3.3, are
described with regard to their ability of making an injection work successful.

26

Grout injection of masonry, scientific approach and modeling

Figure 3.3: Modern injection installation with the necessary components

C
Mixing installation [Figure 3.3, A]
This device mixes the materials. The binding agent, the water, admixtures and additives are put
together and thoroughly mixed. A thorough deflocculation of the particles of the binder is at
least as important as producing an homogeneous grout. This requires a mixing procedure that
provides high shearing. High turbulence mixing is one possibility, ultrasound mixing is even
better [E.E. Toumbakari, 1999(1)]. It has to be mentioned that ultrasound mixing installations
are not commercially available yet. According to Miltiadou [A. Miltiadou, 1991] ultrasound
mixing has important and observable advantages. Therefore, it is advised to obtain a suitable
grout in two steps: a first high turbulence mixing to produce an homogeneous liquid, followed

Chapter 3 - Injection as a consolidation technique for masonry

27

by an ultrasound mixing to deflocculate. The latter increases the stability and the injectability
of the grout significantly. After mixing the grout enters the collector.
C
The collector [Figure 3.3, B]
This recipient contains the grout, previously mixed in the mixing installation. To keep the
cement particles in suspension, not more not less, the collector continuously stirs the grout. It
has to be mentioned: the action of the collector is insufficient to mix the grout properly.
Because of dosage, the mixing takes place in discrete batches, whereas the pumping is a
continuous process. This requires a buffer. The collector acts as a buffer to the pumping
installation.
C
Pumping installation [Figure 3.3, C]
The pumping installation is fed by the collector. The pumps can either be of volumetric of
continuous type. The volumetric pumps provide a pulsating pressure, just as the heart beat.
One might think that this ram effect helps the injection. This seems not to be the case. Binda
[L. Binda, 1993] indicates in her research that a constant injection pressure provides the best
results. The pumping installation should therefore be able to supply a uninterrupted discharge
of grout at constant pressure. A volumetric pump, at the moment of recharging the grout stops
flowing and thixotropic mechanisms act. This phenomenon is enhanced by the water absorption
out of the grout. Both actions reduce the penetration of the grout in the masonry.
Furthermore, the pumping installation needs to be sufficiently strong to overcome the
hydrostatic pressure. For high buildings, such as church towers, this hydrostatic pressure can
be quite important. The density of the grout is about 1.6 kg/dm3. This means that the hydrostatic
pressure reaches 8 bar to reach the injection holes at the top of the building. The installation
has to overcome the pressure losses in the conduit to maintain the injection pressure at the inlet
of the injection hole.
For large volumes, an electric pump is recommended. It is essential that this device enables a
rapid control with a facility to stop the pumping in a second. An suitable admission system can
provide the same safety without halting the pump. Manual pumping provides a better control of
pressure. They are more compact and lighter, so they can be positioned next to the injection
place.
C
The conduits [Figure 3.3, D and E]
Flexible tubes lead the grout from the ground installation to the infection hole. These tubes can
be considerably long, depending on de size of the building and the mobility of the ground
installation. To limit the charge losses it is recommended to take the diameter of the tubes large
enough. To decrease the amount of grout in the tubes, and hence the time it takes for the grout to
reach the injection hole, a smaller diameter is preferable. In a professional equipment a double
conduit is used. The use of a return conduit is strongly recommended as it prevents the grout to

28

Grout injection of masonry, scientific approach and modeling

stand still too long in the tube. Thus, the grout will not set inside the tube and the workers do
not need to stop and restart the pump all the time. The return tube ends in the collector.
C
The admission system [Figure 3.3, F, G and H]
The admission system is equipped with a joining part that enables a fast and watertight coupling
of the injection conduits with the injection hole. An adjustable three way gate valve limits the
injection pressure. If the counter pressure is too high the three way gate valve will direct the
grout to the return pipe. If the counter pressure is below the injection pressure the grout will
enter the injection hole. The use of this three way gate valve has two major advantages. The
pressure at the inlet of the injection hole is controlled close to the injection hole. This is better
for every control system, since it shortens the reaction time. The pump does not need to be very
complex. The pressure will never, even not for a short period of time, exceed the allowed
injection pressure and hence, the risk for further damage due to over pressure reduces. On the
other hand, the grout will never stop flowing. Whether the counter pressure is too high or
whether one has to switch from one completed injection hole to the next one, the grout will be
directed to the return pipe. By this double pipe system with three way gate valve, the grout is
renewed all the time. During injection, the three way gate valve will allow grout in both
conduits: part into the injection hole and part into the return pipe.
3.2.2. Reality about injection equipment
Above the ideal injection installation and its components are described. It can not always
economically be justified to bring into action such a complex and expensive installation. For
small scale projects one could prefer a manual installation or even gravity grouting could be
considered. For electric pumping a device to limit the pressure inside the masonry should be
present in any case. A pressure of one bar corresponds to a load of 100 kN/m2! One can
imagine that for higher pressures inside the masonry structure additional damage might occur.
For longer distance between pumping installation and injection holes, a return conduit should
be present. If there is no return conduit, the grout will stand still in the main conduit while
switching from one injection hole to another. This will cause the time depending features of the
grout to occur: instability and thixotropy. It depends on the quality of the grout to what extent
these phenomena will arise.
If the installation does not include a separate mixing installation, one has to stop the injection
after finishing a batch of grout from the collector, produce another batch of grout, and restart the
injection. In the chapter about the experimental program, it will be shown that it is nearly
impossible to restart an injection after a period of standstill. The experiments, as well as the
simulations show that a long stand still is one of the worst things that can happen.

3.3. Realization of an injection work

Chapter 3 - Injection as a consolidation technique for masonry

29

In this chapter the essential steps in a consolidation are described. The general procedure for
structural grouting of uncovered massive or double leaf masonry is considered here. In case of
covered masonry (masonry covered with frescos or plastered masonry) extra precautions are
necessary. They are not discussed in this work.
3.3.1. General Requirements
Before discussing in detail the different steps of an injection, it is appropriate to provide a
schematical overview of the requirements for the materials used and the technique. It will be
obvious from the next overview that one has to compromise between different requirements.
Depending on the character and the state of the masonry, the importance of some requirements
can change.
3.3.2. Diagnosis of the masonry
It seems quite logical: before there are any further steps to take, one should wonder if
consolidation injection is necessary or can be of any help to maintain a building. Maybe there
are other techniques, more suitable for the situation at hand. If consolidation can be part of the
restoration, one should determine the procedure to follow. To be able to answer this question
a thorough diagnosis is necessary. Presently, non destructive testing methods are, by
themselves, only able to provide a qualitative evaluation of the masonry. This implies that their
use alone is not enough to justify a grouting operation and to support the design of this
operation. Therefore, the diagnosis of the masonry is most of the time a mixture of nondestructive techniques and classical destructive techniques. The visual charts, produced by the
non destructive testing enable to locate the relevant areas where destructive tests should be
carried out. Non destructive tests are very suitable to obtain a qualitative picture of the
masonry structure. They are ideal to compare the initial state with the injected one.

30

Grout injection of masonry, scientific approach and modeling

Table 3.1: Overview of the general requirements for consolidation injection of masonry
Relevant properties of the grout

Requirement

viscosity, critical shear stress

as low as possible

Stability (no bleeding, no seggregation)

as high as possible

Adhesion to the injected material

as high as possible

Chemical and physical compatibility with the

original material

this is the main criterium to choose between

polymers, cement grout or lime grout

Penetration and injectability

As good as possible

Water retaining properties

As good as to avoid excessive absorption

Mechanical properties

Comparable to the original material

Execution of the injection

Remarks

Preparation of the masonry surface

to avoid leakage

Drilling of the injection holes

closest pattern is preferrable, density can

vary depending on the quality of the masonry

Prewetting of the masonry

only if absolutely neccessary to improve the

injectability

Introduction of the grout either by gravity or

pressure

avoid large internal pressure values

Modern diagnostic tools

Non destructive techniques for the preliminary investigation to decide about consolidation
and for the quality control after injection
Numerical and probabilistic methods to quantify the need, the benefit and the efficiency of an
injection

Destructive techniques
Among the destructive techniques, coring is probably the most frequently used. The coring
enables furthermore to judge the quality of the inner masonry by visual inspection of the cores.
Compressive and splitting tests can be executed on the cores to get an idea of the mechanical
properties of the masonry. Eventually the core hole can be inspected using an endoscope.
Another destructive technique is offered by the resistographical method. A hole is drilled in
the masonry and the force needed to advance is plotted versus the depth. Hard parts of solid
and sound masonry need a high force to advance whereas soft and deteriorated parts only need
a small force. This method causes less damage to the structure: only some small drilling holes,
and provides a transverse scope of the inner masonry.

Chapter 3 - Injection as a consolidation technique for masonry

31

Non-destructive techniques
Since coring is an intensive job but especially since coring is actually damaging the building,
one has to minimize the number of cores to take. Destructive testing therefore can only serve as
a calibration of the data from the non-destructive analysis.
Popular non-destructive testing methods are the sonic [M. Schuller, 1995] and ultrasonic
technique, likewise called acoustic tomography, measuring the transmission speed for a
longitudinal displacement wave inside the material. Sonic measurement is a low cost method
enabling a quick qualitative information about the masonry state. The basic principle is that the
velocity of the sonic wave depends upon the state of the material. The pulse will be transmitted
through the material. Cracks and voids imply that the wave has to find another, longer way and
this increases the transmission time, or that the wave has to pass the crack through a layer of
air, which has a much lower transmission velocity than the surrounding material. Sonic tests
are suitable for detecting voids and cracks in walls of great thickness due to the strength of the
signal.
Ultrasonic signals are much more sensible to surface conditions and the loss of energy, that is
typical for a high frequency wave, leads to a rapid attenuation of the waves. Therefore, the
ultrasonic measurement is only suitable for limited transmission distances.
As for most other non destructive testing methods, no constant relation exists between the
(ultra)sonic measurements and the mechanical parameters or porosity values. However, the
sonic measurements provide an idea of the state of homogeneity of the masonry. Therefore, the
technique is suitable to compare the initial state with the state after injection. It is then possible
to see to what extend the injection was able to overcome the heterogeneity in the masonry.
The Building Material Division has concentrated on the electrical resistivity measurement
technique. The technique has successfully been applied to judge the condition of Het
Gravensteen in Ghent [D. Van Gemert, 1988]. It consists in measuring the electrical
resistance of the masonry, based on the equation of Ohm-Pouillet:
V=R.I
The presence of cracks inside the masonry increases its electrical resistivity. Different
configurations are possible, but the equipment needs at least two electrodes and two probes.
The electrodes introduce a current in the structure and the probes measure the potential
difference between them. The values are a function of the masonry properties in-between them.
A circle with a larger diameter corresponds to a configuration with a larger distance between
the electrode S1 and the measuring electrodes. Hence, the bigger the distance between the
probes, the more the measurements reflect the state of the masonry that is located deeper inside.

32

Grout injection of masonry, scientific approach and modeling

Figure 3.4: Schematic presentation of the Schlumberger configuration for geo-electrical

measurements
The information on the internal state of the structure, visualised in an electrical resistivity map,
is partly masked due to the influence of the limited dimensions of the analyzed masonry. This
observation led to the idea of making maps in which the influence of the boundaries would be
eliminated [H. Janssen, 1993] [K. Haelterman, 1993]. The resulting map can then directly be
correlated to the real properties or anomalies of the structure [D. Van Gemert, 1998].
The Arenberg Castle is surrounded by a brick masonry wall. The technique to filter the geoelectrical resistivity measurements from the geometrical information was applied to this wall.
Figure 3.5, Figure 3.6 and Figure 3.7 provide an example of the filtering of the geometrical
influences on the resistivity measurements. As can be seen from the map of relative
differences on Figure 3.7, the only heterogeneity that is left, is a thin horizontal layer. This
horizontal layer physically corresponds to the interface between two kind of stones with a
different resistivity. This interface gives rise to an accumulation of iso-resistivity lines.

Chapter 3 - Injection as a consolidation technique for masonry

33

Figure 3.5: Electrical resistivity map of part of the wall in the Arenberg Park (Horizontal
distance = 0.00, vertical edge)

Figure 3.6: Electrical resistivity map for a homogeneous wall, horizontal distance = 0
corresponds to the left edge

34

Grout injection of masonry, scientific approach and modeling

Figure 3.7: Map of relative differences, reflecting only the relevant information: the interface
between two kind of stones used to erect the wall

Other influences disturb the relevant information. The humidity influences to a high extend the
resistivity values. The more humid the masonry, the lower the resistivity. The filtering of the
moisture content from the resistivity maps and the uncertainty about the accuracy and the
resolution has been studied by Kathleen Venderickx [K. Venderickx, 1996]. She worked out a
method to filter the information from these unwanted disturbing influences.
Radar techniques that use electromagnetic signals for the investigation of masonry structures
provide charts that are similar to the ones obtained from electrical resistivity measurements.
They also enable to judge the homogeneity of the structure. Nevertheless, the technique requires
more expensive equipment. The acquired information is not very detailed in the sense that the
resolution is poor [C. Colla, 1995]. The presence of water attenuates the electromagnetic
waves in order to reduce the maximum thickness of the wall that could me measured. Besides,
this attenuation masks the relevant information.
Non destructive test data, related to physical evidence, enable to draw a map showing an image
of the transition time for acoustic testing or the electrical resistivity of the wall that was
scanned. This way the heterogeneity of the wall is displayed. This qualitative information is
calibrated using the test results of the destructive tests. This way the engineer can decide if, and
in which areas, consolidation injections are necessary and what strength gain is desired without
coring the hole building.
Probabilistic methods [L. Schueremans, 1997] as mentioned above in Chapter Chapter 2 allow
to calculate the probability for the structure to fail. If the probability to fail is too high,

Chapter 3 - Injection as a consolidation technique for masonry

35

consolidation has to be considered. Eventually the effect of the consolidation of certain parts on
the probability to fail can be studied.
3.3.3. Preparation of the masonry
Once the areas to inject are determined, the building is prepared for injection. An important part
of this preparation is the sealing of the masonry to prevent the leakage of the grout [J. Ashurst,
1989] [J. Ashurst, 1990]. Therefore, a general repointing is recommended. Besides, a general
deep repointing can be seen as a very effective structural intervention. The repointing as such
already consolidates the masonry. The pointing must be fairly porous to absorb the water of the
injected grout. This will improve the setting of the grout and the adhesion to the masonry. Van
Gemert [D. Van Gemert, 1988 (1)] refers to the casing of the masonry using a cement mortar or
a coating based on soluble gels for outside masonry. Cellulose or clay based temporary
coatings are alternatives to these systems. After consolidation they can be washed by water.
Nevertheless leakage can still occur and should then be stopped using quick-setting cement and
cleaned immediately . Leakages prevent to build up pressure inside the flow channels. If the
stability of the masonry is very doubtful and if hydrostatical pressure is feared to arise, an
external reinforcement can be justified.
Second part of the preparation is the drilling of the injection holes. Preferably the holes are
drilled in the joints. This way they will be less visible afterwards. The holes should incline
downwards. Three parameters of the injection holes are important: the pattern, the density or the
distance between two adjacent holes and the depth of the holes. The parameters depend on the
type of masonry, the overall condition of the masonry, the rheological properties of the grout and
incidence of cracks. A more precarious zone with many cracks will be easier to inject and
hence the pattern in this zone can be somewhat less dense. On the other hand the applied
pressure in this zone should be lower. Therefore and for reasons of simplicity the density is
often kept constant for the whole structure. Existing major cracks can easily be used for
injection.
The pattern can be square Figure 4.7 or staggered Figure 4.8, considering a cylindrical
injection, spacial geometry learns that a staggered pattern theoretically increases the covered
surface with around 11 % in comparison with a square pattern. In literature nothing more than a
few general guidelines are available. The density of the injection holes is expressed as holes
per square m. The recommended number of holes per square meter found in literature are
mentioned in Figure 3.7.

36

Grout injection of masonry, scientific approach and modeling

Table 3.2: Layout of the injection holes pattern found in literature

Reference

number of holes / m2

Pattern

Lizzi
[1982]

2-3

Pume
[1989]

2-4

Oblique holes, lengte equal 2/3 of the wall

thickness, diameter of 33 mm

Zagorcheva
[1988]

3-4

Distance 0.5-0.7m in vertical direction and about

0.5m in horizontal direction, diameter of 12 mm

Tomazvic
[1992]

0.5m to 1.0m interval between holes, at least half

the thickness of the wall deep

Binda
[1991]

Injectors inserted till 2/3 of the walls thickness

The holes should at least reach the middle of the wall. Two-third of the wall would be better.
Above some general rules about injection hole density and depth are mentioned. Case studies
learn that it is hard to find rules that are generally applicable. An experimental program carried
out by Baronio [G. Baronio, 1992] about masonries with cracks and voids irregularly
distributed and of different dimensions showed them difficult to inject properly. Therefore,
Baronio states that the usual 2 to 4 holes per square meter are not enough. A possible
explanation can be formulated intuitively. During injection, when a grout reaches a large void,
no pressure can be built up in the neighborhood of that void. Due to this low pressure, the grout
will enter the fine cracks only over a short distance. Thixotropy, water absorption and
instability of the grout cause the blocking for further injection in these finer cracks. When the
large void is finally filled, the pressure can increase again, but too much water of the grout is
absorbed in the fine cracks to restart flowing. The zone hidden by the finer cracks will never be
injected through this one hole. So she advises to shorten the distance between holes. The
covered area for one injection hole depends on the penetration of the grout inside the masonry.
The calculation of this penetration of the grout starting from the properties of the grout, the
injection pressure and the permeability of the masonry is one of the main goals of this thesis.
3.3.4. Injection pressure
One of the most important parameters of injection was already mentioned above discussing the
injection holes. Of course the type of grout and it rheological properties on one hand and the
quality and properties of the masonry on the other hand stay, by far, the most important
parameters. Their influence on the grouting process will be discussed thoroughly, but not in this
technological chapter. The injection pressure is the pressure at the inlet towards the injection
holes and can be quite different from the pressure generated by the pump. The injection

Chapter 3 - Injection as a consolidation technique for masonry

37

pressure is the driving force behind the penetration of the grout inside the masonry. The higher
the pressure the easier and faster the grout will pass. Because the grout flows faster, the grout
will loose less water by absorption and the particles will remain better in suspension.
However, the pressure is limited to a few bars. The internal pressure of the grout blows up the
masonry introducing tensile stresses that can not be taken by the masonry.

Figure 3.8: The hydrostatic pressure adds to the injection pressure and
might cause additional damage to the masonry
Increasing the pressure would soon cause additional damage to the structure. The pressure built
up inside the masonry is the addition of the injection pressure and the hydrostatical pressure.
The hydrostatical pressure is proportional to the height of the injected grout column that is still
in fluid state, Figure 3.8. Therefore, it is recommended not to inject the injection holes in a
vertical order. InTable 3.3 some recommended values for the injection pressure are given.

38

Grout injection of masonry, scientific approach and modeling

Table 3.3: Injection pressure values found in literature

Reference

Pressure

Type of grout

Site

Feilden
[1982]

2.0 bar

lime + fly-ash + cement

Central tower, Norwisch cathedral,

England, stone masonry

Vogiatzis
[1989]

1.0 bar

cement + fine sand + SP

Church of holy Apostles in Kalamata,

Greece

Tomazevic
[1992]

2.0 bar

cement + puzzolane

Rubble stone masonry buildings,

Kozjansko, Slovenia

Binda
[1993]

3.5 bar

Hydraulic lime +
additives

Irregularly cut stone masonry,

Rovereto, Italy

Typical compressive stresses in ancient masonry are about 1 MPa or 10 bar. Tensile stresses
are close to zero. Internal hydrostatic pressure might cause big tensile stresses or might push
out the outer leaf of the masonry structure.
3.3.5. Execution of the injection
The injection itself starts at the lowest injection holes, in the middle of the wall. Then it
progresses sidewards before the next level of injection holes can be is injected. Thanks to the
feedback conduit, the switch between two injection holes takes place without any problem: the
grout keeps on flowing, there is no risk for setting in the conduit. A quick joint enables a fast
coupling and decoupling from the nozzle to the taps.
The filling of one hole continues until
C
the pressure exceeds permanently the injection pressure
C
the grout emerges freely at adjacent injection holes
C
a predetermined quantity of grout is injected in that hole
The latter stop criterion is used in case the grout flows away through an invisible leakage. An
invisible leakage can not be sealed by the normal procedures, but can be stopped by injecting a
fast setting grout or a grout with a high critical shear strength value (plastic threshold).
Furthermore, it is advisable to register the amount of grout that is injected in every injection
hole. These data can be analyzed in order to check for a complete filling of the masonry, to
control if no grout ran away. In spite of a good preparation of the masonry, it can not be
excluded that leakages occur. A fast sealing of these leakages using fast hardening cement is
advised. These leakages are almost inevitable and they provide information about the flow of
the grout. Three workers are required for a good execution of an injection job. One worker
serves the mixing installation and the pump. He takes care for the grout to be available all the
time in the collector by mixing in time the successive batches of grout in the mixing installation
and to continuously pump the grout towards the injection hole. If necessary, he has to stop the

Chapter 3 - Injection as a consolidation technique for masonry

39

pumping activity or to adapt the pressure. At the end of the day he has an important role in
cleaning the installation. The other men work in the neighborhood of the injection hole. One of
them handles the conduits and takes care of connecting the conduits to the injection hole. He
also handles the three way valve to adjust the injection pressure. The third person helps this
second man in moving the conduits and seals the occurring leakages. If the grout that streams out
of the leakage smudges the facade of the building, it should be washed away before hardening
takes place.
When the work has been completed, the injection holes need to be repointed. For this purpose,
the drilling powder that was collected during drilling the injection hole, can be used to add to
the binding agent to assure the same tint.
3.3.6. Control of quality
Although general guidelines exist, technological parameters and the composition of the grout
depend highly on the specific situation. To insure a high quality work, the masonry building
must be carefully studied. The best way to prevent problems and to avoid a poor quality is a
good preparation of the job. An extensive study should be carried out. This study results in the
determination of the following items:
C
specification of the preparation of the masonry
C
the pattern of injection holes
C
the injection pressure
C
the composition of the grout
C
the mixing procedure of the grout: sequence and mixing time
For every batch of grout one should check the fluidity by a flow-test type Marshall funnel or
Afnor cup. Additionally one could check the stability as described further in this thesis.
In the ideal situation, the quality of the work is controlled on line. This means that during
injection, the flow of the grout is monitored. If things go wrong, one should be able to interfere
as soon as possible. Two important conditions must be fulfilled for the on line system of
control and correction of the injection. In first instance the on line control needs to be worked
out. Until now, these on line control techniques are not operational. Though, it should be
possible to realize on line control using the electric resistivity method. The presence of
humidity and hence of the grout, influences to a large degree the electric resistivity of the
masonry. The penetration of the grout suddenly decreases the electrical resistivity of that part of
masonry. It should be possible to detect and visualize the decrease in resistivity and hence the
arrival of the grout in a certain point. So, let us suppose that it is technically possible to monitor
the grout penetration. In case of problems one has to react promptly. First possibility: the
penetration is insufficient. In that case possible reactions are: the use of a different grout
composition with an improved rheological behavior; an increase of the injection pressure;
additional injection holes etc...For this purpose a reliable model of the grout flow can help to
decide what action to take. The present thesis offers such a model.

40

Grout injection of masonry, scientific approach and modeling

If on line control cannot be done, as it is in the actual state, off line control is necessary. The
injected region should be checked. One possible and reliable method is to use the same non
destructive testing method as was used for the diagnosis of the masonry structure. It is possible
to compare the maps that were obtained before the injection with those obtained during the
control measurements after injection.
Ultrasonic measurements, electrical resistivity
measurements or radar technique were discussed above. In case of electrical measurements one
should take care about the influence of the humidity on the measurements. Since a large amount
of water is brought in by the action of injection, the resistivity will be changed. Research is
going on to filter the obtained information from the influence of water. In case of a successful
injection work, the heterogeneity in the resistivity map will fade away because of the filling of
the voids by the injection. Any non destructive testing is best completed with destructive test
e.g. coring.

Figure 3.9 : Electric resistivity map before and after cement grout injection in two
points: the resistance decreased significantly [D. Van Gemert, 1988(2)]

Chapter 3 - Injection as a consolidation technique for masonry

41

Figure 3.10: A map of the difference in resistance [Ohm.m] before and after the injection of an
epoxy resin, the measurements reflect the material property to one meter inside [D. Van Gemert,
1988 (2)]
Figure 3.11 displays another nice example of the use of non destructive testing for the control of
the effectiveness of consolidation injections. The study is carried out by Atkinson-Noland &
Associates, USA and provides the three dimensional surface plot of through-wall ultrasonic
pulse arrival time measurements for the original, damaged and repaired condition of a
laboratory masonry wallet. For on site application, the first image will not exist. However, the
second image of the wallet in damaged condition can be analyzed in order to decide about
injection. To control the execution of the injection the third image indicates that the injection
provided a uniformly filled masonry structure, with an almost constant pulse arrival time for the
ultrasonic waves. The pulse arrival time is the time that passes between the moment the pulse
was sent and the moment the receptor notices the arrival of the pulse. When the voids are filled
with hardened grout, the pulse arrival time will generally decrease. Furthermore, the pulse
arrival time will show less scatter than before the repair.

3.4. Types of binding agents

3.4.1. Polymers
Polymers are the most recent binding agents used for consolidation injections. Polymers are
pure liquids (no dispersions like all the other binding agents used for injection purposes) and
show a relatively low viscosity and low critical shear value. Since a polymer is a pure liquid,
it will not suffer from thickening due to water loss. The wide range of viscosity values and the
absence of particles that might hinder the flow, make polymers very suitable for injection.
Moreover, polymers show a very good adhesion to dry surfaces and have a high compressive
strength and, even more important, a high tensile and adhesion strength. This makes them very
suitable for particular cases were the penetration of hydraulic grouts is problematic.

42

Grout injection of masonry, scientific approach and modeling

Figure 3.11: Plots of the pulse arrival time for ultra sound measurements, original condition (a),
after loading (b) and after grout injection (c) [R.H. Atkinson, 1991]
Beside these advantages of polymeric grouts, one has to mention the weak adhesion to wet
surfaces. The physical properties (water permeability, thermal expansion) of these grouts are
completely different from the original materials. This could cause problems with the moisture
transport or could cause thermal stresses.
Polymers are very expensive materials when compared to other possible injection materials.
When large quantities are to be injected in a historical building, an injection with polymers will
soon become economically unfeasible. Therefore, the Building Materials Division advised on
different occasions to inject the structure first with a mineral grout. A second injection with
polymers can than fill the remaining voids.
Probably the biggest problem that polymers are facing is that they are considered not to
correspond to the materials that were historically used. Many architects tend to favor materials
that are more compatible: these modern chemical materials do not belong in valuable historical
buildings [The Charter of Venice, 1964]. This is not only a philosophical statement, but can be
motivated with scientific physical arguments: the material has a completely different stressstrain behavior, it is impervious to water and has a different thermal dilatation factor.
Polymeric grouts show certain disadvantages discussed above but regarding strengthening, they
are the most efficient ones. This fact is undoubtedly connected with the outstanding mechanical
properties and the good rheological properties of these grouts that are associated with an
absence of solid particles in suspension.

Chapter 3 - Injection as a consolidation technique for masonry

43

3.4.2. Cement grouts

Cement is the most popular binding agent of modern times. Cement is not expensive, the raw
materials are widely available. From a more technological point of view: the fast hardening of
cement allows to proceed faster in applying new loads to the structure.
Basically a cement grout consists of cement and water. The same combination of cement and
water is applied for soil consolidation. In case of soil injection very high pressures, up to 100
bar, are applied. For the injection of masonry structures, the injection normally proceeds at
pressures lower then 2 bar. The cracks and voids through which the grout has to penetrate can
be relatively small. Such a cement grout needs a good fluidity. To achieve this a
superplasticizer is added to the mix. The superplasticizer enables to reduces the water content
for the grout to an acceptable level without damaging the required fluidity. Besides,
superplasticizers have a deflocculating action. This prevents the flocculation of cement
particles. Superplasticizers based on melamine formaldehyde and sulfonated naphthalene
formaldehyde are the most popular. They decrease the electrostatical forces that keep the
particles together.
In addition to superplasticizer, other additives or admixtures are applied. Because of the nature
of a cement grout, a dispersion of cement particles in water, the bigger cement particles sink.
This phenomenon creates a heterogeneous consolidation. The lower region contains more
cement and will be harder and stronger, whereas the higher region will be weak due to a lack of
binding agent. The top of the injected zone could, in case of severe bleeding, only be filled with
water. Stabilizing agents prevent the bleeding of the grout or the segregation of the cement
particles. An extended study of A. Miltiadou [A. Miltiadou, 1990], A. Miltiadou, 1991] and a
study of A.M. Paillre [A. Paillre, 1986] indicate that the addition of ultra fines improves the
injectability of a grout. These ultra fine puzzolanic admixtures correct the granularity of the
cement in the finer region and influence positively the injectability and the stability of the grout.
A finer dispersion of cement particles in the water will enhance the flow of the grout. The grout
will be able to progress longer inside the masonry before it will be blocked. On the other hand,
the finer the particles, the more water needed for a good fluidity and hence addition of
superplasticizer is inevitable. Depending on the application and the concrete situation, a type of
cement with modified granularity can be necessary. This can be achieved by adding ultra fine
mineral admixtures.
Strictly spoken, cement does not correspond to the binding agent used in most historic masonry
buildings. The nature of the historical binding agents is air hardening lime or natural hydraulic
lime. Therefore, art historians and architects disapprove of the injection of cementitious grouts
in historically valuable buildings. The arguments come down to the aversion of putting a
material, strange to the original materials, into cultural patrimony. To my point of view this
argument is less valuable than in the case of polymers. Cement is a mineral binding agent, just
as lime. The physical properties with regard to moisture transport, thermal expansion,
temperature household etc... are much closer to those of the historical materials than in case of
polymers. Many buildings have been injected satisfactory using cement grouts.

44

Grout injection of masonry, scientific approach and modeling

Sometimes, in specific situations, it is possible to use special kinds of cement. A blast furnace
slag cement is always preferable to ordinary Portland cement because less of its C3A -content is
reducing the risk for the formation of ettringite and with low alkali reducing the danger for
efflorescences. Hydration is a little slower, and less danger exist for possible expansive
reactions. Furthermore, blast furnace cement is somewhat finer than pure Portland cement. If
the masonry to consolidate is hard to inject due to fine cracks and voids, one could consider the
use of micro cement. Micro cement is a fine ground kind of cement frequently used for soil
injections. The extreme fineness, Blaine specific surface value of 8000 cm2/g and higher,
against around 3000 cm2/g for OPC, provide a very good injectability. Based on our experience
more water is required and the addition of more superplasticizer is necessary. During our
experiments, we were not able to produce a stable micro cement grout without the addition of a
separate stabilizer. The additional amount of water will sooner or later evaporate from the
masonry structure. This could cause problems for the moisture household and if soluble salts
are present, could cause crystallization effects and efflorescence. However, the results of other
researchers mention very successful injections using micro cement grouts. Miltiadou studied the
strength increase of a traditional rubble core, through the injection of ultrafine cement based
grouts. The results show a remarkable increase of strength and stiffness. The average
compressive strengthening increase factor was seven, for stiffness it was five. The question
remains what would happen in case of an additional settlement when the stiffness has increased
so much. One could state that micro cement requires more professional skill than normal
cement. Besides, micro cement is more expensive. The use of micro cement is more demanding
for the mixing procedure: a badly mixed micro-cement grout may be worse than a badly mixed
normal cement grout. An extensive study has been done by O. Benhamou on the rheology of
grout using microcement [O. Benhamou, 1994]. From this study the behavior of microcement
grouts seems to be complex and sometimes unpredictable.
3.4.3. Lime based grouts
Lime based grouts are without any doubt the grouts the most compatible with the original
materials for the consolidation of ancient masonry. For centuries lime of both types, air
hardening and hydraulic lime, has been used for the construction of buildings. The use of lime
based grouts for consolidation of masonry should be very popular. Tough there exist several
contra-indications.
The carbonation of thick layers of lime is a very slow process due to the slow diffusion of CO2
through the carbonated layer. The air hardening lime can be given hydraulic properties by
blending with puzzolanic material. Several natural products are available for this purpose:
Trass or Santorini Earth. Also cement provides hydraulic properties to the grout. Hydraulic
lime is an even better alternative. The hydraulic properties tend to give an acceptable early
strength whereas the lime provides the wanted ductility making the grout suitable for the
application in seismic areas.

Chapter 3 - Injection as a consolidation technique for masonry

45

Only few studies have been done on the rheological properties and efficiency of these grouts.
Our own research reveals that the water content of a fluid lime based grout is very high without
the use of an appropriate superplasticizer, that we could not find Figure 10.1. The samples that
we injected with the lime grout obtained very low mechanical strength, although the grout was
very injectable. We will discuss the results in the chapter about the experimental program. Due
to its fineness (Blaine Specific Surface > 8000 cm2/ g) lime based grouts are very injectable in
fine cracks.

3.5. Grouting improves the load bearing capacity of the masonry

Masonry is a heterogeneous material, composed of a relatively hard component (the bricks or
stones) on one hand and relatively soft mortar on the other hand. By its specificity it is not easy
to predict the behavior of the whole starting from the properties of the components. This
statement is true for the physical behavior (moisture transport, heat transfer or migration of
salts) as well as for the mechanical behavior. The mechanical behavior includes the functioning
under loading, the attitude when a setting occurs or, for seismic regions, the interaction between
soil and structure. Grouting mainly intends an improvement of the mechanical behavior, though
the impact on the physical behavior may not be disregarded.
Many rules for dimensioning masonry structures are based on empirical rules. Those rules are
only approximately applicable. The reader will understand that it is nearly impossible to
calculate the efficiency of a consolidation injection. Literature provides some global
consolidation factors, deduced from experiments, part of which are presented above. The load
bearing capacity after injection is compared to the load bearing capacity of the structure in
undamaged condition (Eq 2.7). Depending on the original state, the injection and the
composition of the grout, the strength improvement is 40 % to 150 %. Extrapolating these
values to real masonry is not very reliable. The factors give a notion of the possibilities of
injection consolidation. All authors agree that the consolidation factor increases as the initial
condition of the structure is poorer.
The strengthening effect is due to:
C
Internal voids cause tensile stresses in the material. By filling the voids, those tensile
stresses can not give rise to extra tensile cracks.
C
The injection of additional binding agent into the structure will enhance the internal
cohesion. Especially the zones that show poor cohesion due to fine cracks and loose
material will be infiltrated by the grout and be strengthened. The resistance to splitting
and to shear will improve.
C
Monolithical behavior after injection: this is the way masonry structures are designed.
Generally, a lime grout injection provides less strength improvement than a cement based grout.
But it has to be mentioned directly that strength is not the only property that is involved when it
comes to the consolidation of masonry. The strength improvement is most of the time connected

46

Grout injection of masonry, scientific approach and modeling

to a loss of ductility. This loss of ductility is much higher in case of cement grouts and hence,
the masonry fails to maintain its self-healing property and its ability for auto-equilibration of
forces. However, it is clear that grouting is an effective method to improve the load bearing
capacity of solid or double leaf masonry. The original mechanical strength can easily be
achieved. Concerning the masonry support, there is a clear difference between the efficiency of
grouting a double leaf masonry wall with a rubble core and a solid brick or stone masonry. The
explanation to this fact remains on the much higher level of porosity in the first case. In double
leaf masonry, not only the voids are bigger but also the channels of communication between
voids are much more than in the second type. The importance of porosity is easily understood
considering that the main principle of grouting is to fill the inside voids. Concerning the grout
itself one important factor is its injectability - it explains why ultra fine cement based grouts are
more effective than cement based grouts with SP and these more effective than pure cement
grouts. However, this parameter by itself is not enough to explain why modified lime based
grouts are less efficient than cement based grouts or the different results between polymeric
grouts and ultra fine cement based grouts. Lime based grouts modified with SP have probably a
similar rheological behavior of cement based grouts. Moreover, the injectability of epoxy
grouts and ultra fine cement based grouts is similar. Thus, the explanation must lie in the
intrinsic mechanical properties of the grout. Concerning lime based grouts, the fact that
hydraulic lime or hydrated lime plus puzzolanas are used, enables the slow process of
carbonation of lime by a puzzolanic reaction to be overcome. However, there is always free
lime that needs carbonation to set. As CO2 has difficulties in reaching the inside of the masonry,
the process is slow, explaining the low efficiency of these grouts, and suggesting that it will
slightly improve in long term [K. Van Balen, 1991]. With polymeric grouts, although the
adhesion properties are similar to ultrafine cement based grouts, their intrinsic mechanical
properties are significantly superior and this factor only can explain their outstanding behavior.
Nevertheless, a grout with high intrinsic resistance is not always needed, since the final
resistance of masonry is also a function of its own mechanical properties.
If the results of polymeric grouting are compared with those obtained by fine cement grouts, it
can be seen that the latter can be a real alternative to polymeric grouts, when their demanding
mixture procedure will be available on site. Firstly, because the actual knowledge of polymeric
grout behavior clearly suggests it is too early for their unlimited use in historic buildings.
Secondly, the cost of cement grout is significantly lower, which is particularly important for
consolidation of high porosity masonry, such as double leaf masonry with a rubble core. For the
moment, whenever a significant mechanical improvement of a historical building is needed,
cement based grouts with SP are a good choice. They are effective and the mixing procedure is
not so demanding as for ultra fine grouts. Concerning lime based grouts, they can restore the
original mechanical properties of a double leaf masonry. Therefore, their use is a real option,
whenever only a repair operation is envisaged. However for solid brick masonry, if the
relation that exists for the other grouts between the strength factors of the two types of masonry
is applied, a average strength increase factor of only 1.3 is obtained, suggesting that they are not

Chapter 3 - Injection as a consolidation technique for masonry

47

adequate to this type of masonry Table 3.4. Finally, it must be emphasized that these values are
only reference values, once they are a function of parameters that cannot be controlled, and thus
they can never be used on a design phase. The porosity is not constant within a masonry's
global type, nor can the rheological and mechanical properties of the grouts be assured to be the
same. Nevertheless, they enable a first approach of the expected improvement by grouting and
clarify our expectations.
Table 3.4: Average mechanical strengthening factors found in literature [N. Gil, 1995]
Lime based
grout

Cement grout
with SP

Ultra fine cement

grout with SP

Polymeric grout

Double leaf masonry

with rubble core

2,5

Solid stone or brick

masonry

1,3

1,6

2,7

The more damage happened to the structure, the more both effects, increase of average strength
and the reduction of the variance of the strength, can enhance the behavior of the building.
Therefore, the strengthening factor needs always to be related to the specific situation.
European historic masonry is often composed of two leafs with a very open core inside. This
core consists of rubble on which fluid mortar was poured. The load bearing capacity of this
inner part of the structure can be significantly improved by filling the voids. Besides, by filling
the inner part, both leafs become interconnected. This provides a nearly monolithical unit that
withstands very well mechanical loadings.
The mechanical strength of the injected grout for sure has influence on the expected
improvement. Nevertheless, this influence is less important than one might expect. Even if the
voids are properly filled with a grout with relatively poor mechanical properties, the structure
will regain a lot of strength because of the decrease of the variance of the strength. M.
Tomazevic [M. Tomazevic, 1992] injected wallets with four different kind of grouts; the
compressive strength of the grout varied from 7 MPa to 32MPa. Still he found that the
mechanical strength of the hardened grout did hardly influence the final mechanical behavior of
the injected wallets. When using a grout that is too hard or too rigid after curing, additional
settlements might cause splitting forces just above the injected holes. More important for the
effect of a consolidation injection is the complete filling of all the voids and cracks. If this is
not the case, additional concentrations of stresses will occur that can cause additional damage to
the injected or neighborhood structure.
The reasoning developed above implies that the flow properties of the grout, the rheological
properties, deserve more attention than the mechanical strength. We will extensively come back
to this issue later in this thesis. The quality of the carrying out of the injection procedure has

48

Grout injection of masonry, scientific approach and modeling

major influence on the final result of the consolidation. The presence of local stiffness in a
global structure can cause severe damage in case of future solicitations of the monument.
The problem must be studied in a probabilistic way. This is part of the research in the Building
Materials Division. In paragraph 2.3.1 a model for calculating the increase of the reliability of
a masonry structure by injection is presented. The increase of the average strength and the
decrease of the variance on the strength reduce the probability of failure.

3.6. Masonry grouting, code of practice

A reading of codes of practice of several countries, dating from the 1970's, showed a complete
absence of references to the retrofitting techniques of masonry structures.
The early official mentions of repair techniques to masonry buildings dates of 1986, in the
Italian technical standards for construction in seismic areas. These norms had their origin on the
official requirement to guide the immense retrofitting interventions after the earthquakes of
Friuli (1976) and Campania (1980). They define two levels of intervention.
The first level is the upgrading intervention, obligatory in case of vast interventions that can
lead to changes in the global behavior of the building. The norms define it as a substantial
intervention on the masonry building to make it resistant to the official seismic loads.
The second level is an improvement level, obligatory in case of renovation or change of
masonry elements in the structure. The norms define it as the execution of one or more
operations on the masonry structure to improve its original mechanical properties in at least 20
percent, without changing the global behavior of the building. They only mention grouting as one
possible intervention technique~ in the aim of this second level . This code clearly illustrates
the poor state of knowledge about grouting at that time.
ln 1991, the city of Los Angeles developed a detailed specification for the repair of cracks with
grouting, but it only concerns the technological point of view. They give a specific volumetric
grout mix and fully describe the technical parameters, and the quality control measures [City of
L.A., 1991].
A thorough approach to this matter only took place with the issue of the Part 1.4 of the
Eurocode-8, concerning the repair and strengthening (against seismic actions), in 1993
[Eurocode 8].
In this document, criteria for the assessment of the seismic performance of existing structures are
mentioned. The code presents a decision-making process regarding the corrective measures to
take and criteria for the repair and/or strengthening of existing structures.
At last, there is a specific chapter, concerning the monuments and historical buildings, where the
cultural level to preserve is officially mentioned for the first time.
Following, a concise reading of the code is presented, regarding its position to grouting as a
retrofitting technique of existing masonry structures and in particular to monuments and historic
buildings.

Chapter 3 - Injection as a consolidation technique for masonry

49

The code first emphasizes the importance of a correct structural evaluation preceding the
intervention itself it presents a methodology to undertake this evaluation referring the following
items:
C
a historical inquiry must be carried out to help understanding the present structural
scheme and to improve the knowledge of the past behavior, especially with reference to
earthquakes;
C
a detailed recording should be done to correctly evaluate the present situation.
Destructive and non destructive tests must be used to estimate the integrity and
mechanical properties of the masonry.
C
if grouting is considered, the code specifically recommends the determination of the
chemical properties of the mortars, to avoid detrimental reactions, particularly in the
presence of sulphates.
Grouting is mentioned as a possible intervention to mechanically improve masonry, in particular
in the repair of cracks and on the strengthening of rubble core masonry walls.
To repair cracked masonry, the code refers cement grouting, if the width of the cracks is small
(less than 1 mm) and the thickness of masonry is considerable. it recommends the use of
admixtures preventing shrinkage and proposes the use of epoxy grouting for fine cracks.
Lastly, it suggests cement grouting as an efficient method for strengthening rubble core walls,
with the condition of a satisfactory absorption.
Concerning the quality assurance of interventions, it mentions the following control measures for
grouting interventions:
C
protective surface treatments must be assured and final cleaning methods. The
effectiveness of these technologies must be checked on trial areas;
C
inspection of certificates of filling materials and, possibly, acceptance tests
(composition, stability, conditions of use);
C
measurement of local strains and control of deflections produced during grouting
procedures;
C
visual inspection of the final work; possible extraction of cores across selected check-areas, to evaluate the efficiency of grouting.
The reading of the Eurocode-8 clearly shows a full but only qualitative official acceptance of
grouting as a retrofitting technique. However, the present knowledge about the technique enables
a deeper approach instead of only general considerations.
In future, the code could more clearly present the differences between the grouts available,
namely between the cementitious grouts and polymeric grouts. It should present the several types
of cementitious grouts available and the levels of injectability associated. The code simply
ignores the existing (micro) cement grouts with SP, when it only makes reference to the epoxy
grouts for injection of fine cracks.

50

Grout injection of masonry, scientific approach and modeling

The remark concerning the efficiency of grouting to strengthen masonry walls with rubble core
is correct. However, the absorption remark is very vague and it may lead to misunderstandings.
The recommendations about the methodology to follow in an intervention, although very
concise, are pertinent and point the main factors to take into consideration.
The chemical incompatibility of normal cement grouts with a sulphated environment is
mentioned, but the risk to form the expansive ettringite or to create efflorescence should be
explained. Also, for quality assurance, the code concisely points the most important aspects to
take into consideration.
The lack of quantitative data over the effectiveness of the masonry strengthening reflects the
prudence philosophy subjacent to the whole code. This attitude is also reflected on the sole
mention of core extraction to evaluate the efficiency of grouting. This prudence can be explained
as masonry grouting works with a parameter not yet well understood - the masonry. Therefore, it
cannot be limited to a design in the office, but it utterly demands calibration of the results on site
and a tight quality control to ensure the final efficiency.
However, a future code should emphasize the existence of non destructive methods to control
the global efficiency of masonry grouting, when associated with core removal. This way, it
would help to change a still common attitude of suspicion towards grouting, mostly supported on
ignorance of its present state of knowledge.
Concerning the historical buildings and monuments, Eurocode 8 mentions the cultural level to
preserve, beyond safeguarding the human lives involved, defining a new safety concept.
Thus, a "Monument's safety level" corresponds to a situation where the maximum probable
earthquake B is only expected to produce repairable and not fatal artistic damages.
The intervention techniques proposed for a monument should fulfill the following requirements:
C
Effectiveness
Shown by qualitative or numerical proofs.
C
Compatibility
From the mechanical, chemical, technological and architectural point of views.
C
Durability
Comparable to that of the other materials of the building, unless a periodic
replacement is foreseen.
C
Reversibility
As far as possible, to allow for different future decisions.
Concerning the effectiveness of masonry grouting, it can be clearly stated that this is presently
guaranteed, as long as the correct methodology in design is followed and the quality control on
site is assured.
As shown in Table 3.4, the average masonry increase strengthening factors are known,
according to the general masonry type and the type of grout. Moreover, beyond core sample
removal, other non destructive methods enable the efficiency evaluation. The results of research
programs carried out worldwide during the last twenty years assure the mechanical and
chemical compatibility and durability of cementitious grouts within time, since a correct

Chapter 3 - Injection as a consolidation technique for masonry

51

designing of the grout's formulation is undertaken. However, the present lack of knowledge
about these main questions concerning polymeric grouts, definitely limits their use on historical
buildings.
From the technological viewpoint, grouting technology is presently mastered to avoid any kind
of damage to a historical building.
In terms of architectural and structural compatibility, the fact that a binding element is
introduced where in the past another cementitious or lime mortar existed, enables grouting to
preserve the authenticity of the building.
Finally, the fact that grouting is not a reversible technique, does not hinder its use on historical
buildings. However, it demands from the technicians responsible for a grouting operation, a
complete diagnosis and a careful evaluation of all the parameters involved, as the only way to
achieve a qualified restoration project. Moreover, a permanent supervision of the works is a
vital condition to ensure a demanding level of quality.

3.7. Conclusions of chapter 3

This technological study on grouting shows that the technical requirements are presently
understood. There are solutions to overcome the difficulties inherent to interventions on historic
buildings.
The fact that grouting involves working with masonry - a media far from being completely
understood - makes it difficult for the present to achieve an analytical modeling. It strongly
demands that a grouting design must be fully calibrated on site. Moreover, a continuous
supervision of the works is the only way to ensure an effective quality control. Concerning
cementitious grouts, they are compatible with the traditional masonry fabric, if the predefined
requirements are followed.
Lime based grouts are poorly studied in rheological terms and provide a weaker mechanical
improvement. In particular, this urges a further study over the effectiveness of adding a
plasticizer. Nevertheless, whenever a given repair operation of a double leaf masonry is
envisaged, they can be an alternative.
Cement grouts with superplasticizer added, are better studied and they offer a satisfying
injectability capacity, with a simultaneous not demanding mixing procedure. Moreover, the
provided mechanical improvement enables to already use them on strengthening operations.
Ultrafine cement grouts, obtained by grading correction, are undoubtedly the most effective .
However, their use is rather postponed for the moment, as high speed mixers to use on site are
not easily available, and mixing procedure is complex, even in laboratory conditions.
Concerning micro cement grouts, the state of the present knowledge is not yet adequate to enable
their use in masonry grouting. They are associated with a high W/C ratio, which leads to
unanswered questions. What mechanical improvement and shrinkage can be expected?
Polymeric grouts offer the highest effectiveness but their compatibility with masonry and the
behavior in long time is not completely known. These facts should strongly limit their use in
historical buildings.
52

Grout injection of masonry, scientific approach and modeling

The mechanical effectiveness of grouting is mainly a function of two parameters:

the injectability of the grout and the type of masonry. Two main types can be distinguished, due
to the significant different porosity levels associated: double leaf masonry with a rubble core
and plain stone or brick masonry. Moreover, sole grouting should not be used in buildings
whose first mode of collapse concerns the rotation of the walls out of their plane and a
strengthening operation to horizontal loads must be performed.
As the grouting principle is to reinforce the mechanical properties, it poorly strengthens the
building behavior to that first mode. In these cases, grouting must be carried out in parallel with
other strengthening techniques, such as tie-rods or ring beams.
Concerning the control of effectiveness, it presently demands destructive tests to be carried out,
although minimized by non destructive techniques. Coring is the most effective destructive
method, while either electrical resistivity or sonic measurements are effective non destructive
techniques. The fact that destructive tests must be carried out, strongly points to the need of
assuring a close interdisciplinary work. Grouting answers to the demanding conditions of
interventions in historic buildings. Presently, its effectiveness is guaranteed as well as the
durability of cementitious grouts. it is compatible in the physical, structural and architectural
viewpoints, preserving the authenticity of the building. The fact grouting is not reversible
implies that it should only be performed, whenever the means to ensure a complete and qualified
operation are present. Grouting is still unpopular and in particular cement based grouting is seen
as undesirable to intervene in historic buildings. On the basis of this common suspicious ideas,
there is lack of knowledge over the matter. It is hoped that this work helps to clarify technicians
in charge over the powerful resources of this technique. Thus, the porosity of the media and the
rheological behavior of the grouts are main parameters to be correlated with the mechanical
improvement.
The rheological behavior and the injectability are studied in detail by the LCPC in France. As it
is shown in Chapter 4, most of the problems that consolidation injections are meting are related
to a poor penetration of the grout in the masonry. This provides a non uniform filling of the
voids resulting in a poor or uncertain strengthening. This research program has focused on the
modeling of the flow of the grout inside the masonry. The experimental program helped to
understand most of the physical mechanisms taking place when the grout penetrates the masonry.
The model is built, using this information, to simulate the penetration of the grout. This helps to
judge which combination of process parameters, such as the injection pressure, the bore hole
diameter, the pattern and density of the bore holes on one hand and the rheological parameters of
the grout determined by the composition of the grout on the other hand can lead to a satisfying
and uniform filling of the internal voids in the masonry.
One could resume that the effectiveness of grouting of historic masonry implies a structural
improvement and an improved reliability. The strength increases because of the better internal
cohesion and the filling of the voids make the structure. However, the structural improvement
does not only depend on the strength. The ductility of ancient masonry might be endangered if
the grouts composition is chosen without competence. The creation of a solid stiff part in a hole

Chapter 3 - Injection as a consolidation technique for masonry

53

that has certain ductility may lead to additional damage during the following solicitation. The
more the injection leads to an even (re)distribution for stresses the more efficient it will be.

54

Grout injection of masonry, scientific approach and modeling

Chapter 4.
solutions

Problems faced during injection - possible

4.1. Wrong materials, chemical, physical and structural incompatibility

A consolidation injection is always an important operation. As clarified above it is a multi
disciplinary task to accomplish. During injection many problems can occur. Sometimes, these
problems reveal clearly during the injection itself, for instance when the injection blocks very
soon after it started. The grout does not penetrate inside the masonry. On other occasions, the
problems appear only many years later. A known example of the latter is efflorescence or salt
crystallization because of the wrong injection material. Similarly, the use of high strength grouts
can cause severe damage during the next mechanical solicitation (eg. Italy). This chapter treats
some known mistakes and tries to formulate some guidelines to avoid them.
The possible injection materials, all binding agent in liquid form, are generally discussed in
paragraph 3.4. They all have their peculiarities, their advantages and disadvantages. A
cautious selection of the binding agent is of major importance for a successful execution of the
injection. The injection material has to be selected on three criteria. In the order of importance
these criteria are: chemical and historical compatibility, rheological properties and mechanical
properties.
To my opinion the chemical compatibility is the most important requirement. If the injected
materials are not chemically compatible, sooner or later the injection will turn out to be a big
problem for the building.
We already mentioned the possible formation of ettringite when injecting cementitious grouts.
The CA3 in the cement reacts with the gypsum in ancient masonry. This is way blast furnace slag
cement is preferable to ordinary Portland cement. The alkali in cement might cause
efflorescence. This is why E. Toumbakari [1991(1)] recommends low alkali (marked LA on the
bags) cement for grout injection. For similar reasons the injection of cement grouts in
monuments that are erected using gypsum is to be a avoided.
However, chemical incompatibility is the most important form, other possible incompatibilities
must be considered. The structural and physical compatibility justifies a lot of attention.
Ancient masonry is astonishingly capable of taking settlements due to its very ductile behavior
(Figure 4.1). In part 3.4 it was indicated that cementitious grouts provide very good
strengthening, but also mean a loss of ductility to the structure. This loss of ductility
jeopardizes the self healing properties of masonry. Although the injected structure becomes
stronger and is able to carry higher load, it is possible that foundation settlements cause an
earlier damage than in case of a more ductile binding agent providing a less strong injected
building.

Chapter 4 - Problems faced during injection - possible solutions

55

Figure 4.1: A nice demonstration of the ductile behavior of masonry: the settlement of the
central part is supported without major cracking [Van Balen, 1991]

Injecting cement grout in an ancient building that was erected using lime mortar, introduces
locally stiff parts between relatively soft surrounding material. These hard parts can cause
splitting forces as was shown above on Figure 2.3. The soft mortar is more compressible than
the hard cement grout. In case of additional loading, bending occurs.
The above problem of introducing some hard material must be a concern, but one has to realize
that the bricks are also stiff material. Replacing the relatively soft mortar by stiff grout might
cause problems.
Physical compatibility is important for the temperature and moisture household of the building.
The physical compatibility becomes a bigger concern when polymers are used. Polymers have
a completely different physical nature than mineral binding agents. The thermal expansion is
different from the thermal dilatation of the mineral components in masonry. The facade of
buildings is subjected to rather severe temperature fluctuations. They cause thermal stresses
and by the dynamic nature of the fluctuation they can damage the consolidated material.
The moisture household can be changed dramatically by injecting polymers. Polymers are
impervious to water. This means that they form a barrier for rising damp and that in specific
situations, the water can not escape any longer from the masonry structure. The water will, by
necessity, find another way to evaporate. Through this new way it is harder to expel the same
amount of water. Therefore, the overall moisture content of the building will increase. This can
cause additional damage, for instance to decorations that are sensitive to humidity or can cause
a different salt migration problem. Similar problems can happen in case of mineral mortars.

56

Grout injection of masonry, scientific approach and modeling

Mostly, the moisture transport properties are less different than in case of polymers, but it has to
be mentioned that the pore structure of a cement mortar differs a lot from the pore structure of a
lime mortar or an hydraulic lime mortar. Figure 4.2 shows an example of a repointing, done
without considering the problem of moisture evacuation through the pointing layers. The water
is stopped right behind the repointing and the freezing water has soon pushed away the newly
placed mortar. Thermal incompatibility between laying mortar and pointing mortar can cause
the same damage phenomenon.

Figure 4.2: Loss of bond between repointing mortar and the brick due to freezing water behind
repointing. The poor evacuation of water is related to the different pore structure of
the new repointing mortar. [Van Balen, 1999]

4.2. Incomplete filling of the voids

It has already been mentioned that the reduction of the risk of failure depends highly on the
degree of homogeneity of the masonry after injection. At a uniform filling the variance on the
strength decreases and this way the reliability is improved. Similarly, the increase of the
average strength reduces the risk of failure. Indications exists that the uniform filling of the
masonry is more important than the mechanical properties of the injection grout. Tomazevic
performed an experimental program where four different cement based grouts were used
[Tomazevic, 1992(1)]. The quantities of cement were varied and in some grouts a hydrophobic
additive was added. Although the compressive strength of the different mixtures varied from 7
to 32 MPa, there was no significant difference in the final mechanical strength of the injected
masonry samples. This proves again that a complete filling and a good penetration are
essential. Filling all the voids inside the deteriorated masonry is not an easy challenge. Several
Chapter 4 - Problems faced during injection - possible solutions

57

parameters are involved. The distance between the bore holes, the injection pressure, the
rheological properties of the grout and the water absorbing properties of the masonry.
Furthermore, the overall condition of the masonry is important, the amount of cracks and their
width. The model, described later in this thesis, is able to estimate the influence of these
parameters.
4.2.1. Rheology
There exist different causes for incomplete filling of the voids inside the masonry. The most
obvious one is a lack of fluidity of the grout. Considering the final aim of this research program
of predicting the flow of the grout inside the masonry, some rheological considerations about
dispersions are more than justified. This topic is treated extensively in Chapter 6. The most
important findings are listed here. Many of the expected problems involve the nature of grout: it
is a dispersion of cement particles in water.
C
If the flow channels are too narrow, the particles will get stuck in these channels.
Hereby one has to realize that flocculated grains act as big grains. The concentration of
grains is relatively high and hence they hinder each other during motion.
C
The water absorption causes cement particles to stick to the wall of the flow channel.
Therefore it can be stated that the flow channels narrow by the suction of water out of the
grout and that it is not true that there is an overall increase of concentration of grains that
would cause a dramatic increase of the rheological parameters. The grout that keeps on
flowing however, has the same properties as the grout that is injected. Experiments
indicate no changes in the rheological properties between the grout that leaves the test
samples and the grout that was injected. Of course, the result will be the same: in both
hypotheses the grout will stop flowing. In the first assumption, the grout will stop
flowing because the flow channel becomes too small, whereas in the second assumption
the grout flow will halt because of differences in rheological properties. Both
phenomena bring the correction factor for Bingham fluids in the Buckingham formula (Eq
4.1) closer to zero.
Q '
Where

p D4 ? P
4 4t
L
1 4t L 4
[1 & ( c
)% ( c
)]
128 L
3 D ?P 3 D ?P
Q
?P
D

L
tc

(Eq 4.1)

= discharge [m3/s]
= pressure difference [Pa]
= diameter of the flow channel [m]
= dynamic viscosity of the grout [Pa.s]
= length of the flow channel [m]
= shear stress of the grout [Pa]

Figure 4.3 shows a slice of the injected cylinders, filled with crushed bricks. In 1, the
test configuration will be explained as well as how the test injections were done and
58

Grout injection of masonry, scientific approach and modeling

what the results were. As can be seen, there is a dark layer around each fragment of
crushed brick indicating a zone with a higher cement concentration. This is caused by
the water absorption out of the grout forcing the cement grains to stick to the walls of the
flow channels. The water absorption is beneficial for a good bond to the existing
masonry. It creates an interfacial layer with high cement content. Besides, since the
water is absorbed inside the capillary pores of the bricks, some very fine cement grains
also move into these pores.
When the grout penetrates further inside the masonry, the pressure gradient decreases.
At a certain moment, in case of a Bingham fluid, the critical shear stress will not be
reached any longer. Then the grout stops flowing. Firstly in the finer channels, but
finally also in big flow channels. This is a pure rheological phenomenon, the same
would happen in a plastic tube. This proves again that it is not correct to treat the grout
as a Newtonian fluid since a Newtonian fluid has no critical shear stress.

Figure 4.3: The dark layer around each fragment of crushed brick indicates a
zone with a higher cement concentration. This is caused by the water
absorption out of the grout forcing the cement grains to stick to the walls of
the flow channels.
4.2.2. Stability of grouts
Another phenomenon can give rise to zones that are insufficiently strengthened. Those zones
were filled during injection, but due to instability of the grout, the cement particles sink. This
causes a strength gradient in the injected zone: a stronger zone downwards, because of a higher
cement concentration and a weaker zone on top of this. This strength gradient was clearly
noticed when analyzing the strength results from the test injection that will be discussed in 5.4.4.

Chapter 4 - Problems faced during injection - possible solutions

59

The finding about strength, stiffness and density gradient are discussed there in detail. The
consequences can even be worse in case of a very unstable grout. The segregation of the cement
particles occurs in the collector and so a grout is injected with a much higher water content as
was meant. Or the sinking happens in the injected masonry. In the worst case the upper layer of
the injected zone contains no cement at all. This situation is shown in Figure 4.4. Similarly,
properly injected regions are emptied by invisible leakages. The grout flows away and leaves
the injected zone partly empty.

Figure 4.4: A very unstable grout, 30 % bleeding, was injected resulting in an upper zone with
hardly any cement left [Chandra, 1993][Van Rickstal 1999]

Stability is thus a very important property of the grout. Another argument to try to compose a
stable grout is the injectability. When flow slows down, the cement particles in an unstable
grout sink to the bottom of the flow channel. This narrows the channel and finally blocks further
injection.
Addition of stabilizing admixtures (bentonite, ultra fines) significantly improves the stability and
the injectability of the grout [Miltiadou, 1991][Paillre, 1986]. For this reason it is important to
check the stability of the grout and the evolution in time of the stability [Van Rickstal, 1995].
The newly developed test (see paragraph 5.2.5 for more information about the test method)
enables a detailed analysis of the stability of the grout. The data that are recorded are plotted
against time. As an example, Figure 4.5 shows how powerful the technique is to quantify the
impact of the dosage of a stabilizer on the stability. It has to be mentioned that for neither of the
dosages, except for the one without stabilizer, a visual difference could be observed. According
to the classical existing test method, which consist of simply pouring some grout in a measuring

60

Grout injection of masonry, scientific approach and modeling

jug and observing if any color difference is occurring between the toplayer and the rest of the
grout, they would all be classified being stable since the segregation is not visible.

Influence of % Bentonite on Stability

% of Initial Density [%]

100
98
96
94
92

0 % bentonite
1 % bentonite
1,5 % bentonite
2 % bentonite
3 % bentonite

90
88
86
84
82
0

200

400

600

800

1000

1200

1400

1600

1800

Time[sec]

Figure 4.5: The stability test is used to check the influence of bentonite as a
stabilizing admixture.

4.3. Solutions
4.3.1. Improving the injectability
Obviously, the composition has a major impact on the injectability of the grout. A research
project for the composition of a consolidation grout for the consolidation of the Basilica of Our
Lady in Tongeren was set up [Chandra, 1993][Van Gemert, 1989][Van Gemert, 1990]. It turned
out to be nearly impossible to formulate a well injectable grout without the use of
superplasticizer. All the grouts that are considered contain a certain dosage of superplasticizer.
The PhD study of A. Miltiadou [1990] deals with a great number of grout compositions. Most
important findings are related to the selection of the type of cement, the fineness of the cement,
the water content and the mixing procedure.
She studied the relation between the diameter of the bigger particles of the cement and the
dimension of the cracks to insure the injectability of the mixture. If necessary, the cement has to
be sieved to eliminate the coarse fraction. The study of the cement grading enabled to define the
absolute condition for a pure cement grout to be injectable in a sand column. The dimensions of
the sand grains in the sand column can be used to calculate the approximate dimensions of the
width of the cracks. Although grading demands are normally not possible in practice, the
granularity can be adapted by adding ultrafines. Ultrafine cement is another possibility. The
grading conditions are listed in Table 4.1.

Chapter 4 - Problems faced during injection - possible solutions

61

Table 4.1: Conditions of injectability for pure cement grouts [Miltiadou, 1990]
Dimensions of the
cracks [mm]

Granularity of the cement: percentage of refusal

160 m

80 m

64 m

32 m

0,1 - 0 2

#1

# 12

0,25 - 0,4

#7

#8

# 23

The water content has a large impact on the rheological behavior, as will be discussed in
Chapter 6. The rheological behavior of grouts has extensively been studied by O. Benhamou
[1994]. A higher water content improves the rheological features of a cementitious grout. The
viscosity and the yield stress decrease more than linear as can be seen on Figure 4.6. For other
kinds of cement, with other admixtures or other type of superplasticizer, the figures will differ
from this case, but the trend will remain similar. Intuitively one might find the solution to a
better injectability by increasing the W/C ratio. But doing so has two negative consequences:
firstly, the mechanical strength of a grout with a high water content is poor and secondly the high
water content has detrimental effects on the stability of the grout. A better option is to use
additional superplasticizer.

Rheology of grouts
0.16

120

0.14
Viscosity (literature)
Viscosity (own experiments)
Shear stress

Viscosity [Pa s]

0.12
0.1
0.08

80

60

0.06

40

Shear stress [Pa]

100

0.04
20

0.02
0

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1.1

W/C ratio

Figure 4.6: Rheological properties of grouts in function of the water content [Gil, 1995]

62

Grout injection of masonry, scientific approach and modeling

Ultrasonic mixing [Miltiadou, 1990][Toumbakari, 1999(1)] improves dispersion, especially

when ultra fines are added to the grout. It permits the use of a water content lower than in case
of high turbulence mixing, still providing the same penetrability of the grout. Penetrability
performance does not depend only on the maximum diameter of the particles, present in the
grout. It is known that fine materials in suspension coagulate very easily due to interparticle
interactions. The use of superplasticizer permits the development of repulsive forces due to the
adsorption of the ionic polymers on the surface of the grains. However, their action might not
be sufficient when the grout contains very fine materials such as silica fume and lime. Those
particles tend to coalesce in flakes of different size. The penetration capacity of such grouts is
significantly decreased. Thus, for a required penetrability performance, either the water content
of the grout must increase, with detrimental effects on the stability of the suspension and the
mechanical properties of the hardened grout, or the mixing procedure must be able to
deflocculate the particle clusters formed in the suspension.
It had been demonstrated earlier that an ultrasonic treatment can easily disperse fine substances
in water. More recently, the ultrasonic dispersion technique has been applied to the
development of cement grouts for the repair of masonry structures. In the study high turbulence
mixing was compared to ultrasound mixing. The conclusions of those studies [Toumbakari,
1999(1)] can be formulated as follows. The high turbulence mixing procedure was found to be
unable to ensure a constant penetrability of grouts composed of cement and fine materials. It is
not capable to deflocculate all the formed flakes. This does not necessarily hinder the injection,
because the heavy flakes settle rather soon after mixing in the recipients. However, obstruction
of pumping cannot be excluded and the water content of the injected grout will be higher than
planned. Furthermore, if the suspension contains silica fume, the high turbulence mixing
procedure is not able to produce an injectable grout unless the water content is increased or the
dosage of superplasticizer is increased. The ultrasonic mixing procedure on the contrary
permits to produce a high penetrability grout with a limited water content, even if silica fume is
used. This is due to the high dispersion capacity, which permits to deflocculate even very small
particle clusters.

4.3.2. Improving the stability

The stability can be corrected by changing the W/C ratio, by adding stabilizing agent such as
bentonite, by adding ultafine admixtures. The testing method that was developed, has been used
to measure the effect of the above parameters on the stability of the grout. The results of these
test are discussed in paragraph 5.2.5.
4.3.3. Injection holes
An intelligent layout of the injection holes can reduce the uncovered zone. Actually there are
three parameters involved about injection holes. The injection hole diameter, the injection hole

Chapter 4 - Problems faced during injection - possible solutions

63

depth and the pattern of injection holes. The depth of the injection hole should at least be half of
the wall thickness. If the cracks are narrow, or if the cracks are rather rare, the drilling hole
should reach at least two third of the walls thickness. The diameter of the hole should be
sufficiently large. Especially if the wall is thick, the discharge through the injection hole can
reach significant values. The smaller the diameter, the more important the pressure loss will be.
If the pressure loss is large, the injection time will increase and the grout will stick more to the
walls of the flow channels as will be demonstrated in paragraph 9.3. During injection, when the
grout reaches a large void, the time to fill that void will be much higher in case of a small
injection hole. During that time the grout will settle in the fine neighbourhood channels and
when the void is finally filled, the grout in the fine cracks lost too much water to restart flowing.
The zone, hidden by the fine cracks will not any longer be injectable through that particular
injection hole. The injection holes are the only flow channels that one has under control, so they
should be as perfect as possible.
In the hypothesis that the action radius is equal for all injection holes, the closest pattern (Figure
4.8) provides and injected area that is more than 90 % of the total area. The square pattern
covers only 78.5% of the global area. Making the pattern denser until no area is uncovered, one
gets some overlap. This gives the situation as drawn in Figure 4.9 and Figure 4.10. In case of
the closest pattern this overlap is 20.9 %, in case of the square pattern the overlap is 57.1 %.
From the figures and from the calculated sections of overlap it will be clear that the closest
pattern will be more economical than any other pattern. In other words the closest pattern
requires less drilling holes for a complete covering of the area than any other configuration.

Figure 4.7:Square pattern, 21.5 % not

covered

64

Figure 4.8 : Closest pattern, 9.3 % not

covered

Grout injection of masonry, scientific approach and modeling

Figure 4.9: Square pattern without

uncovered zone, 57.1 % overlap

Figure 4.10: Closest pattern without

uncovered zone, 20.9 % overlap

From theoretical point of view, the denser the injection hole pattern, the more homogeneously
the masonry will be injected, the lower the injection pressure can be kept and the better the final
result. But there are economical constraints that reduce the possible number of injection holes.
First of all, those holes need to be drilled. Secondly, a switch from one hole to another means
extra work, extra time needed to complete the job. Apart from economical reasons, also
technical reasons limit the density of the injection holes. If not sealed during the injection of a
hole, the neighboring holes may act as leakages that prevent the pressure to build up. In
particular cases a denser pattern can be used for zones that are problematic to inject. Locally
additional holes can be drilled.
4.3.4. Chemical and mechanical compatibility
The use of polymers for consolidation should be limited for those cases where the structural
challenges are such that cementitious grouts can not achieve them. In those cases polymers
could mean the answer to the problem. The discussion about chemical compatibility is limited
to cementitious grouts.
A justified method is trying to imitate the original mortar. Therefore, the original mortar is
analyzed with regard to the binding agents nature and eventually the ratio of binding agent on
inert material. When the analysis is done, the results can be used to see if it is possible to
compose a similar mixture that can provide the required strengthening. By using an imitation
mortar, the mechanical compatibility should not induce further problems. A very liquid form of
this mortar, consisting of pure binding agent, modified using an suitable superplasticizer, is then
suitable for injection purposes. Eventually, some cementitious material, that will not be present
in the original mortar, could replace the hydraulic portion of the binding agent. It will provide
the initial strength, needed for immediate consolidation and if used cautiously and in limited
dosage, it will provide a stronger material without compromising the ductility. Anyway,
because of the chemical compatibility blast furnace slag cement is preferred above OPC.

Chapter 4 - Problems faced during injection - possible solutions

65

Natural puzzolanic materials, such as Santorini Earth or Trass are alternatives. Experiments
were carried out by E. Toumbakari [1997], that showed the requested possibilities of these
blended materials.
Anyhow, the building should be checked on soluble salts. The formation of ettringite or any
other expansive crystallization caused by using chemically incompatible cementitious grout
should be prevented It is also worth looking at possible efflorescence, although efflorescence
will not endanger the stability of the building. Aesthetically however, it should be avoided.

4.4. Subject of research

Chapter 2 illustrates that some massive masonry structures require consolidation due to various
causes and deterioration mechanisms. In many cases a consolidation injection is able to
provide the wanted strengthening. From the probabilistic evaluation method [Schueremans
1997, 1997(2),1998], it is clear that both, the increase of the average strength and the decrease
of the variance of the strength, can improve the reliability of masonry structures. Paragraph
2.3.2 indicates how a uniform filling gives rise to a limited variance on the strength of the
masonry. Chapter 3 gives the state of knowledge about injection as a consolidation technique
for masonry. The technological possibilities are described, the correct way of analyzing a
masonry building with regard to consolidation is given. In paragraph 3.4 gives a closer look to
the possible grout compositions. Chapter 4 mentions problems that are regularly encountered
during consolidation injections. From this list of problems and from the description of the
benefit of an injection for the building, it is obvious that it would be useful to possess a model
that could simulate the penetration of the grout inside the masonry. Knowing if the grout is able
to uniformly fill the voids in the masonry, could mean a breakthrough in increasing the efficiency
of grouting. The above introduction and positioning of the injection technique is the result of a
literature study, the first pillar of this research program. The following chapters contains the
two other pillars that lead to the model used for simulating the penetration of grout in masonry.
Chapter 5 describes the experimental methods, newly developed for or adapted to the
peculiarities of grouting, lists the results of the experiments and draws conclusions for modeling
the grouting process.
The third pillar is a study of the relevant theories for the grouting process. In Chapter 6 the
rheology of dispersions is discussed. It is explained how the fact that a hydraulic grout is a
dispersion has an impact on the grouting process, more specifically the penetration speed and
penetration depth of the grout in the masonry. The grout flows mainly through the big cracks of
the masonry. For that reason attention is paid to the flow of dispersions through cylindrical
channels. The material that is surrounding the main cracks is porous masonry. Although, the
transport of the grout through the masonry is not treated using a continuum approach, some
aspects of the broad theory about flow of fluids through porous materials are selected and
discussed in Chapter 7. Darcys law, already mentioned in paragraph 5.4.3, is analyzed and the
mathematical finite element formulation is given. It is used for incorporating the water
absorption [paragraph 8.2.4] in the model.
66

Grout injection of masonry, scientific approach and modeling

The input from these three pillars - literature study, theoretical study and the experimantal
program - will lead to the discrete modeling of the penetration of the grout inside the masonry.
Chapter 8 explains the options that were taken in building the model. And finally the use of the
model is demonstrated for some relevant applications in paragraph 9.3.

Chapter 4 - Problems faced during injection - possible solutions

67

68

Grout injection of masonry, scientific approach and modeling

Chapter 5.

Experimental program

5.1. Aim of the tests

The experimental program splits up in three parts. One part is dedicated to appropriate testing
methods for the relevant properties of the grout. A second part discusses methods to investigate
the masonry structure, especially with regard to injectability. Of major impact is the size
distribution of cracks and voids. Current state of non destructive testing is hardly able to map
the major cracks. A method will be presented that results in a global permeability value for
each injection hole. The last and most important part treats a large amount of test injections. In
this last part masonry properties and grout properties are combined to come to an understanding
of the interaction of both components on the complex injection process.
As was explained before, a homogeneous filling of all gaps and voids is the most important goal
for achieving a successful consolidation of the masonry structure. The injectability of the grout
will therefore obtain all the attention that is needed in this chapter. When talking about
injectability, two kind of grout properties can be seen: the rheological properties in strict sense
and the additional properties of the grout related to the fact that a grout is a dispersion of cement
particles in water.
The experimental program points out the basic aspects of grouting, and forms the basis for
further mathematical modeling.

5.2. Testing the grouts properties

5.2.1. Mixing procedure
The characteristics of the materials used for the grouts composition are given in Annex 1 of this
book. Since mixing the composition to obtain a stable, injectable and well dispersed grout is a
complex procedure in which every step is important for the final performance of the grout, the
same mixing procedure is used throughout all the experiments.
The bentonite, Bentonile CV 15, supplied by Denys N.V., was mixed in advance with water,
taking the proportion of 1 weight unit of bentonite to 9 weight units of water. (This amount of
water is taken into account for the final W/C ratio.) After the swelling of the bentonite for more
than 24 hours, this mixture obtains the outlook of a putty. Then, the bentonite water mixture is
able to provide the required stability enhancing action. The mixing procedure contains the
following steps:
C
C
C
C

mixing at low speed the bentonite putty with half of the water
adding the cement (and the mineral admixtures such as silica fume)
mixing at low speed the cement, half of the water and the bentonite
2 minutes of waiting time

Chapter 5 - Experimental program

69

C
C
C

adding half of the superplasticizer and mixing at high speed during 2 minutes
2 minutes of waiting time
adding the other half of the superplasticizer and mixing at high speed during 2 minutes

5.2.2. Rheological properties of the grout

As extensively described in 1, a grout, being a dispersion of cement particles in water, has a
complex rheological behavior. It is justified to treat a grout as a fluid obeying the Bingham law
of flow. This means that there exists a critical yield stress t 0 . If the shear stress does not
exceed this critical value, no shearing will take place. Apart from this critical shear strength,
the relation between the shear stress and the shear rate is assumed to be linear as expressed by
the Bingham formula:
t ' t c % ?0
(Eq 5.1)
where
tc
?0

dynamic viscosity [Pa s]

critical shear strength [Pa]
shear rate

Relatively simple testing methods exist to determine both rheological values:

Viscosity was measured using the Brookfield viscometer. The reading on the Brookfield
viscometer is actually the torque, necessary to realize a certain rotation speed. Depending on
the rotation speed, the torque needs to be multiplied by a factor providing the dynamic viscosity
in cP (1 centiPoise = 0.001 Pa s). The viscosity at different shear rates was estimated. By
extrapolation to shear rate being zero, the critical shear strength was computed. This way of
working can be criticized, since the behavior of a dispersion when approaching a shear rate of
zero seems to be very uncertain [Toorman, 1992]. One might find an sudden increase or
decrease of the relation shear rate vs shear stress. Generally, the obtained results corresponded
fairly well with values found in literature. Figure 5.1 displays the possible error made by
extrapolation of the value near to zero. This way the critical shear stress could be
underestimated.

70

Grout injection of masonry, scientific approach and modeling

Figure 5.1:

Possible mistakes made by linear extrapolation of the measurements

[Toorman, 1995]

5.2.3. Dynamic viscosity

The dynamic viscosity measurements were performed on four kinds of grouts. These four kinds
of grout cover the complete gamut of mineral grouts that can be used for injection: the cement
grout using normal blast furnace slag cement (CEM III/A 42.5, LA from Obourg), hydraulic lime
grout Lime B-fluid (a very fine hydraulic lime meant for repairing the adhesion of frescos to the
supporting wall), micro cement grout (using spinor, an ultra fine cement that was commercially
available) and fine grout using normal blast furnace slag cement with an adapted grading by
adding condensed silica fume. The results are listed and discussed in this paragraph.
1 Influence of water content
Grout 1
binding agent
CEM III/A 42.5, LA
superplasticizer
1.5 % of Rheobuild 716, sulfonated naphthalene formaldehyde
stabilizer
2 % of bentonite, Bentonile CV 15

Chapter 5 - Experimental program

71

Table 5.1: Viscosity measurements of grout 1

W/C factor
40
0.50
0.60
0.67
0.7
0.8
0.90
1.00

Grout 2
binding agent
superplasticizer
stabilizer

Rotation speed
(opm)
3
30
60
30
60
30
60
30
60
30
60
30
60
30
60

Viscosity
(cP)
520
93
82
78
72
42
39
40
28
30
19
27
20
25
20

Hydraulic lime, Lime B-fluid 0, from Unilit

none
none
Table 5.2: Viscosity measurements of grout 2
W/L factor
1.2

1.5

Grout 3
binding agent
superplasticizer
stabilizer

72

Rotation speed
(opm)
12
30
60
12
30
60

Viscosity
(cP)
56.25
35.5
27.5
16.25
9
6.875

Microcement,
1.5 % of Rheobuild 716, sulfonated naphthalene formaldehyde
2 % of bentonite, Bentonile CV 15

Grout injection of masonry, scientific approach and modeling

Table 5.3: Viscosity measurements of grout 3

W/C factor
0.67
1.00
1.20

Grout 4
binding agent
superplasticizer
stabilizer

Rotation speed
Viscosity
(opm)
(cP)
to viscous to measure
30
22.5
60
15
30
5.25
60
4.5

80 % CEM III/A 42.5, LA + 20 % condensed silica fume

1.5 % of Rheobuild 716, sulfonated naphthalene formaldehyde
Silica fume has stabilizing action
Table 5.4: Viscosity measurements of grout 4
W/C factor
0.75

0.85

Rotation speed
(opm)
12
30
60
12
30
60

Viscosity
(cP)
137.5
62
37
91.25
45
26

Table 5.4 presents the viscosity in function of the W/C ratio for the first grout. The viscosity
decreases remarkably when increasing the W/C ratio from 0.47 to 0.7. A further increase of the
water content of the grout has no significant influence on the viscosity. Besides, the stability of
the grout gets worse. This is why a W/C ratio of 0.67 is chosen for the basic composition of the
grout. This W/C is used to check the influence of the other parameters of the composition.
Einstein and Brinkman propose a numerical relation between de volumetric concentration of
particles and the viscosity. For Einstein, who assumes that the particles doe not interfere, the
increase of the viscosity related to the viscosity of the pure liquid is linear, Brinkman has an
exponentially increasing relation. Both relations are represented in Figure 5.2. Our results tend
to confirm the exponential relation. Besides, the above theoretical relations suppose no
interaction of the particles. In case of cement grouts this condition is not fulfilled. Cement
particles interact by their electro statical charge. Experimental results prove that the viscosity
is increasing faster than the theoretical relations predict.

Chapter 5 - Experimental program

73

Table 5.5: Theoretical and experimental values for the viscosity in function of the W/C ratio
with a dosage of 1,5 % of superplasticizer
W/C Concentration
Viscosity [cP]
of solids
Theoretical relations
Experimental data
Einstein Brinkman Ish-Shalom Vom Berg Van Rickstal
0,4
0,442
21
96,8
39
300
520
0,5
0,388
19,7
63,9
16
120
82
0,6
0,346
18,6
48,2
12
45
72
0,7
0,312
17,8
39,2
8,5
20
28
0,8
0,284
17,1
33,4
5,8
9
20

Table 5.5 compares the theoretical values for Einstein and Brinkman with experimental findings
of Ish-Shalom, Vom Berg and our own experimental results. Generally, the viscosity is
decreasing more than linearly if the dosage of superplasticizer increases. The injectability will
depend highly on the rheological properties. Therefore, it can be stated that superplasticizers
are highly efficient in improving the injectability of the grout. Besides, Petrie has stated that for
a cement suspension with a high content of superplasticizer, the viscosity in function of the
concentration obeys an exponential law of type Brinkman. Since the general theoritical values
are much lower than the experimental ones, these values are referencing to the second axis.

Theoretical and experimental relations

between W/C and Viscosity
600

120

Ish-Shalo
Vom Berg
Van Rickstal

400

100

80

Einstein (2nd axis)

Brinkman (2nd axis)

300

60

200

40

100

20

Viscosity [mPa s]

Viscosity [mPa s]

500

0
0.3

0.4

0.5

0.6

0.7

0.8

0.9

W/C ratio

Figure 5.2: Two theoretical relations (Einstein and Brinkman) and three experimental
relations between viscosity and W/C ratio

2 Influence of stabilizer

74

Grout injection of masonry, scientific approach and modeling

It is obvious that the stabilizer will affect the viscosity and critical shear strength. The
stabilizer thickens the liquid to keep the particles in suspension. One has to compromise
between stability and viscosity. Bentonite CV 150 has been used as a stabilizer.
The composition for testing the influence of the dosage of bentonite on the viscosity is the
following:
binding agent
superplasticizer
water content
stabilizer

CEM III/A 42.5, LA

1.5 % of Rheobuild 716, sulfonated naphthalene formaldehyde
W/C = 0.67
parameter, Bentonile CV 15 supplied by Denys N.V.
Table 5.6: Viscosity measurements, influence of stabiliser
bentonite
% of cement mass
0
1
1,5
2
3

rotation speed
60
60
60
60
60

viscosity
cP
17,5
23,2
28,8
43,3
135,5

Viscosity measurements
influence of stabilizer

viscosity [cP]

160
140
120
100
80
60
40
20
0
0

0.5

1.5

2.5

3.5

bentonite [%]

Figure 5.3: Influence of dosage of stabilizer on the viscosity of cement

grout
3 Influence of ultra fines

Chapter 5 - Experimental program

75

A. Miltiadou [1990] states that the addition of ultra fines improves the injectability of a grout.
The assumption is that this improvement is due to a better stability of the grout and water
retaining properties. On the other hand, the extreme fineness of these admixtures increases the
water demand. Since ultra fines have stabilizing action, no additional bentonite is added to the
grout. The composition for testing the influence of the dosage of ultra fines on the viscosity is
the following:
binding agent
superplasticizer
water content
ultra fines

CEM III/A 42.5, LA

1.5 % of Rheobuild 716, sulfonated naphthalene formaldehyde
W/C = 0.67
parameter, dosage is percentage of cement + ultra fines
Table 5.7: Viscosity measurement, influence of ultra fines
ultra fines
% of Cem + SF
0
5
7.5
10
12.5

rotation speed

viscosity
cP
17.5
39.2
58.3
151.7
223.3

60
60
60
60
60

Viscosity measurements
influence of ultra fines
250

viscosity [cP]

200

150

100

50

0
0

10

12

14

silica fume [%]

Figure 5.4: Influence of dosage of silica fume on the viscosity of cement grout
5.2.4. Thixotropy, non linear behavior and time dependent properties

76

Grout injection of masonry, scientific approach and modeling

Most compositions were tested for two different values of the rotation speed: 30 rpm and 60
rpm. The results reveal that viscosity at lower speed is slightly higher than viscosity at higher
speed. This shows that grouts have a pseudo plastic behavior. A pseudo plastic fluid shows
lower viscosity values for higher shear rate. This phenomenon enhances the blocking process
when an injection slows down. This might explain the practical findings that a constant
injection pressure is preferable to a discontinuous injection with a periodical increase and
decrease of pressure.
The addition of bentonite provides some thixotropical aspects to the grout. Thixotropy means
that the viscosity decreases at sustained shear deformation, which occurs when the grout is
mixed during a longer period of time. Also, the initial shear strength increases after a moment
of rest. These aspects are not visible from the above measurements, since the grout was
thoroughly mixed after each period of stand still. Thixotropy increase the problems to restart an
injection that was halted.
5.2.5. Stability
Stability is a first requirement for a grout to be injectable. When the flow slows down, the
cement particles in an unstable grout will sink to the bottom of the flow channel. This narrows
the channel and finally blocks further injection. Stability also means that the grout is able to
retain the water. When the grout passes through the relatively dry masonry, some water will be
absorbed from the grout. The more water is absorbed the less fluid the grout will become and
hence injection will slow down and finally stop. These two blocking mechanisms show the
importance of the stability of the grout.

Figure 5.5: The classical stability test provides only little information
Stability of a grout used to be checked by putting some grout in a measuring glass (Figure 5.5).
After a period of time, it was visually checked if any segregation or bleeding appeared. This

Chapter 5 - Experimental program

77

way only very poor information is obtained about the evolution in time of the stability of the
grout. An other disadvantage is that segregation must be visible before it can be recognized. The
classical tests are thus not very suitable for detailed analysis of the stability. For those reasons
a more powerful test was developed.
Instability of the grout means that the heavy particles of the dispersion sink to the bottom of the
recipient due to gravity. This means that the density of the grout in the top region decreases
because of this loss of heavy particles. The more unstable the grout, the more the density will
decrease. The developed experiment makes it possible to measure the evolution of the density
at any height in the recipient.
An object, hanging in a liquid undergoes an buoyance force according to Archimedes law. This
force can be expressed by the following formula:
F = ? . g .V

?
g
V

= density of the fluid

= gravity acceleration
= volume of the object

When the fluid around the object becomes less heavy due to the settlement of the cement
particles, the density will decrease and hence the force decreases in the same proportion. By
recording the upward force one gets information about the evolution in time of the density and
the stability of the grout. This buoyance force is recorded using the setup in Figure 5.6 and
Figure 5.7. The balance is indicating the upward force, which is recorded every 5 seconds by a
computer. This method is analogue to the Andreasen Method in soil mechanics, used to measure
the content of very fine parts in a soil sample. The balance is put to zero when the recipient
with the grout is put on the balance. Once the object hangs in the grout the reading of the balance
gives the mass corresponding to the buyance force. Since the density ? varies when the
particles sink due to instability, the recording of the buoyance force implies the recording of the
stability.

78

Grout injection of masonry, scientific approach and modeling

Figure 5.6: Principle of the newly developed stability test for grouts

Figure 5.7: Configuration for the stability test

Chapter 5 - Experimental program

79

1 Influence of water content

The composition used to test the influence of the water content on the stability has a normal blast
furnace slag cement. Blast furnace slag cement is preferable to Portland cement because of the
lower CA3 content. This reduces the risk for the formation of ettringite. The low alkali content
(LA) reduces the risk for efflorescence [Toumbakari, 1997]. The dosage of 1.5% of Rheobuild
is chosen because it provides a relatively low viscosity. Two percent of bentonite turned out to
provide good viscosity values.
binding agent
superplasticizer
stabilizer
Water content

CEM III/A 42.5, LA

1.5 % of Rheobuild 716, sulfonated naphthalene formaldehyde
2 % of bentonite, Bentonile CV 15
parameter

Stability
influence of W/C ratio

% of initial density [%]

100
99
98
W/C = 0,5

97

W/C = 0,6
W/C = 0,7
W/C = 0,9

96

W/C = 0,67
W/C = 0,8

95
94
0

200

400

600
Time [sec]

800

1000

1200

Figure 5.8: Stability of cement grout in function of W/C ratio

As can be seen from Figure 5.8, the stability of all grouts is good. Bentonite performs very
well as a stabilizing admixture. On first sight the behavior of W/C = 0.8 grout is surprising.
This composition was tested a few months later then the other series. The cement was partly
hydrated through moisture uptake from the air. This way the normal mixing procedure could not
deflocculate the cement clusters as intense as before. Those heavy clusters sink quickly to the
bottom of the recipient, resulting in a decreased stability. This result proves how delicate the

80

Grout injection of masonry, scientific approach and modeling

formation of a well injectable, stable grout is. The cement has to be of good quality. The
smallest lack of deflocculation causes a complete change in stability.
2 Influence of stabilizer
The composition for testing the influence of stabilizer on the stability is the following:
binding agent
superplasticizer
stabilizer
Water content

CEM III/A 42.5, LA

1.5 % of Rheobuild 716, sulfonated naphthalene formaldehyde
parameter, % of cement mass, Bentonile CV 15
W/C = 0.67

One would expect the grout to become more stable when the dosage of bentonite is increased.
This seems to be the case until 2 % of bentonite. Adding more stabilizer does not improve the
stability behavior any longer. As can be seen from Figure 5.9, some stabilizer is absolutely
necessary to provide a grout that is sufficiently stable. A grout without stabilizer loses 15% (!)
of its initial density after half an hour.

Stability
influence of stabilizer
100
% of Initial Density [%]

98
96
94
92

0 % bentonite

90

1 % bentonite
1,5 % bentonite

88
86

2 % bentonite
3 % bentonite

84
82
0

200

400

600

800

1000

1200

1400

Time [sec]

Figure 5.9: Stability of cement grout as a function of stabilizer addition

Chapter 5 - Experimental program

81

3 Influence of ultra fines (silica fume)

The composition for testing the influence of ultra fines on the stability is the following:
binding agent
superplasticizer
Water content
ultra fines

CEM III/A 42.5, LA

1.5 % of Rheobuild 716, sulfonated naphthalene formaldehyde
W/C = 0.67
parameter, % of the sum of cement and ultra fines

Stability
influence of ultra fines
100
% of Initial Density [%]

98
96
94
92

5 % SF
7,5 % SF
10 % SF
12,5 % SF
0 % SF, bent
0 % SF

90
88
86
84
82
0

200

400

600

800

1000

1200

1400

Time [sec]

Figure 5.10: Stability of cement grout in function of dosage of ultra fines

Again we notice that a grout without any ultra fines is not stable. The stabilizing features of
silica fume are obvious from Figure 5.10. After all, bentonite is still a better stabilizer then
silica fume. As we will see later in this text, silica fume corrects the granularity of the cement
in the finer region and therefore, it improves the penetration of the grout in zones with fine
cracks. Another remarkable fact appears when too much ultra fines are added. The stability of
a grout with 12.5 % of silica fume is less stable then a grout without stabilizing agent. This is
similar to what was noticed when testing the influence of the dosage of bentonite on the stability.
A possible explanation is the presence of many very fine particles. The distance between those
particles inevitably becomes very small. Therefore, flocculation might happen although a
deflocculating superplasticizer is used.
The different efficiency of both stabilizing admixtures can be explained by their different shape.
Silica fume consists of almost perfect spheres whereas bentonite has a lamellar shape. Besides,

82

Grout injection of masonry, scientific approach and modeling

bentonite has electrostatic charges providing a better stabilizing action than silica fume [Van
Den Berghe, 1994].

Figure 5.11: SEM picture of bentonite

Figure 5.12: SEM picture of condensed silica

fume

5.2.6. Flow time measurements

To determine the rheological parameters of a fluid, one disposes of different families of
viscometer [Blom, 1988]. For the determination of the viscosity we used the Brookfield
viscometer from the family of coaxial viscometers. Other type of instruments exist, all have
their own advantages and disadvantages. Regarding our research field, it would be logical to
consider the family of capillary viscometers Figure 5.13. The instrument is basically a
recipient for a calibrated volume of liquid equipped with an opening to let the liquid flow out of
the recipient.

Figure 5.13:

Comparison between a flow cup (on the left) and a capillary viscometer (on
the right)

Chapter 5 - Experimental program

83

An ideal capillary viscometer has a long capillary cylindrical conduit with a small diameter.
This way the discharge is mainly determined by the flow resistance of this conduit through
which a laminar flow takes place. The available viscometers do not have this long conduit to
let the liquid flow out of the recipient. A capillary viscometer has a ratio length over diameter
of at least 15, whereas the flow cups have a ratio L/D of only 1,5. The Afnor cup, viscometer
according to the Franch code NF T 30-014, has only an opening. The Marchall cone (ASTM D
1084) has a small cylindrical outlet, but far too short to force the flow to be laminar or to limit
the discharge according to known rheological laws such as Haegen-Poisseuills law or
Buckingham formula [Hinch, 1975]. The charge losses due to the the transition from the
recipient to the conduit master the flow completely. As a consequence of this it is not possible
to calculate the rheological parameters out of the flow time measurements and to use these
values in the model. For that purpose, a capillary viscometer would be perfect. Though the
measurements enable to judge the fluidity of the grouts. Indeed, a comparison between all
measurements of flow time and Brookfield viscosity for the basic grout with varying W/C ratio,
shows that there is only a poor correlation between both (Figure 5.17).

Correlation Viscosity - Flow time

80
W/C ratio

75

R2 = 0,9407
70

Flow time [sec]

65
2

R = 0,9893

60

Stabilizer

R2 = 0,1804

55
50

Ultrafines

45

R = 0,8906

40
35
30
0

50

100

150

200

250

Viscosity [cP]

Figure 5.14:

The correlation between the flow time and the viscosity depends on the
investigated parameter

The overall correlation is only 0.18 whereas the correlation for varying W/C ratio, bentonite
dosage and ultra fine dosage is 0,97; 0,98 and 0,89 respectively. When only one parameter of

84

Grout injection of masonry, scientific approach and modeling

the basic grout composition ( CEM III A, 42.5 LA, 1.5 % of SP, 2% of bentonite, W/C=0.67) is
changed, the correlation is good as can be seen from Figure 5.15, Figure 5.16 and Figure 5.17.
One could state that the flow time measurements do not allow to derive the exact value for the
rheological parameters. They only give an indication about the fluidity. For on site control
measurements they are very suitable: a chronometer and the recipient are the only devices that
are required for the measurement.
Flow time measurements
influence of water content
250

80

Flow time
70

150

60

100

50

50

40

Flow time [s]

Viscosity [cP]

Viscosity
200

30
0.5

0.6

0.7

0.8

0.9

W/C ratio

Figure 5.15: Influence of W/C ratio on the flow time and the viscosity
Flow time measurements
influence of stabilizer
250

80

Viscosity [cP]

Flow time

70

150

60

100

50

50

40

Flow time [s]

Viscosity

200

30
0

0.5

1.5

2.5

dosage [%]

Figure 5.16: Influence of stabilizer on the flow time and the viscosity

Chapter 5 - Experimental program

85

Flow time measurements

influence of ultrafines
250

80

Viscosity [cP]

Flow time

70

150

60

100

50

50

40

Flow time [s]

Viscosity

200

30
0

10

12

14

dosage [%]

Figure 5.17: Influence of ultra fines on the flow time and the viscosity
It is remarkable that the influence of the water content on the flow time is proportionally as big
as the influence on the viscosity, whereas the influence of the dosage of bentonite on the flow
time is only half of the influence of the stabilizer on the viscosity. For the influence of ultra
fines, this phenomenon is even worse: the influence on the flow time is almost none, the
influence on the viscosity, on the contrary, is extremely big. This proves that different physical
effects are playing an important role in both types of measurements. It is remarkable that the
flow time of a grout with ultra fines remains low although the viscosity increases. This could be
an explanation for the good injectability of a grout with ultra fine admixtures. The physical
phenomenon that takes place inside the flow channels of deteriorated masonry is closer to what
happens during the flow time measurement than it is to what happens during the coaxial
viscometer test. Apparently the addition of fine admixtures has a completely different effect on
the viscosity than on the flow time.

5.2.7. General observations about testing the grout properties

The first observation about testing the grout properties is a rather critical one: standard methods
generally used in practice to find out about rheological properties of fluids are not very suitable.
In case of the stability measurement, the existing method is rather simplistic and not very
efficient providing only very limited information. A completely different method was
developed providing the evolution in time of the density and hence the stability. The method
enables to judge admixtures with regard to their stabilizing capacities, the efficiency of mixing
procedures etc... For the measurements of the rheological properties, the standard equipment is
in fact not very suitable for dispersions, nor for the viscosity nor for the critical shear strength of
the dispersion. The flow time measurements enable to judge the fluidity of the grout, but as was

86

Grout injection of masonry, scientific approach and modeling

proved by the experiments, the relation with viscosity is poor. Therefore it is recommended to
adapt the existing flow cups. Using a longer capillary conduit would probably solve the
problem about the correlation flow time to viscosity. Further research in this is beyond the
scope of this thesis.
Second observation is the significant increase of flow time and viscosity when the water content
decreases. This observation is a confirmation of theoretical and experimental findings in
literature. This fact makes the water retaining capacities of a grout composition very important
for deep penetration.

5.3. Testing the masonrys properties

5.3.1. Diagnosis of the masonry
The penetration of the grout inside the masonry structure is determined by the interaction of
grout and masonry. Tests to check the quality of the grout were described above. Although some
problems were mentioned and some adaptation of existing testing equipment seems to be
necessary to obtain reliable results, grout properties can be analyzed in the laboratory.
Different compositions can be compared. The test results can be used to quantify the parameters
in the model that will be explained later.
The internal structure of the masonry, the permeability, the crack size distribution and the
moisture content are the most important properties of the masonry with regard to injectability.
The overall condition of the masonry used to be determined by coring and analyzing the cores
with regard to mechanical properties, cracks, condition of the bricks and the mortar. These
results are often used to judge the possibilities of consolidation. Less destructive is the
combination of a non-destructive testing method, i.e. (ultra) sonic measurements or electric
resistivity measurements, with coring. This combination enables to quantify the qualitative data
obtained from the non-destructive testing. These data help to make the diagnosis: does the
building need consolidation and can grouting be a part of this consolidation? For the grouting
two questions are important: first, what strength improvement is required and what grout can
provide it; and secondly how will the grout progress in the masonry, to what extend will the
voids be filled?
The required strength improvement is relatively easy to judge. Fact is that filling the cracks and
the voids will cause the masonry to behave monolithically. Research indicated [Tomazevic,
1992] that the intrinsic mechanical properties of the grout do not or hardly influence the final
strength of a deteriorated masonry wall in case of comparable rheological properties of the
grout. This actually reduces the first question to the second one, provided the mechanical
properties of the grout are not too low. The flow of the grout through the masonry will depend
on the permeability of the masonry and even more on the crack size distribution. For the model
that was developed, a map of cracks would be the ideal information to make correct
simulations. The existing non destructive techniques however, are not able to provide that

Chapter 5 - Experimental program

87

information. None of the techniques has a resolution that is accurate enough to find out about
individual pores and cracks, even to give an idea about the crack size distribution.
If one wants to simulate the transport of grout by a continuum description of the masonry,
average permeability values are the required information.
5.3.2. Reproducible masonry samples: physical model
First problem is to compose a reproducible masonry sample suitable for test injections. Real
masonry structures have two disadvantages. Firstly they can not be remade or copied. Every
test zone, after injection is lost for further tests with other grouts or other parameters.
Secondly, it is hard to control or visualize what is happening inside the masonry structure. It is
not possible to see if the grout has penetrated the complete injection zone. Eventually cores can
reveal some of the desired information.
To improve the knowledge about the physical mechanisms that take place during injection and
that determine the penetration of the grout in the masonry, plexiglass cylinders, diameter 100 mm
and height 450 mm, are filled with a fraction of crushed bricks and are injected unidirectionally
from bottom to top. At the bottom, a layer of fine stones ensures a good distribution of the grout.
Crushed bricks are preferred over sand grains since crushed bricks have, just as masonry, a
water absorbing action. The crushed bricks were bought from a company constructing tennis
courts. Then they were sieved to obtain different grain size distributions to enable the
simulation of different permeability for the masonry Figure 5.18. The cylinders are made of
transparent plexiglass to make observation and thorough analysis possible.

Figure 5.18: The reproducible masonry sample: crusher bricks inside a plexiglass tube
5.3.3. Permeability of samples
Permeability of the masonry determines the practical injection process. The injection hole
pattern, the necessary fluidity and hence the composition of the grout depend upon the

88

Grout injection of masonry, scientific approach and modeling

permeability of the masonry. The more permeable the masonry the coarser the injection hole
pattern may be that guarantees a successful injection.
In constructing a model to describe the injection process, the permeability will be one of the
main parameters of the masonry that influence the evolution and distribution of the grout in the
masonry mass. Once again, before being able to determine the influence of the permeability
parameter, it must be possible to determine it as exactly as possible. The test setup to obtain the
permeability of the reproducible masonry samples will be explained first. Afterwards it is
indicated how this test could be adapted to obtain permeability values on real walls.
The samples used to perform the test injections are plexiglass cylinders, filled with different
fractions of crushed bricks. These fractions are densified by vibrating the cylinder after filling
one third and two third of it. That way a reproducible compaction is realized.
The setup on Figure 5.19 is used to determine the permeability of these test samples. Different
discharges are established and the pressure difference over the tube is measured. The formula
(Eq 5.3), based upon Darcys law for laminar air flow through porous media, provides the
permeability of the tested tube.

Figure 5.19: Setup for permeability measurements using air flow.

Re '

Where

Re
v
d

=
=
=

v d
? q d
'
' 8,46
?
Af

(Eq 5.2)

Reynolds number
velocity
diameter of channel

Chapter 5 - Experimental program

89

A
q
d
f

=
=
=
=
=
=

kinematic viscosity [m2/s]

dynamic viscosity [Pa.s]
cross section of sample
air discharge
average grain size
porosity

Darcys law can be used here because the Reynolds number for the most turbulent configuration
equals only 8.46, which is in the beginning of the transition region from laminar to turbulent
flow (1 < Re < 10) [Collins, 1965] and hence air flow through the crushed bricks can be
considered being laminar.

discharge q

area A

porous medium

Figure 20: Geometry and pressures for the calculation of the permeability

90

Grout injection of masonry, scientific approach and modeling

q '

Where

=
=
=
=
=
=
=

K
A
L
?p

K.A
( ? p)
.L

(Eq 5.3)

discharge [m3/s]
Dynamic viscosity of air
17,1 x 10-6 Pa.s
Permeability [m2]
Area of the tube [m2]
Length of the tube
pb - pa

For each discharge (10, 20, ..., 80 l/h) the pressure difference over the plexiglass cylinder is
measured taking into account the pressure loss over the tubes. In this way 8 data points are
obtained. Linear regression is applied and from this curve the K-value is calculated. Figure
5.21 shows the relation between discharge and pressure difference over the cylinder. The
charge losses in the conduits are taken into account by doing a measurement with an empty
plexiglass cylinder. The permeability of the samples is considered to be constant. However, in
precise terms, as the bottom layer of crushed bricks is more vibrated than the top layer, the
permeability at the bottom is probably slightly lower than at the top, for the same brick fraction.
In the contact zone of a finer fraction with a coarser fraction, a locally lower permeability will
be present because of the finer particles entering and filling the voids of the coarser zone.

Permeability using air flow, fraction 2-4 mm

4,5

Pressure [Pa]

4
3,5
3
2,5
2
measurements
regression

1,5
1
0,5
0
0

10

20

30

40

50

60

70

80

Discharge [l/h]

Figure 5.21: Permeability measurements for the fraction 2-4 mm, using air flow

Chapter 5 - Experimental program

91

This test provided the permeability values of the different fractions that were used to determine
the influence of the permeability on the injection process and on the evolution of the grout inside
the masonry. Table 5.8 lists the obtained values.

Table 5.8: Listing of the permeability values in several units

Fraction

Permeability K

K? g/

[mm]

[ m2 ]

[1 darcy = 10-12 m2 ]

[m/s]

1-2

7.28 E-10

0.728 E+03

7.11 E-03

2-4

68.7 E-10

6.87 E+03

67.1 E-03

4-10

150 E-10

15.0 E+03

147 E-03

1-4

61.9 E-10

6.19 E+03

60.5 E-03

2-4 / 1-2 / 2-4

29.5 E-10

2.95 E+03

28.8 E-03

5.3.4. Permeability of masonry structure

The above setup enables us to determine the permeability of the test samples in the laboratory. It
would be interesting to adapt the setup for in situ measurements of the permeability of the
masonry to inject. Of course, the laboratory situation is completely different from reality.
Nevertheless it is possible to use the above principle to blow air into the masonry and to
measure the pressure needed to realize a certain discharge to obtain indications about the
permeability.
Figure 5.22 shows a possible setup for this purpose. Through a drilled hole, a certain discharge
of air is blown into the masonry. For each discharge the air pressure is measured just before
entering the masonry. Inside a homogeneous masonry a radial flow will be generated. Using
Darcys law in differential form and in cylindrical coordinates it is possible to get a good
estimation of the permeability by integration between the hole radius and infinity [Collins,
1965][Dullien, 1979].

92

Grout injection of masonry, scientific approach and modeling

Figure 5.22:Possible setup for in situ permeability measurement

Precaution is needed due to possible leakage of air trough big holes. The masonry will never be
homogeneous, so it is possible that the main flow will occur through one big hole instead of
through the cracks and voids of the masonry and this way the measured permeability would be
higher than the average permeability of the masonry.

5.4. Laboratory injection tests

5.4.1. Description of the tests
The reproducible masonry samples that are characterized in paragraph 5.3.2, are used to
improve the knowledge about the physical phenomena that take place during the penetration of
the grout. During the first series of test injections, the progress of the grout in the plexiglass
sample is recorded on video tape. During the second series, the indication of the balance on
which the reservoir was put is also taped on video. This provides data about the discharge that
could be realized. Both recordings enable to analyze the evolution of the injection in time. The
first recordings provide the progress of the grout whereas the second series enables to quantify
the amount of injected grout in function of time. In addition to this analysis, a thorough
observation is possible. The complete injection of a sample of 45 cm heigh, takes only 30
seconds to one minute, so the time to observe in detail is too short. The recorded images can be
looked at more exhaustively.
At first the layout of the samples was straightforward: the cylinders were filled with only one
fraction of crushed brick. In a later stage, the samples layout was more complex. Several
fractions of the crushed bricks were combined to study what happens when a high permeable
fraction is in parallel with a low permeable fraction. The first series of injections were stopped

Chapter 5 - Experimental program

93

when the grout comes out of the sample. No further information can be obtained. The second
series however, are continued to see how the discharge changes with time. Most of the injected
samples are emptied right after recording the injection to save the plexiglass tubes. Only in
some particular cases, the mechanical strength of the hardened sample is tested. The sound
velocity is checked in order to calculate the dynamic Youngs modulus of the hardened sample.
The injection pressure for all these experiments was 1 bar = 100.000 Pa. If not mentioned
differently, the injections are done using the grout mix that came out best from the rheological
study, with proportions:
binding agent
superplasticizer
stabilizer
water content

CEM III/A 42.5, LA

1.5 % of Rheobuild 716, sulfonated naphthalene formaldehyde
2 % of bentonite
W/C = 0.67

5.4.2. Flow charts

The flow charts are the graphical presentation of the results from analysis of the video images of
the test injections. During the analysis, the time is noted to reach a certain height. The results
for the samples filled with one fraction of crushed brick are listed in Table 5.9. The same
information is given by the flowchart on Figure 5.23. Looking at the permeability of each
fraction as listed in Table 5.8, there seems to be some correlation between the permeability and
the progress of the grout in the porous medium. However, the ranking is not the same for
permeability and penetration speed.
Table 5.9: Progress of the grout in the samples filled with only one fraction
Height
cm
0
5
10
15
20
25
30
35
40
45

1/2
CEM III A
0
2,7
6,1
9,1
12,7
16,2
19,7
25,1
29,5
33,6

1/4
CEM III A
0
3,6
7,5
10,2
15,5
19,1
23,7
28,8
33,2
37,6

2/4
CEM III A
0
2,9
5,8
9,2
12,3
15,4
18,2
22
25,2
27,9

4/10
CEM III A
0
2,8
6,2
9,4
13
16,3
20,1
24,1
27,5
31

The 1-2 mm fraction, that has by far the smallest permeability, is only the second slowest to
reach the top of the sample. So, other parameters are involved besides the permeability of the
porous medium. One of these parameters is the total volume to inject. Before a grout can
continue to progress, the voids needs to be filled. The fraction 1-4 mm contains a bigger
volume of voids than the fraction 1-2 mm. Other possibility might be that it is easier to fill the

94

Grout injection of masonry, scientific approach and modeling

voids of the fraction 1-4 mm. This makes further progress impossible because no pressure can
be build up before the voids are filled. When the grout arrives at the top, the complete cylinder
is filled. If it is difficult to fill the voids, some pressure can be build up. The grout is able to
proceed, leaving part of the voids empty because the resistance to fill these voids is big. Hence
the denser and less permeable fraction requires less time to reach the top because a smaller
amount of grout is to be injected.

Flow charts of the Grout Injections

Injection pressure = 100.000 Pa
40
35

4/10
2/4
1/2
1/4

Time [sec]

30
25
20
15
10
5
0
0

10

20

30

40

50

Height [cm]

Figure 5.23: Flow charts of the samples filled with only one fraction of the crushed bricks
The flow curves are almost linear. Because of the increasing resistance of that part of the
sample that is already injected, one expects an evolution of the penetration depth that is ever
slowing down. This seems not to be the case. There are two possible explanations for this
behavior. First possibility: the resistance for the grout to penetrate is located at the injection
front. The rest of the resistance can almost be neglected with regard to this front resistance.
Some simple simulations at the end of this chapter show that by incorporating a front resistance
value, the simulations fit well with the experimental data.

Chapter 5 - Experimental program

95

Figure 5.24: Layout of horizontally split sample (a) and vertically split sample (b)
A second series of samples contains three layers of fractions of crushed bricks. The middle
layer is finer than the first and the last layer. The layout of such samples is shown in Figure
5.24 a Aim of this layout is to analyze what happens when a grout meets a zone with lower
permeability that is in series with a zone with higher permeability. Figure 5.25 clearly shows
how the grout slows down when arriving in the denser zone. The penetration speed decreases
significantly. It is strange however, to see that the penetration speed increases again to the
initial value ones the low permeable zone is passed. The front resistance theory is able to
explain also this observation.

96

Grout injection of masonry, scientific approach and modeling

Flowchart 2-4 1-2 2-4 mm

Injection pressure = 100.000 Pa

40
35

Time [sec]

30
25
20
15
10
5
0
0

10

15

20

25

30

35

40

45

Height [cm]
Figure 5.25: Grout penetration in a horizontally split cylinder, 2-4 1-2 2-4 layout

A similar effect happens when the low permeable zone is put at the bottom and at the top. Then
the penetration speed increases somewhat when the more permeable zone is reached.
Afterwards the penetration slows down again (Figure 5.26).
In order to analyze the behavior of the grout when a less permeable zone is in parallel whit a
high permeable zone, some samples were split vertically. The layout is then as represented in
Figure 5.24 b. Although the 1-2 mm fraction is on its own well injectable as shown in the
previous flow charts of the horizontally split cylinders and of the uniform cylinders, in
vertically split layout this fraction could not at all be injected at a first attempt. Therefore, it
was decided to make two more tests of the same layout. In one other sample the zone parallel to
the 2-4 fraction was only partially injected. In a third attempt, the zone was injected
completely, but somewhat later than the 2-4 mm zone. This is of course not obvious from the
flow chart (Figure 5.27), but is shown in Figure 5.28. The overall penetration is much slower
than in all previous cases. The slowing down is much more explicit. The penetration speeds up
somewhat after the split zone, but does not regain its original speed.

Chapter 5 - Experimental program

97

Flow chart 1-2 2-4 1-2 mm

Injection pressure = 100.000 Pa
35
30

Low speed

Time [sec]

25
20

High speed
15

Low speed

10
5
0
0

10

15

20

25

30

35

40

45

Height [cm]

Figure 5.26: Grout penetration in a horizontally split cylinder, 1-2 2-4 1-2 layout

Flow chart vertically split, 2/4 1/2 2/4

Injection pressure = 100.000 Pa
100
90
80

Time [sec]

70
60
50
40
30
20
10
0
0

10

20

30

40

50

Height [cm]

Figure 5.27: Grout penetration for the vertically split cylinder

98

Grout injection of masonry, scientific approach and modeling

Figure 5.28: Picture of the partly injected, vertically split cylinder

Due to its fineness, the lime grout, consisting of hydraulic lime, B-fluid 0 from Unilit and a
water/lime ratio of 1.5 without any admixtures nor superplasiticizer seems to have less hinder
from the finer fraction. No slowing down can be noticed. The total time to reach the top of the
cylinder however, is smaller for the cement grout. The lime grout shows a more Newtonian
behavior: the penetration speed continuously decreases resulting in a parabolic evolution
Figure 5.29. A second attempt was not successful in the sense that the grout blocked at 15 cm
from the top. Again, the parabolic behavior is clearly visible.

Lime grout vs Cementgrout

Injection pressure = 100.000 Pa
60
2/4 1/2 2/4 CEM III A

50

2/4 1/2 2/4 Lime - B

Time [sec]

40

Water

blocked

2/4 1/2 2/4 Lime - B

30
20
10
0
0

10

20

30

40

50

Height [cm]

Figure 5.29:

Comparison between the injection of cement grout, lime

grout and pure water of the horizontally split cylinder

Chapter 5 - Experimental program

99

Since it was expected that the dry masonry would absorb water out of the grout, the crushed
bricks were wetted in advance by simple injection with water. Then, the valve at the bottom of
the cylinder was opened and the water flew out of the sample. After 30 minutes the same
sample was injected with the cement grout. The injection of the prewetted sample behaved
completely linear. The time needed to reach the top was half of the time for the dry masonry
sample. Compared to water, the grout injection was only 3 seconds slower. Obviously, the
resistance to flow has been reduced significantly by the water injection. It is possible that this
reduction has to do with the front resistance that has disappeared. Besides, the conductivity of a
porous medium is related to the water content as is displayed in Figure 5.30. This means that
the conductivity to water of a dry porous medium is smaller than the conductivity to water of a
wet sample. The same reasoning goes for the conductivity to airflow: a porous medium with a
high moisture content will provide a greater resistance to flow of air than a dry sample [Groot,
1993].

Figure 5.30: The diffusivity for water of porous material is related to the moisture
content. Curve 1: experimental data, curve 2: zone with constant relation, curve 3:
exponential fitting Figure 10.1
At the water injection, the voids are already filled with water. Then the water is let out of the
sample. After half an hour, not all the injected water has flown away. Additionally, the water
injection has washed out and cleared the major flow channels.

100

Grout injection of masonry, scientific approach and modeling

Injection of Dry and Pre-wetted Cylinder

Injection pressure = 100.000 Pa
100
90

2/4 1/2 2/4 CEM III A split

2/4 1/2 2/4 CEM III A split/wet
2/4 1/2 2/4 Water

80

Time [sec]

70
60
50
40
30
20
10
0
0

10

20

30

40

50

Height [cm]

Figure 5.31: Behavior of the injection of the prewetted, horizontally split cylinder
From this experiment it is clear that prewetting can solve penetrability problems. But, since
there is no water absorption out of the grout, the mechanical strength of these particular samples
is very poor. Therefore prewetting has to be used with precaution.
As mentioned before, during some experiments the mass flow has been recorded. This enables
to filter the influence of unfilled voids. Besides, when numerical or analytical simulations are
done, the mass flow is calculated and not the injection height. Furthermore, since there is an
assumption that the main resistance to flow is situated at the injection front, once the sample is
filled, this resistance should disappear and the discharge should increase suddenly. Figure 5.32
indicates such a discharge increase. The vertical line indicates the moment when the cylinder is
completely filled. The increase however, is too small to be the consequence of the disappearing
of a major resistance to flow. Though, the discharge does increase and hence the total
resistance to flow decreases somewhat. The discharge gradually decreases after the small
increase when the cylinder is completely filled. This can be explained by two phenomena that
amplify each other. The flow channels are narrowed by the water absorption out of the grout.
Cement particles stick to the walls of the flow channels. By this, the overall resistance to flow
increases. Second phenomenon is the clogging up of the fine channels by the big or flocculated
cement grains that pass. Once a few particles are caught, the constriction acts as a filter for ever
smaller particles and finally the channel is completely blocked. Again, the overall resistance to
flow increases and hence the discharge decreases.

Chapter 5 - Experimental program

101

Mass flow through fraction 1-4 mm

Injection pressure = 100.000 Pa
10

0.08

0.07

Total Mass [kg]

0.06
7
0.05

6
sample is filled

Total Mass
Mass Flow

0.04
0.03

increasing discharge

Mass Flow [kg/s]

0.02
2
0.01

1
0
0

100

200

300

400

500

0
600

Time [sec]

Figure 5.32: Evolution of mass flow for the injection of the 1-4 mm fraction, basic grout

5.4.3. Simplified mathematical model

The experimental observations are used to build a first mathematical model. The experiments
also improved the understanding of the physical mechanisms taking place during the penetration
of the grout through the masonry. At first we tried to model the injections using Darcys law,
governing the laminar flow of fluids through porous materials. The flow is considered to be
laminar since the Reynolds numbers, that characterize the nature of flow for all used fractions of
crushed brick (Table 5.10) are far below the range that indicates turbulent flow.
Table 5.10: Reynolds number at the test injections for the used fraction

102

Fraction [mm]

Reynolds number

1-2

2,7

1-4

4,2

2-4

6,7

4 - 10

13,6

Grout injection of masonry, scientific approach and modeling

Figure 5.33 shows the inadequacy of Darcys law to model the injection tests. The average
velocities of the injections calculated using Darcys law are about three times higher than the
experimental ones. The qualitative difference in shape between the experimental and the
theoretical flow chart (the theoretical flow is not as linear as the experimental one) proves that
there is also a physical dissimilarity.

Darcy's law, uniform cylinders

Injection pressure = 100.000 Pa

50

Experiment (1-2 mm)

Darcy (1-2mm)
Experiment (1-4mm)
Darcy (1-4mm)
Experiment (2-4mm)
Darcy (2-4mm)
Experiment (4-10mm)
Darcy (4-10mm)

45

Time [sec]

40
35
30
25
20
15
10
5
0
0

0.05

0.1

0.15

0.2

0.25

0.3

0.35

0.4

0.45

Height [m]
Figure 5.33: Inadequacy of Darcys law to simulate the experiments
The observations mentioned above enabled us to develop the front resistance theory, based upon
the hypothesis that beyond the overall media resistance, there is an additional resistance to the
flow at the grout front. This front resistance is characterized by an equivalent height (Leq ) of
saturated crushed bricks, having the same resistance as the thin layer of dry masonry.
It is calculated by curve fitting of numerical simulation of the progress and the experimental
data. The optimized magnitude of this front resistance is not a constant but depends on the
fraction of the crushed bricks. The relation between the permeability of the crushed bricks and
the magnitude of the front resistance is found to be almost linear and can be expressed as:

Leq [m] ' 0,27 .E09 K [m 2] & 0,21

where K

(Eq 5.4)

= permeability of crushed bricks

The higher the permeability, the more important and the more dominating is the front resistance
(Table 5.11).

Chapter 5 - Experimental program

103

Front ResistanceTheory

50

Injection pressure = 100.000 Pa

45

Experiment (1-2 mm)

Fr. Res.(1-2mm), Leq=0,13 m
Experiment (1-4mm)
Fr. Res. (1-4mm), Leq=1.52 m
Experiment (2-4mm)
Fr. Res. (2-4mm), Leq=1.44 m
Experiment (4-10mm)
Fr. Res. (4-10mm), Leq=4.14 m

40

Time [ sec ]

35
30
25
20
15
10
5
0
0

0.05

0.1

0.15

0.2

0.25

0.3

0.35

0.4

0.45

0.5

Height [ m ]

Figure 5.34:

The combination of Darcys law and the front resistance theory fits very well
with the experiments

Although it is often used, curve fitting is not a very scientific method. In order to verify the front
resistance values, obtained by optimization for the short uniform cylinders, and to see if the
theory is able to reproduce and predict injections in longer cylinders and other configurations,
an injection was carried out on a split cylinder with a height of 0.9 m. The cylinder is
horizontally split in three parts of 30 cm each, filled with the 2-4 mm, 1-2 mm and 2-4 mm
fraction of the crushed bricks.
The front resistance is incorporated in the model. Pressure at the bottom of the cylinder is
supposed to remain constant. The grouts viscosity is supposed to stay constant according to the
finding that the grout coming out of the samples has the same rheological properties as the
injected grout. Further, the hydrostatic pressure is taken into account. The total resistance for
one dimensional flow through dry crushed bricks can thus be expressed as the Darcy resistance
(first term) and the front resistance (second term):

R '

with

L
Leq

Leq
L
L
% RFront '
%
K A
K A
K A

(Eq 5.5)

= length of injected zone

= equivalent length

and the corresponding mass flow equals:

104

Grout injection of masonry, scientific approach and modeling

Table 5.11: Relative magnitude of the front resistance

Fraction [ mm ]

1-2

1-4

2-4

4 - 10

Permeability [ m 2 ]

0.73 E-09

5.8 E-09

6.5 E-09

14.0 E-09

Leq [ m ]

0,12

1,52

1,44

4,14

Leq / K [ m-1 ] (x 109)

0,168

0,262

0,222

0,296

Rfront [Pa . s / m3]

1,07e+08

1,67e+08

1,41e+08

1,88e+08

q [

kg
?p
] ' ?
s
R

(Eq 5.6)

The integration is done numerically by means of finite differences.

The combination of (Eq 5.4) and (Eq 5.5) indicates that the front resistance is proportional to
Leq / K (Eq 5.7). For the Leq values found from the injections of the short cylinders (Eq 5.4),
and the permeability values listed in Table 5.8, this factor can be calculated. Table 5.11 lists
these values.

RFront '

Leq
K

(Eq 5.7)

In (Eq 5.7), /A is independent from the permeability or the granularity of the media. The
relation between the front resistance and the permeability and the linear regression line are
shown in ?.
Figure 5.35 shows that the calculations according to the front resistance theory, with the Leq
obtained from the optimization for the injections on the short cylinders, fit very well with the
experimental data for the long cylinder.

Chapter 5 - Experimental program

105

One Dimensional Injection, Time vs Mass

With Front Resistance
Injection pressure = 100.000 Pa
250
Front Resistance
Darcy's Law
Experimental Data

Time [ sec ]

200
150
100
50
0
0

Total Mass Flow [kg]

Figure 5.35:

Incorporating the front resistance theory in Darcys law makes the calculated
data to fit well with results from theinjection of the basic grout in the long
cylinder

5.4.4. Mechanical properties

In the first part of this thesis, it is mentioned that the mechanical properties are not as important
as one might think. Nevertheless, it is interesting to get an idea of the obtained mechanical
strength of the hardened grout. Since pure grout will not be present in the injected masonry, we
preferred to test the mechanical strength of the injected samples. For this purpose some of the
samples were not emptied. They were kept in normal laboratory conditions and tested after 28
days of curing. First, the plexiglass cylinder was removed and then the cylinder was cut in
slices. For these slices the sound velocity was measured to calculate the modulus of elasticity,
for some specimen the compressive strength was determined, for others the tensile splitting
strength.
The crushed brick particles are completely embedded in the grout. The crushed bricks play the
role of the aggregates in concrete. These particles absorb the water out of the grout. This way

106

Grout injection of masonry, scientific approach and modeling

the interfacial zone between the cement grout and the particles is not a weak zone as it is in
concrete. The compressive strength of most samples is very high. The strength is probably
much higher than the strength of a real masonry structure that is injected with the same cement
grout.
Because of the partly different curing nature of natural hydraulic lime based grouts, the
mechanical strength of the samples that were injected with hydraulic lime grout, was tested only
after sixty days. Only five specimen were tested in compression according to the Belgian
standard NBN B15 220. The results are listed in Table 5.12.
Table 5.12: Compressive strength of the injected cylinders
Description of the specimen

Compressive strength [MPa]

Cement grout, H = 20 cm, fraction 1 - 2 mm

41,1

Cement grout, H = 20 cm, fraction 1 - 4 mm

33,5

Cement grout, H = 20 cm, fraction 4 - 10 mm

22,4

Cement grout, H = 35 cm, fraction 1 - 4 mm

50,8

Lime grout, H = 20 cm, fraction 1 - 2 mm

11,8

The position of the samples is indicated by the height where they were cut.
The tensile strength of the injected samples is more important and hence most of the specimen
were tested in tension. The Brazilian splitting test according to NBN B 15-218 was applied.
The height of the samples is only 50 mm, which is lower than prescribed in the standard in order
to be able to test as many samples as possible. For each cylinder the tensile strength was tested
at different heights. The results are shown graphically in Figure 5.36.
The modulus of elasticity is an important property. The compressive strength is directly related
to the modulus of elasticity. When a massive material is loaded, that part that has the highest
modulus of elasticity carries the biggest part of the load. Besides, the modulus can be obtained
from the non destructive testing of the sound velocity in the sample.

Chapter 5 - Experimental program

107

Tensile strength of grouted cylinders

8

Tensile Strength MPa

previously wet
Fr 4 - 10

Fr 2 - 4
5

Fr 1 - 4
Fr 1 - 2

Linear (Fr 1 - 2)
3

Linear (Fr 4 - 10)

Linear (Fr 2 - 4)

Linear (previously wet)

1

Linear (Fr 1 - 4)

0
0

10

15

20

25

30

35

40

45

Height [cm]
Figure 5.36:

Results of the splitting test for the various specimen in function of the
height location in the test cylinders

The measurement of the sound velocity in a material is a relatively simple and fast test. The
transmission time is measured for a sound wave to traverse the specimen. The relation between
velocity, modulus of elasticity and density is given by equation (Eq 5.8).
? [m/s] '

Where

v
E
?
?

=
=
=
=

E
. f (v)
?

(Eq 5.8)

velocity of the sound

Young modulus
density
Poissons ratio

The velocity is measured on each sample.

108

Grout injection of masonry, scientific approach and modeling

Young's modulus' variation in grouted cylinders

35000

Fr 4 - 10
Fr 2 - 4
Fr 1 - 4
Fr 1 - 2
Linear (Fr 2 - 4)
Linear (Fr 1 - 2)
Linear (Fr 1 - 4)
Linear (Fr 4 - 10)

E [MPa]

30000

25000

20000

15000
0

10

20

30

40

50

Height [cm]
Figure 5.37: Modulus of elasticity for the different fraction in function of the height

It is obvious that the mechanical results highly depend on the position of the specimen in the
original sample. The strengths of the specimen taken at the base of the cylinder are similar.
Then, for some samples the relation between mechanical strength or modulus of elasticity show
a positive gradient whereas for other there is a negative gradient. When this was noticed, the
density variation was also investigated. The same relation was found. The density variation in
function of the height is given in Figure 5.38. Two phenomena can cause a density gradient.
While the grout is going up through the cylinder filled with crushed bricks, its W/C ratio
decreases because the bricks absorb some of the water. This water absorption leads to an
increase of the density of the grout.
The relation between water content of the grout and density of the grout can be specified
analytically. Let us assume that the density of cement ?cement = 3050 kg/m3. Let us further
assume that the grout is only composed of cement and water without air. The density of the
mixture of water and cement equals:
? grout '

Chapter 5 - Experimental program

1 % W/C
1
W/C
%
?cem
? wat

(Eq 5.9)

109

The grout that penetrates gradually loses water and becomes denser. This creates a positive
density gradient.
As the crushed bricks are finer, the specific surface is bigger and consequently the water
absorption is higher. This explains why the density gradient increases as the pore size gets
smaller. If the pores are greater, this phenomenon is masked by another process. The
gravitational forces that are acting upon the cement particles in the grout are superior to the
electro statical and viscous forces that keep the particles in suspension. When the injection
stops, the particles start falling towards the bottom of the cylinder. This causes a negative
density gradient. In this case, when the pores are greater, the water absorption is also smaller.

Density variation in grouted cylinders

2100

E [MPa]

2000

Fr 4 - 10
Fr 2 - 4
Fr 1 - 4
Fr 1 - 2
Linear (Fr 1 - 2)
Linear (Fr 1 - 4)
Linear (Fr 4 - 10)
Linear (Fr 2 - 4)

1900

1800

1700

1600
0

10

20

30

40

50

Height [cm]
Figure 5.38: Density variation in function of the height for the various specimen

It is logic that when the grout is heavier, because richer in cement, the mechanical results are
better. It is normal that these phenomena are more significant if the grout is not very stable and
if the water retaining properties of the grout are poor. This observation has important
consequences for practical jobs. The upper zone of an injected piece of masonry might lose too
much cement and therefore become weaker than the bottom zone. The mechanical strength and
the stiffness are negatively affected. This observation shows that the maximum injection height
is not only limited by the hydrostatic pressure, but also by the strength gradient that arises.
The mechanical strength of the pre-wetted samples is very poor. The specimen taken at the top
can be crushed by hand. Apparently they lost almost all the binding cement. Also the specimen,
taken at the bottom of the cylinder are significantly weaker than those that were not prewetted.

110

Grout injection of masonry, scientific approach and modeling

5.5. Important findings

5.5.1. Blocking mechanisms
Granularity of the cement
The first mechanism depends on the granularity of the cement and the crack size distribution of
the brickwork to inject. In literature, there are different empirical formulas expressing the
requirements for the granularity of the cement with regard to the smallest crack diameter of the
brickwork. An incompatible grain size distribution of the used binding agent results in a sudden
obstruction of the injection. The biggest grains should at least be smaller then 0,1 to 0,15 times
the smallest crack width [Benhamou, 1994][Miltiadou, 1990]. Flocculated cement grains act as
big cement grains. Therefore, a proper mixing procedure is necessary. For the same reason it
seems to be impossible to produce an injectable grout without superplasticizers. They bring the
W/C ratio to acceptable values and more important, they have a deflocculating action.
Stability of the grout
The second mechanism depends on the stability of the grout. When flow slows down, the
cement particles in an unstable grout sink to the bottom of the flow channel. This narrows the
channel and finally blocks further injection. Addition of stabilizing admixtures (i.e. bentonite or
ultra fines) significantly improves the stability and the injectability of the grout [Miltiadou,
1990][Paillre, 1993]. For this reason it is important to check the stability of the grout and the
evolution in time of the stability [Van Rickstal, 1995(1)].
Water absorption out of the grout
The third possible blocking mechanism is the result of water absorption out of the grout into the
masonry Figure 5.39. When meeting a dry and absorptive masonry it is important to use a grout
with excellent water retaining properties. Here again, a stable grout will be more able to retain
the water.

Chapter 5 - Experimental program

111

Figure 5.39: The dry bricks absorb water out of the grout

A prewetting of the brickwork by water injection is not recommended. First of all it can be
dangerous when the stability of the construction is doubtful. The water will decrease the
internal cohesion and friction and could cause a collapse of the building. By injecting water the
masonry will be saturated and, although the grout will pass easier, the W/C ratio of the grout
remains very high and produces a weak binding material. The little absorption needed for a
good adhesion between grout and masonry does not occur and the mechanical improvement will
be poor. For the test injection on the plexiglass cylinders the tensile strength of the pre-wetted
samples was only 1 N/mm2 or lower, whereas the tensile strength of the dry samples varied
between 3 N/mm2 and 7.5 N/mm2 (Figure 5.36).
This third blocking mechanism produces a slow obstruction of the injection. When the grout
loses too much water the viscosity and the yield stress become too high to make any further
penetration possible. A stand still of the flow allows locally a big amount of water to be
absorbed and this can be crucial for further penetration. Therefore it is recommended to inject
continuously keeping the pressure as constant as possible.
Pressure losses, thixotropy
The resistance of the fine cracks and voids to the flow increases at greater distance form the
injection hole. Due to this increasing resistance finally the yield stress will not be reached at
the front of the injection. Only gravity could compensate this pressure losses for downward
flow and make the grout to keep on flowing. Due to the limited discharge of the injection pump
and the radial injection, the flow velocity will drop and thixotropic behavior of the grout will
enhance the above phenomenon. Also this mechanism causes a gradual stop of the injection.
Decreasing the yield stress by an appropriate composition and mixing procedure of the grout can

112

Grout injection of masonry, scientific approach and modeling

retard this last blocking mechanism. This is one more reason why it is not possible to formulate
a well injectable grout without superplasticizer.
5.5.2. Observations
Many test injections were carried out, varying the parameters of the grout, the injection pressure
and the fraction of the crushed bricks. The following observations were made:
C
C
C
C

when an injection stops, it is not possible to restart the flow by increasing the injection
pressure;
there are two different blocking mechanisms: a sudden obstruction of the flow or a
gradual obstruction Figure 5.28;
when a permeable zone is in parallel with a less permeable zone, the latter will not be
injected, although this medium is, on itself, perfectly injectable;
when injecting an unstable grout, it can be noticed that cement particles are sinking to the
bottom of the cylinder, causing a strength gradient and heterogeneities in the
consolidated masonry;

Figure 5.40: Heterogeneous consolidation of the reproducible samples

because of instability of the grout
when the sample is prewetted by water injection, the grout injection is executed many
times faster than through a dry sample. However, the strength after curing and the
adhesion of the grout to the original material is very poor. Therefore, prewetting the
masonry structure can only be made after cautious preliminary research.

5.6. Conclusions from the experimental program

Chapter 5 - Experimental program

113

This chapter about the experimental program explains a number of test methods that are newly
developed or adapted to the peculiarities of the grouting process. The methods that are testing
the rheological properties of the grout, in the laboratory and on the site should be part of a
scientific approach of the consolidation injection. An injection grout basically consists of a
binding agent, water and admixtures eg. a superplasticizer and a stabilizer. The influence of
every component on the viscosity and the stability is analyzed.
Furthermore, the experimental program, especially the numerous test injections that where
performed on the reproducible masonry samples, provide a better understanding of blocking
mechanisms that occur. The influence of a prewetting procedure is investigated and although
prewetting could improve the penetration of the grout inside the masonry, it has dangerous
effects on the mechanical strength, the strength gradient and the stability of the injected grout.
The water that is absorbed out of the grout provides a better adhesion to the flow channels.
The analysis of the results of the mechanical tests, reveals exciting findings about the strength
gradient and the density gradient that originates from the sinking of the cement grains.
Finally, the experiments provide data about the penetration speed and depth of the grout inside
the reproducible masonry sample. These data are used to calibrate and check the numerical
simulations.

114

Grout injection of masonry, scientific approach and modeling

Chapter 6.

Rheology of grouts

6.1. Introduction to the rheology of dispersions

Considering the final aim of this research program of predicting the flow of the grout inside the
masonry, some rheological considerations about dispersions are justified. Rheology of
dispersions is an important research field on its own. The complete characterization of the
rheological behavior of grouts: the rheological parameters in function of the concentration, the
influence of mixing procedure or more generally the history of the grout, the problems to
measure both the viscosity and the shear strength of grouts in function of the shear rate is beyond
the objectives of the thesis. Many books are available about the rheology of dispersions. Most
researchers in the field of grouting or in the field of grout compositions, spend a larger or
smaller part of their work on the measurements of the rheological parameters of grouts and try to
discover some trends in the influence of the composition on the rheology of the grouts. Their
qualitative results will be used in the present study. Some assumptions will simplify the
complex rheological behavior of cement grouts. If necessary, the implications of these
assumption will be mentioned.
A liquid with small particles scattered all over is called a dispersion. If the magnitude of the
particles is very small in comparison with the other dimensions in the given situation and if the
particle concentration is small, one can characterize the behavior of the dispersion in the same
way as the dispersing fluid. The particles present in the dispersing fluid do not affect the flow
in any way. This is certainly not true for injection grouts. If the concentration increases from
zero % the particles will initially not hinder each other. The viscosity for such a dispersion can
approximately be expressed by the relation, formulated by Einstein (Eq 6.1) [Wilkinson, 1960].

' 1 % a?
0

(Eq 6.1)

If the concentration increases further, generally a second order term appears in the description
of the rheological behavior (Eq 6.2) [Wilkinson, 1960].

' 1 % a ? % ? 2
0
where

?
0

a,

(Eq 6.2)

volumetric concentration of particles

the viscosity for the pure fluid without any particles [Pa.s]
viscosity at concentration ? [Pa.s]
constant factors, depending on the type of fluid and particles

The importance of particle size is linked to the presence of Brownian motion. The influence of
the particle size on the rheological behavior is incorporated in the parameters a and . This is
Chapter 6. Rheology of grouts

115

the omni-present randomizing process which will always seek to generate a statistical
distribution of the positions and orientations of particles, while any impose of motion (shear or
extension) seeks to impose some organizing effect on the particles. Thus when speaking of size
we must realize that it is the ratio of viscous to Brownian forces that matters most. Also the
shape of particles influences the viscosity in two main ways: first it gives the particles a defined
orientation vector. Both, the Brownian motion and flow will compete to influence this
orientation vector. Secondly, irregular shape implies that the effect will be felt in the liquid to a
greater distance than with a sphere of equal volume. Therefore, one instinctively expects a
larger a-factor in (Eq 6.1) for an irregular shape than for a spherical shape. If the particles
carry a charge, the electro-viscous effect plays. Each particle is surrounded by a charge cloud
of counter-ions and any deformation of the suspension tends to deform this cloud. Thermal and
electrical forces try to counter this distortion by moving ions relative to the fluid, thereby
dissipating extra energy and therefore increasing the viscosity.
A dispersion can be stabilized by surfactant stabilization. Surfactants not only play a large role
in the stabilization of formed dispersions but also in their original formation by breaking down
of the solid to reduce the primary particle size. This is achieved by both the action of the
superplasticizers and by the ultrasonic mixing procedure. The surfactant leads to a strongly
adsorbed water of fluid layer which can cause stabilization of the particle distribution in the
dispersion. If the adsorbed surfactant is of the nonionic type then it comes into the class of
stearic stabilization. If the adsorbed polymer is ionic, then irreversible flocculation is
prevented by the repulsive forces generated from the presence of an electrical double layer in
between the particles and the particle solution interface. In this way, (super)plasticizers are
active.
Whereas the approach, described above, gives an insight in the motion of dispersions, most used
for calculations are empirical relationships between viscosity and shear rate. In growing order
of complexity one can list these relationships in Table 6.1.
Table 6.1: Listing of frequently used rheological models [Wilkinson, 1960]
Name

formula

number of parameters

Newtonian

t ' ?0

Power law

t ' ? ?0 n

Bingham

t ' t 0 % ?0

Herschel/Bulkley

t ' t 0 % ? ?0 n

Cross

116

t ' 4 ?0 %

0 &4

1 % ? ?0 n

Grout injection of masonry, scientific approach and modeling

In the search for a useful expression one works up from the simplest. Many mistakes are made
in the choice of a model, not because the model was to simple but because the range of shear
rates over which the data were fitted to determine the parameters was inappropriate to the
situation it was used for. Frequently occurring mistakes are: calculating the shear stresses by
extrapolating from only a few relatively high shear rates to obtain a simple Bingham equation
[Toorman, 1995]! Another inaccuracy arises by considering the dispersion as a continuum,
disregarding the particles inside. These particles give rise to three phenomena. These
phenomena create a gradient in concentration. Therefore, the dispersion can not be considered
being a homogeneous liquid. The phenomena are:
C
In many situations the continuous fluid phase in dispersions moves relative to the
suspended phase. In case of instability, particles move down and liquid moves up when
no three dimensional structure of particles or yield stress of the continuous phase can
prevent the particle from sedimenting under gravity.
C
When a water based dispersion is flowing through a medium with capillary action, this
medium sucks out part of the continuous phase. The concentration increases and hence
viscosity increases too. The experimental observations indicate that this is what
happens when water is absorbed out of the grout.
C
And finally, it is possible that it is easier for a system to appear to flow with the
continuous phase moving through the dispersed phase. The remaining particles form a
filter cake through which no flow is possible. This narrows the flow channel because
the filter cake sticks to the wall of the flow channel (Figure 6.1).

Figure 6.1: The continuous phase moves through the dispersed phase, a filter cake is formed and
finally the flow stops
As in a grout the particles are slightly charged, there is a repulsive force. The attractive forces
are the London dispersion forces arising from the coupling of the fluctuations in charge

Chapter 6. Rheology of grouts

117

distribution of the electron clouds of the atoms making up the particle. The repulsive forces
grow very fast when the particles distance decreases. A system for which the repulsive forces
win is called a dispersive structure.
The dispersion can be concentrated to such an extent that the average inter-particle separation is
of a similar magnitude as the working range of the inter-particle forces. As there is often a
preferred state for these systems: the state of minimum energy with the particles being as far
from their neighbors as possible, there is the possibility of elasticity. Because the forces
holding the particles in these positions are not very large, we can still cause flow once we
overcome the critical shear stress. So we can expect a classical Bingham type of fluid.

6.2. Non-Newtonian behavior of aqueous dispersions

A liquid is called to be Newtonian if it obeys Newtons Law for fluid motion:
du
t '
' ?0 where is constant
(Eq 6.3)
dy
Almost every aqueous dispersion shows non-Newtonian behavior in simple shear flow, which
may originate from any of the following:
C
forces of attraction or repulsion existing between the particles, which are modified by
the applied flow. Therefore, the material exhibits a yield stress in order to overcome the
forces of attraction present between the particles in that state.
C
the adsorption of the suspending liquid onto the particle surface. Therefore, such a
dispersion shows shear thinning behavior. It probably originates in the progressive
shearing off of layers of adsorbed water from the particles, reducing the apparent
volume concentration with a consequent decrease in viscosity.
C
mechanical interference between particles. Such interference will increase with
increasing shear rate, with increasing concentration and with decreasing particle
isometry. A shear thickening behavior will be the consequence.
C
mechanical fracture of particles. This results in the removal of irregularities that are
hindering the motion. Again a shear tinning effect will occur.
C
time dependent behavior. Flocculation arising from attractive forces normally takes a
finite time to re-establish itself when broken down by shear.
The complexity of rheology of dispersions can be underlined if we add that it is possible for a
dispersion to exhibit a yield stress, shear thinning, shear thickening time dependence and
elastic-viscous behavior in one and the same dispersion depending on the level of shear rate at
which the measurement is made.
Forces of attraction and repulsion will invariably exist between particles in suspension, since
London-van der Waals dispersion forces will always be present, though only significant when
particles are close together. Repulsive forces generally mean that the particles can move as
individuals and the suspension is completely dispersed, while attractive forces mean that the

118

Grout injection of masonry, scientific approach and modeling

particles are less likely to be able to move as individuals, and that an linking structure is present
throughout the suspension.
On top of this, some dispersions show a time dependent behavior. This behavior is due to the
fact that temporary bindings are formed as soon as the fluid motion stops. These bindings need
to be broken if the fluid starts moving again. The longer the fluid rests, the stronger or the more
numerous these bindings are. The following derivations do not take into account this time
dependent behavior, called thixotropy. Thixotropy will only occur after a period of stand still.
Because of the permanent stirring in the collector, thixotropy will not happen there. Besides,
when an injection is properly executed, the grout flow will never stop until the final blocking.

6.3. Flow of a dispersion in a cylindrical tube

6.3.1. General equation
The following derivations are obtained by considering the dispersion as a continuum
disregarding the influence of the particles. We also suppose that the next conditions are
fulfilled:
C
We deal with laminar flow, all fluid parts are moving parallel to the axis of the tube
C
No shear movement occurs at the walls, the speed of the fluid part at the wall is zero
C
The shear rate in one point is only function of the shear in that point

Figure 6.2: Dynamic state of equilibrium for an annulus between radius

r and r + dr

We now consider an elementary volume being a annulus between the cylinders with radius r and
r + dr. We suppose that all transient phenomena are over and that the system is in stationary
motion. This dynamic equilibrium state is expressed by equation (Eq 6.4).
2pL (r t &(r % dr) (t %dt )) % (p1&p2) 2p r dr ' 0

Chapter 6. Rheology of grouts

(Eq 6.4)

119

t % dt ' t %

With

dt
dr
dr

p & p2
d(rt )
' 1
r
dr
L

one gets

t '

After integration

p1 & p2 r
2
L

(Eq 6.5)

(Eq 6.6)

Taking into consideration the boundary conditions, (Eq 6.6) can be reformulated as
t
r
'
tw
R

and

?0 ' &

Where

t
tw
?p
L
D
f(t )

=
=
=
=
=
=

tw '

R ?p
D ?p
'
2 L
4 L

du
r
' f (t ) ' f (t w )
dr
R

(Eq 6.7)

(Eq 6.8)

shear stress
the shear stress at the wall of the tube.
p1 - p2 = pressure difference
length of elementary volume
diameter of the tube
rheological behavior expressing the relation between the shear
rate and t
R

The discharge can be calculated by the integral

Q '

2pr u dr

providing the relation of Rabinowitsch (Eq 6.9) [Midoux, 1985]

tw

Q
1
'
t 2 f ( t ) dt
3
3 m
pR
tw 0

(Eq 6.9)

By specifying the relevant rheological behavior f(t ) one finds an expression for the relation
between pressure losses and the discharge. The Rabinowitsch equation is only valid in case of
laminar flow: the forces of inertia must be weak in comparison with the viscous forces.
6.3.2. Newtonian fluid

120

Grout injection of masonry, scientific approach and modeling

As an of example one can derive the well know formula of Hagen-Poisseuille for the laminar
flow of a Newtonian fluid in a cylindrical pipe. For a Newtonian fluid, relation (Eq 6.10)
expresses the relation between share rate and viscosity.
t
?0 ' f(t ) '
(Eq 6.10)

t
Q
' w . When we take into consideration relation (Eq
After integration we find
4
pR3
6.7), we find the Hagen-Poisseuille relation (Eq 6.11) between discharge and pressure gradient.
Q'

pR 4 ? p
pD 4 ? p
.
'
.
8 L
128 L

(Eq 6.11)

From (Eq 6.11) it follows that for a Newtonian fluid, the discharge will never be zero as long
as there is any finite pressure gradient and as long as the pipes diameter is not zero. The
observations learn that in case of injection the flow can drop to zero. So we conclude that the
flow of a grout through masonry can not be described correctly by using the above equation (Eq
6.11).

Chapter 6. Rheology of grouts

121

6.3.3. Bingham fluid

Generally, dispersions are considered to be Bingham fluids. The basic equation of the Bingham
fluid is (Eq 6.12) as was already listed in Table 6.1.
t ' t c % ?0

(Eq 6.12)

Or written in a different way according to eq. (Eq 6.8)

?0 ' f (t ) '
?0 ' f (t ) '

t & tc

when t > t c

(Eq 6.13)

when t < t c

When the yield stress is less then the critical yield value t c, the fluid is not sheared any longer.
du
The fluid then moves as a block or rigid body:
' ?0 ' 0 for that part of the fluid. In a
dy
cylindrical pipe this gives approximately the velocity profile as shown in Figure 6.3. When the
configuration is such that the critical yield stress is bigger then the yield stress at the wall of the
pipe, the fluid will not move at all, since for the complete cross section of the flow channel,
there is not one location where the shear stress exceeds the critical shear stress.

tc
Vc

Figure 6.3: Velocity profile for a Bingham fluid

Introducing the rheological model from eq. (Eq 6.13) in the equation of Rabinowitsch (Eq 6.9)
gives:
tw

Q
1
'
t 2 (t & t c) dt
3
3 m
pR
t w tc

122

(Eq 6.14)

Grout injection of masonry, scientific approach and modeling

The integration of eq. (Eq 6.14) provides the relation of Reiner-Buckingham [Midoux, 1985]
[Schowalter, 1978].
p D4 ? P
4 4t
L
1 4t L 4
[1 & ( c
)% ( c
)]
Q '
(Eq 6.15)
128 L
3 D ?P 3 D ?P
When the critical yield strength t c equals zero, the second factor in (Eq 6.15) vanishes and the
law of Hagen-Poisseuille (Eq 6.11) comes out again.
6.3.4. Discussion on the Reiner-Buckingham formula
Equation (Eq 6.7) expresses the magnitude of the shear stress at the wall of a cylindrical flow
channel. At that place, the shear stress is the highest. The layer that is making contact with the
wall does not move, this is one of the assumptions. Assuming a constant injection pressure, the
shear stress at the wall will decrease when the grout penetrates the channel. This corresponds
to an increasing L value in equation (Eq 6.15). If the shear stress t p decreases to the critical
shear strength of the fluid, no further shearing off will take place. A shear stress at the wall of
the flow channel which is lower than the critical shear strength will result in a complete stand
still of the fluid. This happens when:
D ?P
(Eq 6.16)
4 L
When the injection pressure is too low in relation to the length of the channel, the shear strength
will not be reached any longer, even not at the wall of the flow channel. This causes the flow to
stop. The discharge drop to zero if the correction factor from equation (Eq 6.15), taking into
account the Bingham behavior of a fluid, vanishes. This happens when:
tc $

4 tc L
D ?p

' 1

(Eq 6.17)

and hence, there is a linear relation between the penetration depth (L), the yield strength (t c), the
pressure difference over the channel (? p) and the diameter of the channel (D). Combining this
with the problem of water absorption by the dry masonry out of the grout, it can be stated that the
penetration depth of the grout will always be limited. The explanation that is generally
mentioned is thickening of the grout because of the water absorption. Thickening magnifies the
shear strength. There are reasons to believe that the limited injection depth is not due to
thickening in the first place but that the water absorption cause a reduction of the diameter of the
flow channel by forcing the cement particles to stick to the wall. Experiments show indeed that
the grout that exits the test samples, has the same properties as the original grout.

6.3.5. Numerical simulation

Chapter 6. Rheology of grouts

123

To clarify the influence of the yield strength on the progress of the grout in a cylindrical pipe
(Figure 6.4), a numerical example has been constructed. For viscosity, yield strength, pressure
and pipe dimensions realistic values for the grouting case are used taken from own experimental
data or literature [Benhamou, 1994]. The simulations, plotted in Figure 6.5 shows how a
Newtonian fluid and a Bingham fluid progress in cylindrical tube. For the Newtonian fluid the
differential equation can be solved analytically. Let x be the coordinate of the position along the
axis of the pipe, x = 0 at the entrance of the pipe.

Figure 6.4: Configuration to simulate the flow of a Newtonian and Bingham fluid
through a one dimensional tube

pR 4 ? p
pD 4 ? p
.
'
.
8 x
128 x

Q(x) '

(Eq 6.18)

Taking into account that ? (x) ' dx ' Q(x) with S(x) being the cross section of the pipe at
dt
S(x)
position x, this can be transformed to
dx
'
dt

p D4
?p
D2 ? p
.
'
32 x
pD 2 x
128
4

(Eq 6.19)

Integration of (Eq 6.19) for the case of one pipe with constant diameter yields
t1

m
0

124

dt '

32x
dx
m D 2? p
0

(Eq 6.20)

Grout injection of masonry, scientific approach and modeling

where t1 is the time when position X is reached by the fluid. The solution of this equation is
given by
t1 '

16 X 2
D2 ? p

(Eq 6.21)

Table 6.2: Numerical values for the worked out example

Name

Symbol

Value

Pipe diameter

2 mm

Inlet pressure

?P

10.000 Pa

dynamic viscosity

0.02 Pa.s

Yield strength

tc

1 Pa to 5 Pa

Position at beginning

x0

0.2 m

Time step

?t

0.2 s

The implicit relation between discharge and pressure difference of equation (Eq 6.15) is not so
easy to solve analytically. Therefore we choose the numerical solution by the finite difference
method. The differential equation can be converted into a finite difference equation. For
calculating the progress we need a non-zero x0-value. In this example we arbitrarily took 0.2 m
as x0.
xn ' xn & 1

D2 ? P
%
32 xn & 1

Chapter 6. Rheology of grouts

4
1 &
3

4t c xn&1
D ?P

1
%
3

4t c xn &1
D ?P

?t

(Eq 6.22)

125

Flow through a cylindrical pipe

3

Distance in pipe [m]

2.5

Newtonian fluid
Bingham fluids

Increasing yield strength

1.5
1
0.5
0
0

10

20

30

40

50

60

Time [s]

Figure 6.5: Influence of yield stress on penetration in a narrow pipe

Figure 6.5 plots the progress of both fluids in the cylindrical pipe. As can be seen the Bingham
fluid slows down much faster. Besides the Bingham fluid with yield stress of 5 Pa (Figure 6.5,
lowest curve) would never be able, under the given conditions, to penetrate further than 1 m. At
that moment the shear stress at the wall of the 2 mm channel would not exceed the critical yield
stress of 5 Pa and the movement would stop as expressed by equation (Eq 6.16).

6.4. Conclusions
In this chapter a theoretical base is given for the rheological behavior of dispersions. The
presence of many small particles influences the behavior of the fluid. Generally, a dispersion is
seen as a Bingham fluid. The equations that relate the discharge and the pressure gradient are
build. Still some aspects of a dispersion are not present in the description of a Bingham liquid:
time dependent features and the influence of the particle size. The equation that expresses the
relation between the discharge and the pressure gradient will be used in the model.

126

Grout injection of masonry, scientific approach and modeling

Chapter 7.

Flow of fluids through porous media

7.1. Introduction
Studying flow of fluids through porous materials is of importance in many fields of engineering,
e.g. petroleum engineering, soil mechanics and ground water hydrology. Also for many
applications in transport phenomena in the research field of building physics, masonry is dealt
with as a continuous porous medium, eventually with varying pore structure properties and
permeability. This continuum approach is used in the calculation of moisture transport,
combined moisture and air transport, soluble salt transport etc... For these purposes the
continuum approach is valid and leads to very good simulations of experiments and to good
predictions for particular situations. The driving force for al these transport phenomena is
capillary action, the transport phenomenon is also called imbibition. In case of injections the
main driving force is the injection pressure. In literature a penetration caused by an external
pressure is called invasion. In this chapter, attention is payed to the description of a porous
medium, Darcys law , the formulation and solution methods of the differential equations of fluid
flow through porous materials. However, the main transport for an injection occurs through
relatively big flow channels. The capillary forces in these channels can be neglected in
comparison with the driving force of external pressure.

7.2. Structure and properties of porous materials

7.2.1. Definitions
In the most general sense, a porous material is a material containing holes. This definition is too
general: a metal part containing a bore hole is definitely not a porous material. Therefore, we
state that a porous material is a solid containing holes or voids, either connected or nonconnected, dispersed within it in either a regular or random manner. It is obvious that great
variations in the size and structure of pores exist.
A fluid can flow through a porous material only if the pores are interconnected. The
interconnected pore space is termed the effective porosity, whereas the whole of the pore space
is termed the total porosity. The total pore space is affecting the density and strength of a
material, but only the effective pore space affects the permeability and related properties. The
smallest pores are termed interstices, the large holes are termed caverns. The intermediate
range of voids are termed the pores. Porous materials can be classified according to the
packing of the pores: ordered porous material and random porous material. Most natural
porous materials contain randomly distributed pores. On the microscopic level their pore
structure must be described in terms of statistics. Even so, it is possible to treat the flow of
fluids through such materials on a macroscopic basis in precise terms. At first stage we will
describe the flow on a macroscopic scale. Many theories have been formulated which attempt
to relate in a detailed manner the macroscopic properties of porous materials to the statistical

Chapter 7. Flow of fluids through porous media

127

properties of their microscopic structure, to relate pore size distribution to the macroscopic
properties of the material. While such theories contribute greatly to our understanding of basic
physical processes within porous media, they do not, in general, contribute to the solution of
problems on a macroscopic scale. Probably, the increased computer power will soon enable
to deal with the microscopic level in order to calculate macroscopic behavior.
The porosity of a porous material is the fraction of the bulk volume of the material occupied
with voids.
V
Volume of pores
f ' P '
(Eq 7.1)
Bulk Volume
VB
The volume fraction not occupied with voids, but with material can be expressed by
1 & f '

Vs
VB

'

Volume of solids
Bulk volume

(Eq 7.2)

As defined above distinction has to be made between effective porosity and total porosity.
7.2.2. Methods for porosity measurement
Several methods can be used for measuring the effective or global porosity. To measure the
total porosity the direct method is the only method. The method consists in determining the bulk
volume, crushing the specimen to remove all pores and then measuring the remaining volume.
The use of a pyknometer provides the necessary accuracy to the measurement of the volume of
the crushed material.
The gas expansion method provides a percentage of effective porosity. This percentage can be
calculated using equation
V B & Va & Vb
f eff '

Where

f eff
VB
Va
Vb
P1
P2

=
=
=
=
=
=

P2
P2 & P1

(Eq 7.3)

VB

effective porosity
bulk volume of sample
volume of sample chamber
volume of second chamber
initial pressure
final pressure

The Mercury intrusion method is based on the fact that, due to the surface tension and the nonwetting properties of mercury, external pressure is necessary to enter mercury in a sample. The

128

Grout injection of masonry, scientific approach and modeling

finer the diameters of the pores the higher the pressure needed to fill the pore with mercury.
Therefore, this method provides a pore size distribution. This distribution is in favor of the
small pores since the volume of a bigger pore that is connected to the outside by smaller pores
will be counted as small pores volume. This volume will only be mercury filled when the
pressure corresponding to the filling of the small pores is reached.
The Imbibition method yields the effective porosity. The technique consist of two weight
measurements. One in dry condition and one after vacuum saturation of the material. After
vacuum saturation all the pores are considered to be filled with water. The effective pore
volume can be calculated with equation (Eq 7.4).
Msat & Mdry
Vp '
(Eq 7.4)
?w
Since weighting can be very accurate this is probably the best method for effective porosity
measurement despite the long time needed for vacuum saturation.
More recently, with the aid of image analysis, microscopical methods are developed.
Advanced image analysis of an enlarged part of the sample can even provide a pore size
distribution. This method gives, although more complicated and expensive, much more reliable
results then the mercury intrusion method. This method does not affect the pore size distribution
in favor of the smaller pore size as is the case of the intrusion method because of the bottle neck
phenomenon. For weak materials, the intrusion method results in the collapse of the pore walls.
7.2.3. Permeability, Darcys law
Permeability is that property of a porous material which characterizes the ease with which a
fluid may be made to flow through the material by an applied pressure gradient. Permeability is
the fluid conductivity of the porous material. The equation which defines permeability in terms
of measurable quantities is called Darcys law. If horizontal linear flow of an incompressible
fluid is established through a sample of porous material of length L in the direction of the flow,
and cross-sectional area A, the permeability K, of the material is defined as:
K '

q
?P
A
L

(Eq 7.5)

In this equation q is the flow rate in volume per unit time and is the viscosity of the fluid. ? P
is the applied pressure difference across the length op the specimen. Dimensions of the
permeability are m2. The permeability is mostly determined using a setup similar to the one in
Figure 7.1.

Chapter 7. Flow of fluids through porous media

129

area A

Pa

porous medium

Pb

Figure 7.1: Possible setup for measuring the permeability

For a plate, the shear force per unit area between the solid surface and a fluid tangent to it is
given by Newtons equation:
F '
where

dv
dz

(Eq 7.6)
solid

velocity of the fluidin the direction of the flow

z
the distance perpendicular to the direction of the flow
It is reasonable to suppose that, for a steady laminar flow, the lateral forces, perpendicular to
the main macroscopic flow direction, associated with the microscopic random variations in
velocity average to zero over any macroscopic volume. However, the inertial forces in the flow
direction will not average to zero and hence will only be negligible for low flow rates. The
only non-zero macroscopic force exerted on the fluid by the solid is that associated with the
viscous resistance to flow. For steady flow this force must be in equilibrium with the external
and body forces on the fluid. The fact that inertial forces can be neglected in comparison with
the viscous forces results in a low Reynolds number (Eq 7.12). Darcys law is only valid for
those situations that show a low Reynolds number. The viscous resistance to the flow as
presented in Figure 7.1 is directed opposite to the flow direction and can be obtained by
integrating equation (Eq 7.6) over the microscopical surfaces in the complete volume. The
velocity v and hence dv/dz must be proportional to the flow rate divided by the cross-section.
Since the total surface involved must be proportional to the volume we get:
F ' BqL

130

(Eq 7.7)

Grout injection of masonry, scientific approach and modeling

The external resultant force is proportional to the pressure difference and the cross-section on
which the pressure is exerted. This cross-section is the overall cross-section multiplied by the
porosity. We obtain
F p ' Pb & Pa f A

(Eq 7.8)

The gravity is the only body force that is acting on the fluid and can be expressed as
Fg ' ? f AL g

(Eq 7.9)

Combining (Eq 7.7), (Eq 7.8) and (Eq 7.9) we can write the equilibrium of forces as
B qL % ? f AL g ' P b & P a f A

q ' &

or

with

K '

KA
Pa & Pb % ? g L
L

(Eq 7.10)

(Eq 7.11)

f
, as a constant characteristic of the porous material.
B

The above derivation of Darcys law could lead one to think that it is only applicable for steady
flow. However, the viscous forces involved in laminar flow through porous media are that
much greater than any inertial forces, that also the inertial forces in the flow direction can be
neglected. For practical purposes, Darcys law is thus also valid for variable rate q. This is
important for the injection of grouts since a steady flow will never happen there.
The laminar flow regime breaks down for sufficiently high flow rates. For high flow rates
Darcys law is not valid. The onset of inertial effects occurs rather gradually in the range of
Reynolds number from one to ten. The Reynolds number definition was already mentioned in
Chapter 5, (Eq 5.2) and is defined as:
q? d
Re '
(Eq 7.12)
Af
where d is a measure for the pore diameter. Since a pore diameter is difficult to measure, the
grain diameter can be used instead or, as an alternative for d, one can use

K
[Collins, 1965],
f

which can more easily be determined.

Chapter 7. Flow of fluids through porous media

131

7.3. Equations governing the flow of fluid through porous materials

7.3.1. Differential form of Darcys Law
Above Darcys Law was interpreted as resulting from equilibrium of the forces acting on the
fluid flowing within a macroscopic sample of porous material. The same law can be written for
an element of volume of length ds and plane cross-section area dA. The equilibrium equation
for the fluid in the elementary volume can then be written as[Collins, 1965]:
& LP %

where

B
vP % ? g iP3
f

f dP
s dA ' 0

iP1

unit vector in the first horizontal direction

iP2

unit vector in the second horizontal direction

iP3

unit vector in the vertical direction, the direction of gravity

(Eq 7.13)

or with K = f / B defined as before

vP ' &

K
K P Mp
Mp
Mp
( LP % iP3 ? g) ' &
% iP2
% iP3
% ?g
i1

Mx1
Mx2
Mx3

(Eq 7.14)

where iP1 , iP2 and iP3 are unit vectors parallel to the respective orthogonal Cartesian axes x1, x2
and x3. This equation is the logical generalization of the linear form given by the Darcy equation
(Eq 7.10).
The integrals of this differential law generally agree well with observations during experiments.
The differential law of flow for incompressible fluids can be expressed in very compact form
by defining a flow potential as:
U ' p % ?gx3
Then the law of flow becomes:
vP ' &

K
LU

(Eq 7.15)

In the above derivations of the different forms of Darcys law we assumed that permeability is
independent of the direction of fluid flow within the medium. This is not generally true for all
porous media. The most general linear relationship between vi and the components of M? /Mxi
that can be postulated takes the form:
vi ' &

132

Ki1

MU
MU
MU
% Ki2
% Ki3
Mx1
Mx2
Mx3

i ' 1, 2, 3

(Eq 7.16)

Grout injection of masonry, scientific approach and modeling

The nine Kij form the elements of a tensor. The three equations from (Eq 7.16) can be written in
matrix form:

v1

K11 K12 K13

v2 ' & ?

v3

K21 K22 K23

K31 K32 K33

MU
Mx1
MU
Mx2

(Eq 7.17)

MU
Mx3

The latter equation will be useful for the finite element formulation of the flow through
anisotropic porous media. In most cases, the K-matrix is a symmetric matrix. The rotation of the
original coordinate system to the so called principal axes of the porous medium produce a
diagonal matrix. (Eq 7.16) simplifies to
vi ' & K i

? MU
Mx i

i ' 1, 2,3

(Eq 7.18)

7.3.2. The differential equations of fluid flow through porous materials

In flow phenomena of any kind, one of the most useful mathematical tools is that obtained from a
conservation principle, that no physical quantity can be destroyed or created. Generally this
law of conservation for any physical quantity can be written as
P % G ' MG
& ?P @ O
Mt

(Eq 7.19)

where G is the quantity that is released per unit volume and per unit time, G is the concentration
of the considered physical quantity.
For the single phase and incompressible fluid, the volume of an element of fluid is not altered by
changes in pressure. Hence, fluid volume is conserved and in the general equation of continuity
P becomes the volumetric flux density vP. For a fully saturated volume, the concentration G
O
becomes the concentration of fluid volume which is just the porosity f .

Chapter 7. Flow of fluids through porous media

133

Therefore, the equation of continuity becomes

?P @ vP '

Mv1
Mx1

Mv2

Mx2

Mv3
Mx3

' G(x1,x2,x3, t)

(Eq 7.20)

If there are no sources or sinks, G / 0. For an isotropic porous medium the components of the
flux density are to be expressed in terms of the components of the potential gradient of equation
(Eq 7.15). Depending on the particular situation, (Eq 7.20) results in other several different
equations. The resulting differential equation for an isotropic porous medium yields
M
Mx1

K MU
Mx1

M
Mx2

K MU
Mx2

M
Mx3

K MU
Mx3

% Q ' n

Mf
Mt

(Eq 7.21)

If the medium is homogeneous and is constant, the above equation reduces for the steady state
to the Laplace equation:
L2 U '

M2 U

M x12

M2 U

M x22

M2 U
M x32

' 0

(Eq 7.22)

Since U ' p % ?gx3, and since g and ? are both constant, this Laplace equation can be further
simplified to
M2p
Mx12

M2p
Mx22

M2p
Mx32

' 0

(Eq 7.23)

For the more general case of a homogeneous anisotropic porous medium, a particular
modification of the coordinate system permits to formulate (Eq 7.20) in a comparable way
[Collins, 1965]. For the particular case of grout injection the masonry is considered to be
inhomogeneous but isotropic. Therefore, the latter case is not worked out in this text.
7.3.3. Boundary conditions
The equations governing the flow of fluids through porous materials are second-order partial
differential equations. It will be necessary to specify the boundary conditions for the dependent
functions or its derivatives.

134

Grout injection of masonry, scientific approach and modeling

C
Closed boundary conditions
At a closed boundary, the fluid velocity normal to the boundary equals zero. Darcys Law gives
v n ' vP @ nP ' &
or simply

K
K M?
L? @ nP ' &
' 0

Ml n

(Eq 7.24)

MU
' 0
Mln

In (Eq 7.24), nP is a unit vector normal to the boundary and ln is the distance measured parallel
to nP.
C
Fluid entry or exit
On any section of boundary through which fluid enters or leaves the porous medium different
conditions may be obtained. For a homogeneous fluid entering from a reservoir at constant
pressure or more generally constant potential, the boundary conditions can be expressed as:
P x1 ,x2 , x3 ,t ' constant

U x1 , x2 ,x3 ,t ' constant

or

(Eq 7.25)

The boundary conditions often applying to boundaries across which flow occurs is the
specification of the velocity normal to the boundary. Thus
vn ' &

K MU
Ml n

(Eq 7.26)

C
Discontinuity in the porous medium
Very frequently, flow occurs in material in which a discontinuity in the permeability exists. The
proper boundary conditions in this case are double. First condition is a consequence of the
pressure. Obviously, since only one value of pressure may exist at any point we can state that
p1 ' p2

(Eq 7.27)

where p1 is the pressure at the boundary location for the first medium and p2 in the second
medium.
Since what enters the boundary from one side must come out the boundary on the other side.
Therefore, the velocities normal to the boundary must be equal on both sides.
K1 M? 1

Mln

'

K2 M? 2

Chapter 7. Flow of fluids through porous media

Ml n

(Eq 7.28)

135

7.3.4. Measurement of the permeability using compressible fluids

In paragraph 7.2.3 it is explained how the permeability of a porous sample can be measured
using an incompressible fluid, or liquid. One could use water for this purpose, but the porous
sample, or in our case, the masonry would be saturated with water. This moisture needs to be
evaporated from the masonry, which takes a very long time. The saturated masonry would
provide a completely different resistance to the grout than the dry masonry. Therefore, we use
air to determine the permeability of the masonry. For air, being a compressible gas, equation
(Eq 7.5) is not valid. The compressibility must be taken into account. For linear flow of an
ideal gas the flow equation becomes
M
Mx

K
Mp
p

Mx

' f

Mp
Mt

In the steady state the pressures are stationary or independent of time and hence

(Eq 7.29)
Mp
' 0.
Mt

Klinkenberg [1994] investigated the fact that gasses do not stick to the pores walls as is
required for Darcys law. Slip occurs and this gives rise to an apparent dependence of
permeability on pressure. This dependence can be expressed by the Klinkenberg relation:
K ' K4

1 %

b
p

(Eq 7.30)

where K4 is the permeability as observed for incompressible liquids: the permeability we want
to know, p is the mean pressure and b is a constant characteristic of both the gas and the poreus
medium.
d
b
K4 p 1 %
dx
p

dp
' 0
dx

(Eq 7.31)

After integration, rearranging and again integration between 0 and length L we get
1 %

K4 p

?p
qp
'
L
A

(Eq 7.32)

Where p is the mean pressure, q the mean discharge and K4 the mean permeability given by:
1
K4

'

1
dx
L m0 K4 (x)

(Eq 7.33)

For this reason, the pressure that is applied in the experimental setup to determine the
permeability of the reproducible masonry samples is kept very small. Furthermore, a
depression is applied. This reduces possible errors for the permeability value.

136

Grout injection of masonry, scientific approach and modeling

7.3.5. Radial flow between concentric cylinders

Steady state, radial flow

12000

10000

Pw

Pressure [Pa]

8000

6000

4000

2000

Pe
Rw

0
0

0.1

Re
0.2

0.3

0.4

0.5

0.6

0.7

Radius [m]

Figure 2: Pressure distribution for a steady radial flow from inner radius Rw = 0,1 m to
external radius Re= 0,6 m
The steady flow of a fluid into a well or out of an injection hole is often represented as plane
radial flow between concentric circular boundaries. In first stage the medium is considered
isotropic. The interior circle represents the wall of the bore hole and the outer circle represents
a boundary of constant potential called action radius. To solve this problem the technique of
conformal mapping is made use of. We use the following transformation of the variables from
the x, y plane, being the original coordinates, to the u, v plane.
u ' ln (x 2 % y 2)1/2
v ' tan&1(y / x)

(Eq 7.34)

This conformal mapping preserves the form of Laplaces equation. So the Laplace equation for
the potential function remains of the form:
M2 U
M2 U
%
' 0
Mu 2
Mv 2

Chapter 7. Flow of fluids through porous media

(Eq 7.35)

137

The boundary conditions are U(u = ln Rw ) = Uw and U(u = ln Re ) = Ue . The solution to the
above equation yields:
U ' A % Bu

(Eq 7.36)

with the boundary conditions corresponding to the u, v plane:

U ' Uw on u ' ln Rw
U ' Ue on u ' ln Rre

(Eq 7.37)

The boundary conditions enable to solve for A en B the above solution in equation (Eq 7.36).
U ' Uw % (Ue & Uw)

u & ln R w
ln re & ln R w

(Eq 7.38)

Or after back transformation we get

U ' Uw %

Ue & Uw
ln R e / R w

ln

x 2 % y2
Rw

(Eq 7.39)

The above solution for the pressure gives rise to a discharge that can be calculated as
Q '

K A(r) ? P
K A(r) dP(r)
.

?r

dr

(Eq 7.40)

Replacing the function P(r) by the solution found in equation (Eq 7.39) gives a discharge
independent from r:
Q '

K 2 p h Pw & Pe

ln(R e/Rw)

(Eq 7.41)

Although equation (Eq 7.41) provides the solution for the steady state situation, it can be read
that the discharge will decrease with growing penetration depth (increasing Re -value) of the
fluid into the porous media. Nevertheless, the discharge will never drop to zero for a limited
external radius for a fluid with constant viscosity, without critical yield strength as explained in
paragraph 6.3.4.
According to equation (Eq 7.39), the discharge Q is a function of the bore hole radius Rw . For
injection purposes this is one of the technological parameters influencing the quality of the
consolidation. The discharge increases almost linearly with increasing bore hole radius
(Figure 7.3). The dimensions of the injection hole will of course be limited, but precautions
should be taken not to use too small bore hole diameter.

138

Grout injection of masonry, scientific approach and modeling

Theoretical discharge vs bore hole radius

0,3

0,2

Discharge [dm/s]

0,25

Values for calculation

0,15

K = 1e-10 m2
Re = 0,6 m
Ue = 0 Pa
Uw = 10 000 Pa
mu = 0.01 Pa s
h = 1m

0,1
0,05
0
0

0,01

0,02

0,03
0,04
bore hole radius [m]

0,05

0,06

Figure 7.3: The discharge increases with increasing bore hole radius

The results in for Figure 7.3 are calculated for an isotropic porous medium, injected with a
Newtonian liquid without critical yield strength. As mentioned above, the rheological behavior
of a grout can not correctly be modeled by a Newtonian fluid. Furthermore, the masonry is far
from being an isotropic porous medium. Still the remark about the bore hole diameter holds.
Actually, the effect will even be amplified by the non-Newtonian nature of a grout. The slower
the grout will flow through the masonry, the more water can be extracted from the grout, the
more both viscosity and yield strength will increase. Besides, the low velocity of the grout flow
will cause the thixotropic behavior to take place sooner. As a consequence, the penetration
depth will decrease with a smaller injection hole.

7.4. Conclusions
Together with the previous chapter about rheology of grouts, this chapter forms the theoretical
base that is used to build the final model. The experimental set up for measuring the
permeability using a compressible fluid is worked out theoretically in paragraph 7.3.4. The

Chapter 7. Flow of fluids through porous media

139

theory about fluid flow through porous media is used to illustrate the influence of the bore hole
diameter on the penetration depth of the grout inside the porous medium.
However, the formulas that are derived in this chapter describe the laminar flow of a Newtonian
fluid through a saturated porous medium. This is not exactly what is happening when a grout
enters the flow channels inside the masonry structure. First of all, grouts are no Newtonian
liquids. Grouts rather show a Bingham behavior. Besides, grouts are a dispersion of the
particles of the binding agent in water. Therefore, they show a more complex behavior when
flowing through the porous medium. The particles might block part of the pores. Sedimentation
phenomena and thixotropy imply that the application of Darcys law for describing the
penetration of the grout inside the masonry will lead to inaccurate results.

140

Grout injection of masonry, scientific approach and modeling

Chapter 8.

Modeling grout flow in masonry

8.1. Introduction
This chapter describes the basic features of the model that was developed to simulate the
penetration of a grout inside the masonry. Firstly, the choice of a discrete model is justified and
it is explained why the continuum approach is not used. After that, the model itself is discussed:
a network of channels, connecting the nodes. The theory governing the flow of fluids is
implemented to define the transient penetration in the masonry. In the next chapter the model
will be used to judge the importance of some parameters of the grout and of the masonry with
regard to the degree of filling that can be achieved.

8.2. Discrete model

8.2.1. Justification
Transport phenomena in porous material are usually described using the continuum approach. A
representative volume of material is considered to be homogeneous and to have the average
values for the relevant properties, e.g. porosity, permeability and density. This approach
disregards the heterogeneity of the materials on the microscopic level with the pores, small
cracks and capillary pores through which the transport is happening on one hand and the solid
material on the other hand. The method uses average values on the macroscopic level defining a
representative volume. When the scale of the representative volume is corresponding to the
phenomena at hand, convergence is reached. This is why capillary absorption and moisture
transport are fairly well modeled using the continuum approach.
The problem of grout penetration in masonry has a different nature. The representative volume
has to become very large. Furthermore, it would take a lot of effort to characterize the
properties of this representative volume with regard to the penetration of the grout through this
volume. The flow of grout is concentrated in the big cracks and openings of the masonry. From
these main vessels, the grout spreads out over the finer fraction of the masonry. The latter is
necessary to obtain a proper consolidation: if only the main cracks are filled, the masonry will
never regain its internal cohesion and its monolithical behavior. Disregarding the physical
reality that the main flow is situated in the big flow channels, will never lead to correct
simulations.
The above reasoning demands a discrete modeling of the injection process. A discrete model
starts from the real flow channels present in masonry: the main cracks with a diameter bigger
than an arbitrarily chosen value (order of magnitude 1 mm). Also, a discrete model will be a
better approximation of reality. Unfortunately, it will be unfeasible to find out about the size
and the position of all these flow channels. The state of knowledge offers no technique that is
able to provide these data.

Chapter 8. Modeling grout flow in masonry

141

8.2.2. A network of discrete flow channels

The masonry is modeled by a three dimensional combination of one dimensional flow channels.
A volume of porous materials is represented by a number of nodes. These nodes are
interconnected by one dimensional channels to form a network. The transition from a one
dimensional pipe to a two or three dimensional network does not require any special
modification.
Strictly spoken, any two nodes can be interconnected, but in most cases only neighborhood
connections make sense. The nodes are idealized nodes: they do not possess any volume and
they do not mean any resistance to the flow.
Default shape of the flow channels is a cylinder, but when the relation between pressure
gradient and discharge is known or can be determined analytically or experimentally, any other
shape can be used. The conductivity of the channel is the required information. The
conductivity is the inverse of the resistance to flow, which is on its turn given by:
1
Pa . s
' R
C
m3

'

? P
Q

(Eq 8.1)

As can be derived, the dimension of the conductivity is m3/(Pa.s). The conductivity is

calculated using the equations that were derived in paragraph 6.3.1. In a similar way the
relation can be derived for the flow between parallel plates, which is suitable for the simulation
of flow through a wide crack.
A volume of porous material that is modeled by nodes and the interconnecting channels has only
a very limited percentage of empty space. The portion of the volume that can be filled by the
grout is much smaller than what is observed in reality. Assume a cube with an edge of 500 mm.
In order not to increase the density too much, the distance between two neighborhood nodes is
50 mm. Between each node and his non-diagonal neighbors (two out of three coordinates are
the same), an interconnection is present. This interconnection is a cylindrical channel, diameter
4 mm. This network consists of 2700 channels. All of them have the same length: 50 mm and
the same diameter. The total volume of the channels equals:
Vol ' 2700 x length x

p d2
' 2700 x 628.3 mm 3 ' 1696460 mm 3
4

(Eq 8.2)

Related to the original volume of the cube this is only 1,4 %. In practice, as a rule of thumb, the
volume to be injected with grout is estimated to be around 20% of the outer volume. This
conclusion proves that the capacity, provided by the volume of the channels only, is far too
small. Increasing the diameter of the flow channels would indeed increase the capacity of the
network, but would provide a conductivity that is too high. Making the network denser by
raising the number of nodes or by raising the number of channels is to be rejected. More nodes
means more intensive calculations. Increasing the number of interconnections is not valid since
only neighborhood connections make sense. To solve the problem of the limited capacity of the
network itself, capacitive elements are added in (some of) the nodes. These capacitive

142

Grout injection of masonry, scientific approach and modeling

elements do not participate in the transport, but take a large amount of grout. When the capacity
is completely filled, it has no further influence on the grout flow, since it does not appear in the
mass balance of that node any longer. Anyhow, there is never a discharge through the capacity
to any of the neighborhood nodes.

Figure 8.1: Conceptual representation of the capacitive elements

The capacitive elements can physically be justified. Weathered masonry eventually contains
large internal holes. When the grout arrives in such an internal hole, this hole is filled before
any further penetration can be realized. Such a hole can be simulated by a capacity with a large
volume and a high conductivity. If the large hole leads to other cracks, a capacitive element can
not be used, in that case the hole has to be simulated by a large channel, providing the same
effect since capacitive elements do not allow any transport to neighborhood nodes.
Finely cracked zones are present in the vicinity of the wide cracks through which te main flow
occurs. These zones take also part of the grout. The penetration of these zones is a slow
process and the grout penetrates further before those zones are filled with grout. Besides, it
might happen that these zones are only partially filled at the end of the injection. Such a zone
does not participate in the global transport and is perfectly simulated by a capacity with a small
volume and a rather small, decreasing conductivity.
Coming back to the numerical example of the masonry cube with edge of 500 mm we can easily
show that a capacity of 23.250 mm3 in each node would bring the total injectable volume to
20 % of the outer volume.

Chapter 8. Modeling grout flow in masonry

143

8.2.3. Boundary Conditions

After the network has been created, the program requires the boundary conditions. A node can
be subject to three different boundary conditions. For a dead end node, no explicit boundary
condition needs to be specified. The program is built in such a way that the incoming discharge
will automatically drop to zero. The two other possible boundary conditions consist essentially
of imposing a constant pressure at that node. This boundary condition is most probable to occur
for nodes situated at the border of the represented volume. A node connected to the injection
pipe will be at constant injection pressure, a node representing a leakage will be at constant
zero pressure since the grout can flow freely out of that node.
It would have been possible to incorporate the boundary condition of constant discharge through
one channel, e.g. the injection tube. This exercise, however, is quite useless since the situation
of constant discharge is never valid in case of grout injection. Only in the beginning of the
injection, it might happen that the capacity of the pump is not able to fulfill the constant pressure
boundary condition. However, since the penetration of the grout inside the masonry is our goal
of study, this short moment of time is only of little interest to us. The injection pressure has to
be limited to prevent that hydrostatic forces split the masonry. For this reason a modern
injection installation is equipped with a three way valve, providing an adjustable constant
injection pressure. This corresponds exactly with the used boundary conditions.
As explained above, the injection pressure has to be limited to avoid that the masonry suffers
from additional damage due to the hydrostatical pressure. Modern injection installations,
discussed in detail in ?, posses an adjustable three way gate valve and give rise to boundary
conditions that are pretty close to the constant pressure condition. There is only a limitation for
the discharge. In the beginning, when the discharge is high, the pressure might be a little lower
than the maximum injection pressure due to insufficient capacity of the pump and pressure losses
in the conduits. Distinction can be made between experiments where one side is at constant
injection pressure and the other side is impenetrable and experiments where some nodes are at
constant ambient pressure. The latter situation arises when leakage occurs. The sealing of a
leakage comes down to changing the boundary conditions for that node into the standard
boundary conditions of impenetrability.
8.2.4. Dealing with water absorption
When the grout enters the relatively dry masonry, the masonry that is surrounding the flow
channels will absorb water from the grout. One would expect that the grout loses water and,
because of that water loss, that the rheological properties of the grout get worse. So, the water
absorption could be incorporated by changing the rheological properties depending on the time
that the grout has spent inside the masonry. However, there are some contra indications for the
above assumption. Translating the fact that the rheological properties of the grout depend on the
time that the grout spend inside the masonry would require a huge book keeping. Besides, at
certain moments, the grout that arrives at a node, comes from different flow channels. The grout

144

Grout injection of masonry, scientific approach and modeling

would then be a mixture of two liquids with different rheological properties since the grout from
one channel probably spent more time inside the masonry than the grout from the other channel.
This would result in a very complex situation. This argument is not a physical one. It just
means that it would be difficult to master that possibility. Fortunately, there is also a good
physical argument not to deal with absorption like that. During the experiments, the grout that
came out of the samples was caught in a measuring jug to check the rheological properties. This
grout has then penetrated the complete sample and it was expected that the rheological
properties would differ from the properties of the injected grout. This seemed not to be the
case. The rheological properties of the grout that leaves the sample do not or hardly differ from
the grout that enters the sample. The above assumption appeared not to be in accordance with
the experimental findings. Therefore, we believe that the water absorption causes cement
particles to stick to the wall of the flow channel as presented in Figure 6.1. The grout that
keeps on flowing however, has the same properties as the grout that is injected. Therefore, it can
be stated that the flow channels narrow by the absorption of water out of the grout and that it is
not true that the rheological properties of the grout change by the water absorption. This
phenomenon is incorporated in the model by narrowing gradually the flow channel once the
grout has entered the channel (Figure 8.2). The narrowing is faster in the beginning than after a
while since the cement particles that stick to the wall form a barrier for further water to be
absorbed.

Figure 8.2: The water absorption out of the grout results in a narrowing of the flow
channel
8.2.5. Special features of the program

Chapter 8. Modeling grout flow in masonry

145

The program allows to build the network in a manual way by defining each nodes coordinates
(x, y and z) and consequently by defining the interconnections and the size of the interconnection.
The standard procedure to place nodes however, is an automated procedure that generates a
regular pattern of nodes with a fix distance in each direction. The possibility exists to alter the
position of a node or to manually add nodes if the regular pattern does not fulfill the
requirements. Not all the interconnections or flow channels have the same diameter. The
program allows a completely manual definition of the flow channels. The input requires the
beginning node, the end node and the diameter of the channel. A second way to define the
interconnection is by specifying up to 9 standard diameter values for each possible
interconnection. Every interconnection is proposed by the program. The users specifies the
predefined diameter by pressing number 1 to 9. Pressing 0 means that the proposed
interconnection will not be created. Finally there is a possibility of random generation of the
diameters. The diameter range can be specified by the user. It is also possible to specify
locally a smaller diameter range to simulate a region that is less permeable than the global
structure.
All the parameters that have to be specified by the user, have a default value. If no value is
determined, the default value will be assigned.

8.3. Structure of the program

Aim of this chapter is to show the global structure of the program and to explain some special
features. This way of clarifying the program is preferred above a complete listing in an
appendix.
The program is, just as most desktop applications nowadays, menu controlled. This makes the
program user friendly. Since the number of users will be limited, this is not a goal as such.
These features make it easier to adapt the parameters value, to save an intermediate situation.
The structure of the menu also provides a good overview of the global structure of the program
showing the different steps that the user has to do in order to obtain a simulation.
References to menu items in this chapter will be written in italic font. Reference to action
buttons, button that call for one or a series of procedures are typed between {action}.

146

Grout injection of masonry, scientific approach and modeling

8.3.1. Menu items

Figure 8.3: Main menu of the program

Figure 8.4: Submenu Network

Network
First menu item deals with the definition of the network of pipes. The sub menu items enables,
just as in a classical finite element program, to define the nodes and the elements in different
ways from manually to fully automated as explained before. The Submenu nodes (Figure 8.5)
contains five different procedures to create, add or remove nodes. Again, this can be done
manually or more or less automated. The subitem elements (Figure 8.6) allows to define the
interconnecting channels of the network. This item is only available if the nodes were already
defined. The elements can be generated randomly, only if a regular pattern of nodes was
defined, semi-automatic or manually. The manual method is the only one that is available when
the network has a non-standard pattern. The properties of the channels of an existing network
can be adapted. Channels can be added or deleted. The routine {display} visualizes the
network on the screen, showing the diameter by varying the line thickness of the channel. To
enable the use of a predefined network or to use the same network for several simulations, it is
possible to save {write} the network to disk or to read it from a file {read}. All relevant
properties of the network are then saved: the nodes with their three coordinates and the
interconnections with the two end nodes, and the diameter of the channel. The capacities are not
saved since they are seen as a parameter rather then as a property of the network.

Chapter 8. Modeling grout flow in masonry

147

Network - nodes
The definition of the coordinates of the nodes can be done manually {manual}: the program
will prompt for the number of nodes and ask for their three coordinates one by one. This is a
long lasting procedure if the network counts many nodes. If the desired pattern of nodes is of a
regular kind, one can use the option {standard}. For each direction x, y and z the distance
between the nodes can be specified. A specific node can be added or replaced {edit}. The
connections to a replaced node remain. If they have to disappear, this can be done in de subitem
elements.

Figure 8.5: Submenu nodes

Network - elements

Figure 8.6: Submenu elements

To define the channels that are connecting the nodes, there are four options in an increasing
degree of automation given by Figure 8.6. A manual definition, the only method in case of a non
standard configuration of the nodes, requires the beginning node, the ending node and the
diameter. The other input methods enable only connections between neighborhood nodes where
the diameter is given by:
- keyboard usage {Semi-automatic}

148

Grout injection of masonry, scientific approach and modeling

- a random choice between several predefined diameters {Automatic}

- random choice by the program between predefined limits {random}

Figure 8.7: Submenu Rekenparameters

Second menu item enables to change the default calculation parameters of the program. The
following list explains the meaning of the parameters and provides the default value of the
parameters.
{Injection pressure}
The boundary condition of constant injection pressure can be specified and the nodes
that are subject to the injection pressure. Default value for the pressure is 1 bar or
100.000 Pa and applied to the first node. Any other non-zero value can be given: one
injection pressure value and then up to ten nodes that are connected to the injection
pressure level.
{Capacity}
Two parameters define the capacity: its volume and the conductivity towards the
capacity. Default there is no capacity present. Choice can be made between constant
capacity volume and conductivity in every node, or variable ones. These capacity
parameters can be generated randomly or manually. The physical meaning of the
capacity is commented before.

Chapter 8. Modeling grout flow in masonry

149

{Leakage}
Just as the injection nodes that are subject to the constant injection pressure, the leakage
nodes are subject to a constant pressure boundary condition. In this case the pressure is
zero, since a leakage is supposed to remain at ambient pressure. The number of leakage
nodes has to be specified and then the node number.
{Viscosity}
The first grout parameter to specify is the viscosity. This rheological parameter can be
determined by laboratory test. Testing the viscosity by either a coaxial viscometer or a
capillary viscometer should be a standard test when adjusting the grout composition.
Default value for the viscosity is 0.005 Pa.s.
{Shear strength}
When shear strength is present in the rheology of the grout, this implies that there exists
non-Newtonian behavior. The shear strength for a mixture with a superplasticizer
should be rather low. It is advised to check the shear strength of a new developed grout.
Default value for the shear strength is 0 Pa. Typical value for the shear strength is about
5 to 10 Pa.
{Density}
The density of the grout is a necessary parameter because the program includes gravity
during the calculation of the penetration. The density is only of little influence since the
gravitational forces are rather small compared to the injection pressure.
{Time step}
In the following paragraph, the different steps in the algorithm will be clarified. The
time step determines the accuracy and the speed. The time step can not be too big for
accuracy reasons and not too small in order to avoid exploding the calculation time.
Default value of the time step is 0,001 second.
{Acceleration}
The small time step is especially necessary for the beginning of the injection. The initial
discharge is very large, but slows down after a few time steps. Therefore, it is possible
to increase gradually the time step from an initial value. The time step will be
multiplied by a factor, somewhat larger then 1, after a chosen number of time steps, e.g.
50 or 100 steps. Without undermining the initial accuracy this feature reduces the
calculation time needed to accomplish a simulation.

150

Grout injection of masonry, scientific approach and modeling

Calculate

Figure 8.8: Submenu Rekenparameters

The next item enables to initialize {initialize}, to start the calculations {start} and to continue
{continue} the calculation if they were stopped by the user for instance to change the time step
or to check for intermediate results. After initializing, all the positions, the pressure values, the
simulated time etc.. are reset to their initial value.
Results

Figure 8.9: Submenu Results

Especially for analyzing the results, the fourth menu item is added. It enables to have a look at
some interesting intermediate or final results. The results are grouped into time related results:
total simulated time, time step at that moment, number of iterations done, and volume related
results: total injected volume, volume of the voids of the network, volume of the channels only,
volume of the capacities only, volume filled with grout. The total injected volume will differ
from the volume filled with grout if there is a leakage. The difference of both provides the
amount of grout that is lost.
The next menu item brings us to the calculations itself. Initialization is required. This action
imposes the defined boundary conditions, initializes the positions of those channels that are in
contact with injection nodes and brings the position in all other channels to zero. Finally it
resets the time step to the initial one, it puts the capacities to zero after a former simulation and
also the factors that take into account the Bingham behavior of the grout. After initializing, the
calculation can start. During the calculation it is possible to halt the calculations for saving, or
viewing intermediate results, to adapt the time step etc... After the adaption the calculations can

Chapter 8. Modeling grout flow in masonry

151

be continued. Finally, it is possible to reset the position in any channel to zero or to fill the
complete network. The latter feature can be useful to calculate the total maximum discharge
through a network where some nodes are at injection pressure and some of the nodes are
leakage nodes.
Output items Screen and Disk

Figure 8.10: Submenu output to screen

Fifth and sixth menu items enable output of results or parameters to the screen (Screen) or to
disk (Disk) for analysis in any other program such as a spreadsheet program or a statistical
program. Among the results or parameters are :
properties of each channel: the diameter, length, ending nodes
discharge through each channel
position of the grout in each channel
Correction factor for a Bingham fluid in each channel
Pressure values for each node
Capacity data for each node
The discharge towards the capacity for each node
Furthermore, it is possible to save the complete result file to disk. The calculations can proceed
afterwards or the situation can be analyzed.

152

Grout injection of masonry, scientific approach and modeling

8.3.2. Calculation algorithm

The global structure clarifies how the requested data input happens. It gives a short overview
of the different menu items and gives an idea of the possibilities of the program. The most
important and critical part of the program is the calculation algorithm. There, the physical rules
are translated into mathematical formulas. There these mathematical formulas are translated
into the Turbo Pascal source. This part discusses the basic features of the program, algorithm
and the mathematical techniques used to achieve the simulations. Some of the mathematical
procedures were found in classical standard books. Tough the author realizes that the main
algorithm is probably not the most powerful one. For sure, it is possible to enhance the
algorithm with regard to calculation speed and memory usage. This enhancement however, is
beyond the goal of the present study. The calculations are formed by a series of steps, repeated
at every time step. Each of these steps is explained.
C

Building the set of equations

The pressure values in every node are the unknowns, except for these nodes where a
boundary condition of constant pressure exists such as the inlet nodes or leakage nodes.
Each equation expresses that the net discharge towards the node is zero. This can be
done for every node that is reached by the grout. The number of unknown pressure
values always equals the number of reached nodes and hence the number of equations.
These nodes are firstly eliminated from the set of equations. Notice that the dimensions
of the set of equations are growing as the grout penetrates further inside the masonry.
For node A, surrounded by four nodes i = 1 to 4, the incoming discharge for each
channel can be expressed as:

QAi ' ( Pi & PA ) CAi met CAi '

tn

where BAi

pD4
1
( BAi
128 Pos t n
Ai

(Eq 8.3)

tn

PosAi
PosAi 4
4 4t
1 4t
)% ( c
) , expressing the influence of
' 1 & ( c
3 D P tn & P tn 3 D P tn & P tn
i
A
i
A

the shear stress in case of a Bingham fluid 6.3.3. If the fluid has no critical shear stress,
BAi equals 1.

The discharge towards the capacity, if present, can be expressed as:

QACap ' & P A CACap where CACap ' the given value

Chapter 8. Modeling grout flow in masonry

(Eq 8.4)

153

Stating that the nett discharge towards node A results in:

j QAi
4

% QACap ' 0

(Eq 8.5)

i'1

In this summation, only the channels that really feed node A are mentioned. This means
that a channel that is not completely filled, will only appear in the summation if that
channel is filled from node A. This way, a channel that is being filled up by a
neighborhood node will not contribute for the equation expressing that the net discharge
for node A equals zero. Channels that are fully filled will always contribute.
A node at constant pressure provides no additional unknown. This constant pressure
value will firstly be eliminated from the equations and appear in the right hand sight of
the set of equations. The set of equations results in a sparse matrix, which bandwidth
equals the degree of connectivity: this is the maximum number of surrounding nodes.
The procedure that builds the set of equations, checks all the nodes that were defined in
the model. If a node is reached by the grout it gets a place in the matrix. Then, the
program expresses for all the nodes that were reached that the nett discharge must be
zero. The conductivity factors C that are related to the unknown pressure values are
placed on the right place in the matrix. By only taking into account the reached nodes,
and hence, by considering only those pressure values related to the reached nodes as
unknowns, all other pressure values will remain zero. This is physically justified if one
accepts that the air in the masonry structure is free to escape from the masonry without
imposing a significant counter pressure. Having in mind the structure of ancient
weathered masonry, this assumption is valid. In a similar way it is assumed that the
capacitive elements are not able to generate a counter pressure. The background for
both these assumptions is the relatively open character of masonry with regard to air.
The transport phenomena that are considered here, take place in relatively big channels,
but the air is able to flow away through the fine capillary pores. In the above
explanation, the gravity was not incorporated. However, gravity can not be neglected,
especially since the height of an injected masonry mass can be relativily high. When
taking into account the gravitational forces, the pressure difference changes. The
discharge through the channel that connects node 1 with node A can be expressed as:
QA1 ' ( P1 & PA % ? g? h) CA1 where CA1 '

pD 4
1
128 Pos t n
A1

(Eq 8.6)

In equation (Eq 8.6), ? h is the difference in height [m] between the reached node and the
position of the grout front in the channel A-1. For a channel that is completely filled this

154

Grout injection of masonry, scientific approach and modeling

corresponds to the relative difference in height between both nodes. Each equation,
expressing the discharge between two nodes that are at different height, contains a
constant term. This term will appear in the right hand side when building the matrix of
equations.
C

Solving the set of equations

Once the matrix is build, the set of equations is solved using the Gauss elimination
method with full pivoting. This method is not the best one available, especially not for
spare matrices, but offers the advantage of flexibility. No limitation have to be made
regarding the interconnections between nodes. Adding a node does not require a
complete reordering of the existing nodes. So, for reasons of general applicability, this
method was chosen. Certainly for the manually build networks with a non-standard node
pattern, this is an important feature. Disadvantage of this method is that it demands a
square matrix. When using a method that uses the spareness of the set of equations, the
amount of memory and the CPU is used more efficiently. The solution of the set of
equations returns the pressure values for all the reached nodes that have not a constant
pressure boundary condition.

Calculating the positions at time tn+1

The positions after the nth time step are known from previous calculations or from the
initialization procedure if n = 0. This initialization procedure sets all positions to zero,
except for the positions in channels that are connected to the nodes at constant injection
pressure. The position of these channels get a small non-zero value. When, in course of
the calculations, a new node is reached, that node will also be lighted. Therefore, the
position of the grout, in case the position is still zero, is increased with a very small
value. This value is calculated by taking the amount of grout that would flow out of the
channel through which the node was reached. This amount is proportionally divided
over the empty channels connected to that node according to their conductivity. Leaving
this small portion of grout out of consideration, small mistakes would be accumulated.
Furthermore, problems would occur due to division by a very small number or division
by zero. Indeed the conductivity values are depending upon the position that results from
the previous calculation step (equations ? to ?. After that, the new position is
determined out of the obtained pressure values, the conductivity and the previous
position as follows. The discharge is calculated using the equation of HagenPoisseuille. Remember, these discharges are such that the nett discharge towards the
node is zero. One could calculate the position at time tn+1 under the assumption that the
discharge does not change during the time step. Then, the position at time tn+1 is given by
xn %1 ' xn %

discharge ( ? t
S

Chapter 8. Modeling grout flow in masonry

(Eq 8.7)

155

S is the cross section of the channel. For small time steps, (Eq 8.7) can be applicable,
but actually the progress of the grout is overestimated each time step, since the discharge
will decrease somewhat because of an increased resistance to flow as the grout
progresses. These small mistakes can finally accumulate to an unacceptable total error.
Therefore, the integration is done analytically and not numerically. Let us assume that
the position tn is xn. The differential formulation of the equation of Hagen-Poisseuille is:
dx
D2 ? P
'
dt
32 x

(Eq 8.8)

Integration between the time limits tn and tn+1 and the limits xn and the unknown value xn+1
provides:
xn%1

tn%1

32 x
dx ' ? p dt
m D2
m
x
t
n

(Eq 8.9)

When all parameters are constant in time and independent from x the integration gives:
?t '

xn2%1 & xn2 16

D2 ? P

(Eq 8.10)

(Eq 8.10) can be rearranged as:

x n %1 '

xn2 % ? t

D2 ( ? p
16

(Eq 8.11)

In the program xn+1 is the new position. For a Bingham fluid a similar expression is
used.

156

Visualization of the new situation

After the calculation of the new positions, the screen is refreshed. When empty, the
channels are displayed in white. Once the channel is reached by the grout, the filled part
is colored red. The coloring happens from the node whit the highest pressure value as it
fills. Together with the coloring of the positions, the pressure is visualized. The virtual
manometers indicate the ratio of the obtained pressure and the injection pressure. This
facilitates the interpretation of the progress of the grout. When no discharge is realized,
the pressure drop over a channel is only the hydrostatical pressure difference. The
capacity is represented by a grey node when not completely filled. Once the capacity is
filled the node colors red. When the complete network is filled, it shows up completely
red.

Grout injection of masonry, scientific approach and modeling

Chapter 9.

Validating and using the model

9.1. Validation of the model

The mathematical model being completely new developed, it requires a thorough testing and
validation of the model before using it for the purpose it is built for. This validation will be
done by simulating situations that can be solved analytically.
9.1.1. Flow of Newtonian fluid through one dimensional cylindrical pipe
Cylindrical pipe with constant diameter
The transient flow of a Newtonian fluid through a one-dimensional pipe with constant diameter
can easily be calculated analytically. The relation time to penetration depth in the pipe was
calculated before and resulted in equation (Eq 6.21) on page 125.
t '

16 x 2
D2 ? p

(Eq 9.1)

The model is used to simulate the above case. Two different models were used. The first
model consists of one pipe, diameter 0.005 m and length 1 m. The second model consists of 50
pipes, 51 nodes with diameter 0.005 m and length 0.02 m. In both cases gravity is neglected and
the inlet pressure is maintained constant at 1000 Pa. For the model with one pipe only, there is
no difference between the analytically calculated results and the results obtained by the model:
there is a perfect match. When the pipe of 1 m is built up with 50 small pipes, 0.02 m long, the
difference between the analytical figures and the model mounts up to 1 %. This difference is
caused by the numerical inaccuracy and by round off errors when initializing a new pipe.
1.2

Position [m]

1
0.8
0.6
Analytic
Model 1 pipe
Model 50 pipes

0.4
0.2
0
0

0.5

1.5
Time [sec]

2.5

Figure 9.1: First validation: laminar flow of a Newtonian fluid through a

cylindrical pipe

158

Grout injection of masonry, scientific approach and modeling

Cylindrical pipe with varying diameter

The analytical solution for the transient flow of a Newtonian fluid through a one dimensional
pipe with varying diameter must be found part by part. The situation is shown in Figure 9.2.
The charge losses due to the sudden change of the diameter are not taken into account. Each part
has a length and diameter specified by Li and Di.

Figure 9.2: Configuration of a pipe with varying diameter, used for validation
Let us suppose that the fluid has reached the mth pipe. The pressure is maintained constant at the
inlet of the one dimensional pipe. This means that the pressure difference between the inlet and
the position of the fluid is constant and equals Pin. Therefore the next equation is valid:
Pin ' ? P '

32 xm
2

Dm

% j

m &1

32 vi Li

i' 1

Di

(Eq 9.2)

Furthermore, the discharge must be the same in each pipe, since there are no sources or sinks in
the closed pipeline. This requirement can be expressed by specifying the relation of the
velocities.
p Di2
4

vi '

p Dm2
4

vm

and thus

vi '

Dm2
Di2

vm

(Eq 9.3)

Combination of (Eq 9.2) and (Eq 9.3) provides

?P '

32 vm

m &1

Dm L i

Dm2

i' 1

Di4

Chapter 9. Validating and using the model

% xm

(Eq 9.4)

159

or

Dm2 ? P

dx
vm '
'
dt

m&1

32

Dm4 L i
Di4

i' 1

(Eq 9.5)

% xm

Integration of equation (Eq 9.5) yields

tx

xm

tm&1

dt '

Dm L i

i' 1

Di4

% x
dx

Dm2 ? P

tx & t m '
m

32

m&1

32

m& 1

Dm

Dm2 ? P

i '1

Di4

(Eq 9.6)

xm %

xm
2

(Eq 9.7)

where tm-1 is the time when the m-1th pipe is completely filled. This time can be calculated
analogously to the above formulas.
For the validation of the model a pipe with a diameter alternating between 0.005m and
0.0025 m is considered. The length of each part equals 0.2 m. The pipe starts with a piece of
0.0025 m. Just as in the case discussed above Figure 9.3 shows the perfect match for the
invasion of the pipe when calculated analytically and the figures obtained by the model.

160

Grout injection of masonry, scientific approach and modeling

1,2

Position [m]

0,8

0,6

Analytic
Model

0,4

0,2

0
0

10

12

14

16

18

Time [sec]

Figure 9.3: Second validation: flow of a Newtonian fluid through cylindrical pipe with varying
diameter

9.2. Flow of Bingham fluid through one dimensional pipe

It has already been mentioned in the previous chapters: a Bingham fluid flows through a
cylindrical pipe according to the Reiner-Buckingham formula [paragraph 6.3.3]:
p D4 ? P
4 4t
L
1 4t L 4
Q '
[1 & ( c
)% ( c
)]
(Eq 9.8)
128 L
3 D ?P 3 D ?P
The above equation indicates that there is no longer a linear relationship between the discharge
and the pressure gradient. The flow stops when the condition of equation (Eq 9.9) is fulfilled.
Physically, equation (Eq 9.9) means that there is no shearing for the fluid. The yield value does
not exceed the critical value tc.
4 tc L
D ?p

' 1

(Eq 9.9)

The third validation for the model checks if the Bingham fluid stops penetrating at the position
?PD
where
' t c . For the calculation we used the values given in Table 9.1
4 L

Chapter 9. Validating and using the model

161

Table 9.1: Parameter values for the calculation of the penetration depth of a Bingham fluid
Property

Value

Property

Value

?P

10.000 Pa

0.002 m

tc

4 Pa

0.003 Pa s

Total length

2m

For the above mentioned situation, the penetration depth equals 1.25 m. At that position the
yield value is reached. No further penetration is possible unless the pressure is increased. The
model was used to simulate the same situation using two different configurations. The first
configuration consists of one pipe as specified in Table 9.1. In the second case the pipe is
modeled using 10 pipes with the same diameter and length of 0.2 m. In first case the grout
stocks at 1.25 m exactly, in the second case the final position is 1.254 m. As can be read from
the model, a pressure gradient is present although there is no discharge. When the final position
is reached the pressure difference over each of the ten pipes equals 1600 Pa. This is exactly the
pressure gradient that is bringing expression (Eq 9.9) to 1. So in case of a Bingham fluid, a
pressure gradient is possible even when there is no discharge. Note that the same situation
happens in reality. Although the pressure is applied, there is no discharge when injection has
stopped. This phenomenon has partly the same cause as the jamming of the fluid in the
cylindrical pipe.
9.2.1. Conceptual validation
The above examples are relatively simple situations for which the progress of the grout can be
calculated analytically. For the second series of examples the penetration depth follows
directly from the correction factor in equation (Eq 9.8) as expressed by equation (Eq 9.9). So
there is no analytical difficulty to calculate the penetration depth. The above simulations only
show that the mathematical formulae are correctly implemented, that the translation to the finite
element code is correct. Still the conceptual approach needs to be checked. Therefore, the
experimental data are used. The experimental program is described in Chapter 5.

There are three parameters that influence the properties of the network, defined in the model.
First property is the variation of the diameter of the flow channels. This kind of networks are
generated randomly by the program. The diameter range can be specified by the user. Second
parameter is the volume of the capacitive elements in the model. For these situations, the voids
are normally spread over the complete volume and hence the capacitive elements are all the
same. For the first series of simulations, their total volume equals the total volume of the grout
injected during the experiments. This means that in each node, there are 90 nodes, a capacitive

162

Grout injection of masonry, scientific approach and modeling

volume of 0.01 liter is added. The total volume, including capacitive elements and channels,
equals 0.907 liter as in the experiments. This volume is somewhat bigger than the 20 % that is
normally encountered in real masonry. The reason for this is that in real masonry complete
zones of healthy and uncracked masonry are present providing a very low porosity. Third
parameter is the conductivity towards the capacitive elements. Its physical meaning is already
discussed before. Two experiments are simulated using the model. Both injection are done on
the low plexiglass cylinders. One of them is filled with the crushed brick fraction between
1 mm and 2 mm. The other is filled with the fraction between 2 mm and 4 mm. The capacitive
elements in both simulations have the same volume, but the conductivity is somewhat smaller in
case of fraction 1-2 mm. Of course also the diameter range for the fraction 1-2 mm is somewhat
smaller than for the fraction 2-4 mm.
Table 9.2: Parameter values for the simulations
Fraction

diameter range

volume of capacity

Conductivity towards
capacity

1 - 2 mm

0 - 3 mm

0.01 liter

2 - 4 mm

0 - 5 mm

0.01 liter

As can be seen from Table 9.2, the simulation fits well with the experimental data. By adapting
the three parameters of the network displayed in Table 9.1 it is possible to obtain a good fitting
between experimental data and the simulations. These results (Figure 9.4) prove that the
conceptual approach is valid, but that the model needs some calibration. This calibration
implies the optimization of the three parameters that are defining the model in order to obtain a
good matching of experimental data and simulation of the model.

Chapter 9. Validating and using the model

163

40
35
30
25
20
15

Fraction 2-4 simulation

Fraction 2-4 experiment
Fraction 1-2 simulation
Fraction 1-2 experiment

10
5
0
0

10

15

20

25

30

35

40

45

50

Figure 9.4: The good fitting of the experimental data with the simulation for two fractions of
crushed bricks proves that the conceptual approach is valid

9.3. Using the model for parameter study

Generally a model is used for two different purposes: a parameter study and the simulation of on
site situations or experiments. The first application provides a better understanding of the
physical process. One can determine the critical situations, and find the limits of both the model
and the real injection. The simulation of theoretical experiments using the model is cheep,
does not require any resources and can be done without any risk for the masonry structure.
Basically it requires only some computer time and the time to analyze the results. When the
model is reliable, those simulations are an excellent alternative for experiments [Van Rickstal,
1999].
9.3.1. Grout parameters
Before mentioning the parameters that will be discussed, two phenomena are commented that
are of major interest in the case of grout injection: thixotropy and water absorption. Thixotropic
behavior as such is not incorporated in the model. Though the effect of thixotropy can be
studied from this study of the critical shear value. Thixotropy means that both viscosity and
yield value will increase after a period of stand still. A longer period of stand still will
enhance the difficulties to restart the flow. A similar reasoning goes for the water absorption
out of the grout. When the grout thickens due to water absorption, the viscosity and the yield
value will increase dramatically. The relation between the viscosity and the concentration of a

164

Grout injection of masonry, scientific approach and modeling

dispersion is known to be of exponential order. Figure 9.5 represents the measurements carried
out during this research program for the standard grout composition with modified water content
used for the experiments.

Viscosity vs W/C ratio

160

Viscositiy [mPas]

140
120
100
80

Own experimental data

60
40
20
0
0.5

0.6

0.7

0.8

0.9

W/C ratio

Figure 9.5: The viscosity is increasing dramatically with decreasing W/C-ratio, with
increasing concentration
Critical shear stress t c
The viscosity can be considered being the most critical for the velocity of flow and for the time
it takes to reach a certain penetration depth. But among all rheological parameters, the threshold
shear value t 0 is probably the most critical with regard to the penetration depth. However, the
yield value is very often not known. Besides, it is quite complicated to measure the yield value
for dispersions with the classical rheological testing devices.
Before discussing the influence of the yield value on the penetration depth of a grout inside a
masonry structure, it is suitable to give some remarks on the difficulties to determine this yield
value. The rheological study of grouts is probably worth spending a complete research project
on it. Rheology of dispersions is a complex matter. The development of a guaranteed correct
way of working falls out of the scope of this research. However, an indication is given of how
acceptable values for the yield stress can be obtained. One might determine the shear stress for
different shear rates with a coaxial viscometer, type Brookfield. This provides a number of
points on the diagram shear stress versus shear rate. By extrapolation one could calculate the
crossing of this curve with the Y-axis. This provides the shear value for the grout.
Dr. E. Toorman showed in his PhD thesis [Toorman, 1992] that this way of working is very
often not valid. Especially in the lower shear rate zone, the shear stress - shear rate curves for
dispersions show a significant non-linear behavior. This way the yield value can be heavily
underestimated using the value obtained by extrapolation.

Chapter 9. Validating and using the model

165

Let us suppose that it is possible to determine the correct yield value. Since this parameter
study is a comparative study the same network is used throughout this study. It consists of a big
entrance pipe connected to a variety of channels representing a masonry cylindrical zone with a
diameter of 2m. The injection pressure is constant and equals 0.5 bar at the entrance of the first
pipe. Other specifications of the grout are listed in Table 9.3. For a specific configuration of
the network, the obtained results are listed.

Figure 9.6: Radial network to quantify the influence of the shear stress of the
grout on the injected volume

Table 9.3:

Rheological properties for studying the yield strength as

parameter

Property

Value

Viscosity

5 mPa.s

Density

1600 kg/m3

Shear stress

= parameter

Injection pressure

100 000 Pa

Taking into account the water absorption as described above by decreasing the channels
diameter, the flow does stop at a certain moment. The model is used to check the influence of
the injection pressure, the critical shear strength and the viscosity on the injected volume.

166

Grout injection of masonry, scientific approach and modeling

Table 9.4: Influence of the shear stress and the injection pressure on the injected volume, taking
into account the effects of water absorption
Pressure
25000 Pa
50000 Pa
100000 Pa

t c = 0 Pa
30,54%
37,17%
50,30%

t c = 5 Pa
14,51%
24,66%
39,95%

t c = 10 Pa
11,38%
21,75%
37,31%

t c = 15 Pa
8,50%
18,88%
34,82%

Table 9.4 shows that a high shear stress has an important influence on the total volume of grout
that is injected. For the given conditions it is possible to fill 30% of the network with a grout
without a shear stress, a pure Newtonian fluid, applying only 25.000 Pa. If the critical shear
stress increases to the normal value found in literature, e.g. 5 Pa, only 14,5 % of the network can
be filled. Increasing the injection pressure solves that problem partly. But since a high
injection pressure might cause additional damage to the masonry, it is preferable to adapt the
grout composition in order to bring the shear stress to an acceptable value. A grout with a shear
stress of 15 Pa requires an injection pressure of 100.000 Pa to inject 1/3 of the volume, whereas
a grout without yield stress requires only one fourth of it: 25.000 Pa.
Viscosity
For one cylindrical channel the influence of the viscosity on the penetration can theoretically be
derived from (Eq 6.11) or (Eq 6.15). The discharge is inversely proportional to the viscosity.
For the model the same is true. Table 9.6 lists the time needed to fill the network displayed in
Figure 9.6.
Table 9.5:

Rheological properties for studying the viscosity as parameter

Parameter

Value

Viscosity

= parameter

Density

1600 kg/m3

Shear Strength

5 Pa

Injection Pressure

50.000 Pa

An increasing viscosity will slow down the flow through a flow channel. But de flow can never
be blocked by an increasing velocity without the presence of a yield value. Mostly a grout will
show a (small) yield value and hence the viscosity will influence the pressure gradient inside
the flow channel. Therefore, when the viscosity increases, the critical shear value will sooner
lead to the blocking of further penetration. Table 9.5 lists the process parameters for this study.

Chapter 9. Validating and using the model

167

Table 9.6: Influence of the viscosity for the injection of the 2D network. The
water absorption effect is not incorporated in this simulation.
Time to fill the network

Inlet pressure = 50.000 Pa

t c = 0 Pa

Viscosity [Pa s]

Time [s]

0,005

112

0,01

225

0,02

452

0,04

905

The pressure loss along a pipe depends strongly on the viscosity.

Grout progress in function of the viscosity

0,12

Injected volume [m]

0,10

0,08

0,06

0,04

0,02

= 0.005
= 0.01
= 0.02
= 0.04

0,00
0

50

100

150

200

250

300

350

400

450

Time [sec]

Figure 9.7: Influence of the viscosity on the progress of the injection of a radial network,
taking into account water absorption
9.3.2. Process parameters
Injection Pressure
The injection pressure is the pressure that is applied at the inlet of the injection hole. For an
ideal injection installation a pressure valve is mounted at the end of the feeding tubes. This
ideal situation is applied for modeling the grout flow inside the network. If in reality the
discharge is very big, the applied pressure might drop due to pressure loss in the feeding tubes.
The latter phenomenon in not taken into account. For the determination of the penetration depth

168

Grout injection of masonry, scientific approach and modeling

this is less important since the penetration will drop at a moment of very low discharge and
hence very low pressure loss in the feeding tubes. Within certain limits the injection pressure
can be changed. The injection pressure, this is not the pressure at the pump, should never
exceed 1.5 bar. If the pressure is too high, the grout, under hydrostatical pressure, might cause
severe damage to the structure. A simple calculation proves this. Imagine a double leaf
masonry structure that is injected. At a certain moment the discharge drops and pressure of the
grout between the outer leaf and the inside climbs to the injection pressure. In this situation, the
hydrostatic pressure applies a load of 150 kN (!) on each square meter. Fortunately the pressure
is applied by a liquid and hence the smallest leakage causes the pressure to drop. But still, if an
electrical pump is applied, caution is needed to prevent further damage.
Apart from the above discussion, the injection pressure is an important parameter to reduce the
injection time, to prevent too much water absorption out of the grout or to achieve a good
penetration in the less permeable zones.
Injection holes
Actually there are three parameters involved about injection holes. The injection hole diameter,
the injection hole depth and the pattern of injection holes. In case of the injection hole pattern,
the discussion will take place without the use of the model. For the depth and the diameter of
the injection hole the model is suitable for giving some indications.
The layout of the injection holes has been discussed in paragraph 4.3.3. From this discussion it
is clear that the closest pattern should be used.
From theoretical point of view, the denser the injection hole pattern, the more homogeneously
the masonry will be injected, the lower the injection pressure can be and the better the final
result. But there are economical constraints that reduce the number of injection holes. First of
all, those holes need to be drilled. Secondly, a switch from one hole to another means extra
work, extra time needed to complete the job. Apart from economical reasons, also technical
reason limit the density of the injection holes. If not sealed during the injection of a hole, the
neighborhood holes act as leakages that prevent the pressure to build up.
The diameter of the injection hole has impact on the injected volume and hence on the filling
rate. Intuitively, this can be understood as follows. If the diameter of the injection hole is large,
the pressure loss will be small, also in the beginning when the discharge is big. This way the
grout is present at the entrance of the cracks at high pressure. If the diameter is small, the
pressure loss will be significant when the discharge is big, as expressed by equation (Eq 6.11)
or (Eq 6.15). A simulation confirmed indeed that the time to inject a fixed amount of grout
increases as the injection hole diameter decreases. The total amount of grout that was injected
dropped from 77,3 liters to 54,6 liters. Figure 9.8, Figure 9.9 and Figure 9.10 compare the
final situation for an injection hole with a diameter of respectively 10, 15 and 30 mm. Figure
9.11 displays the evolution of the injected volume of grout for the different bore hole diameters.

Chapter 9. Validating and using the model

169

Figure 9.8: Injection hole diameter = 10 mm

Figure 9.9 : Injection hole diameter = 15 mm

Figure 9.10: Injection hole diameter = 30 mm

The depth of the injection hole should enable to reach all the major cracks. Since the position
of the major cracks are unknown, it is advisable to drill the holes deep enough. At least to half
of the walls thickness, but preferable to 3/4 of the thickness. The deeper the hole, the more risk
for leakages at the backside of the wall. Nevertheless, these leakages can be avoided by
preparing the back side as is done with the injection side.

170

Grout injection of masonry, scientific approach and modeling

Grout progress
influence of the injection hole diameter
0,090

Injected volume [m ]

0,080
0,070
0,060
0,050
0,040
10
15
20
30

0,030
0,020

mm
mm
mm
mm

0,010
0,000
0

100

200

300

400

500

Time [sec]

Figure 9.11:

Evolution of the injected volume depending on the

injection hole diameter

It is obvious that a hole that reaches only half of the walls thickness, will complicate the filling
of the second half of the wall. Of course this will depend upon the situation such as the
presence of large cracks that extend to the injection hole. On the average the second half of the
wall will be filled less completely. One could carry out a second injection from the second
side, but most of the time that side is not so easy to reach. Besides, the execution of a second
injection campaign means almost doubling the cost and the work. The simulations, where the
major cracks are connected to the injection hole, indicate that, the second half of the wall is
incompletely filled if the injection hole does not enters this zone. Similarly, the total injected
volume is smaller. In case the major cracks do not reach the injection hole, the difference will
even be bigger.

Figure 9.12: Injection hole depth is 2/4

Chapter 9. Validating and using the model

Figure 9.13: Injection hole depth is 3/4

171

Grout progress
influence of the injection hole depth
0.080

Injected volume [m ]

0.070
0.060
0.050

2/4 of thickness

0.040

3/4 of thickness
0.030
0.020
0.010
0.000
0

100

200

300

400

500

Time [sec]

Figure 9.14: The evolution of the injected volume depends on the depth of the injection
hole

9.3.3. Radial network: sealing of a leakage

A last demonstration of the use of the model is the sealing of a leakage. Sealing a leakage in
time is important, not only to prevent the loss of too much grout, but also to enhance the
penetration in the neighborhood zone of the leakage. Since a leakage means that the pressure at
that place drops to zero, no pressure is built up in the neighborhood zone. This implies that this
neighborhood zone will not or only partly be injected. The grout penetrates only slowly through
the channels. The absorption of water will stick the cement particles to the wall, reducing the
flow diameter. If the sealing of the leakage takes place too late, the flow channels that are partly
filled with grout will be so narrow as to prevent complete filling of that zone. Therefore, a fast
sealing is important. The model can be used to prove the above intuitive reasoning. Five
situations are compared: a leakage without sealing, a fast sealing after 30 seconds and a sealing
after 1 minute and after 2 minutes. Also the result in case no sealing is done is displayed. The
evolution of the injection of these five possible attitudes are represented in Figure 9.15. As can
be seen, the immediate sealing provides the best result. Sealing after a longer period of time
causes the loss of some grout and reduces the filled part of the masonry. The longer the time
before sealing the leakage the worse the situation gets. In this case there is almost no difference
between sealing after two minutes and no sealing at all. Table 9.7 provides other numerical data

172

Grout injection of masonry, scientific approach and modeling

about sealing of leakage simulations: the amount of grout that flew away through the leakage and
the injected volume.
Sealing of a leakage
0.060

Injected volume [m 3]

0.050

0.040

no sealing
0.030

sealing after 2'

sealing after 1'

0.020

Sealing after 30 "

Immediate sealing

0.010

0.000
0

50

100

150

200

250

300

350

400

Time [sec]

Figure 9.15: Effect of sealing of a leakage, influence of waiting time on the injected volume

Table 9.7: Numerical results of the sealing a leakage simulations

Time before sealing
[sec]
0,0
30,0
60,0
120,0
-

injected volume
[liter]
52,6
51,1
49,8
47,9
46,5

lost grout
[liter]
0,0
3,1
5,6
8,3
9,7

9.4. Using the model as an engineering tool

Let us assume that reliable information of the inner state of the masonry is available from non
destructive methods. Beside, the inner configuration of the main cracks in the masonry are
unchangable data. Therefore, the have to be considered as the boundary conditions for the
consolidation. The more reliable these data, the better the model will correspond to reality.
Other important input data are the grout properties. The first simulation can be done using the
standard grout with the standard rheological properties and t . The value of the absorption
parameter of the model depends on the water absorption capacity of the masonry and the water

Chapter 9. Validating and using the model

173

retaining capacity and the stability of the grout. The granularity of the cement determines the
smallest channel that is still injectable.

Figure 9.16: The model: tool for engineering a consolidation injection

Once the network is defined, engineering can start. Initial value for the diameter and the depth
of the injection holes are defined, a certain injection pressure is used. The simulation will
provide the necessary information to decide about the possibilities of the combination of
masonry properties, grout properties and process parameters. If the final result of the injection
is satisfying, the same combination can be used for the injection. If the results are not satisfying,
three actions can be taken: using a different (more expensive) grout with better rheological
properties, lowering the water absorption of the masonry by a careful prewetting procedure,
changing the process parameters: applying a denser injection hole pattern or a higher injection
pressure. The influence of these parameters on the injected volume are qualitatively evaluated

174

Grout injection of masonry, scientific approach and modeling

in this chapter. Similar evaluation can be done for the given project. The combination that turns
out to provide the best simulation results, can be applied.
The above engineering approach is schematically presented in Figure 9.16

9.5. Conclusions
The four validation exercises presented in this chapter, prove that the model correctly simulates
the flow of both Newtonian and Bingham fluids through cylindrical flow channels. The
confrontation with the experimental data indicates that, after calibration of the model, the
experiments can be simulated within an acceptable degree of accuracy.
The model incorporates the gradual blocking mechanism caused by the narrowing of the flow
channel due to the sticking of the binding agent particles to the wall of the channel. The model
can be helpful to decide about process parameters. The impact of the injection hole depth, the
injection hole diameter, the injection pressure can be analyzed qualitatively. If the necessary
information to define more accurately the network of flow channels and the capacitive elements
is available, the obtained results can replace test injections. The model becomes a powerful
tool for the complete design of the grouting of the building: composition of the grout, injection
pressure and the injection hole configuration, that can vary locally, can be determined using the
model.

Chapter 9. Validating and using the model

175

176

Grout injection of masonry, scientific approach and modeling

Chapter 10. General conclusions and future research

10.1. General conclusions
The safeguarding of historical monument is considered to be a duty of mankind. By far the
biggest part of human patrimony is built of masonry. Masonry structures suffer from a wide
range of deterioration mechanisms. For structural and mechanical damage, grout injection is
part of the restoration. In the broad field of monument care, grout injection is situated as a
consolidation technique for masonry structures. In this thesis, an overview is given of possible
damaging phenomena. It is indicated in which case grout injection can repair the damage. The
injection technique is portrayed by the description of an ideal injection installation. It is
sketched how grouting can enhance the reliability of the masonry and how should be dealt with
grout injection in a scientific way: from the diagnosis of the masonry over the implementation of
the injection to the quality control after completion of the job is done. Although grouting has
been applied successfully for many times, problems still occur frequently because grout
injection is a complex and multi disciplinary task. The use of wrong materials often causes
problems some time after the restoration. Uniform filling of the voids is an important request,
but very often this seems too hard to achieve.
The experimental program discusses a series of tests, newly developed or adapted to the
peculiarities of the grouting problem, that can be used for a scientific approach of grout
injection, for the diagnosis of the masonry and for the composition of the most suitable grout.
The major achievement of this research however, is the model that enables the simulation of the
grout penetration inside the masonry. This model is based on three pillars: the study of
literature about grout injection of masonry, the theory about the rheology of dispersion and about
flow of fluids through porous materials and finally the experimental program.
Numerical methods are used for the mathematical description of the flow of Newtonian and
Bingham fluids inside channels.
It is shown that the model parameters can be adjusted so that the simulations correspond to the
experiments in the reproducible masonry samples. Several examples are given to show the
possible applications of the model. For successful consolidation injections of masonry, a good
penetration of the grout inside the masonry and a uniform filling of voids is essential. In
practice, expensive test injections are done to see wether or not a specific grout could fulfill the
requirements. The model can reduce the number of test injections and can show the influence of
a different composition of the grout, different rheological parameters, a higher injection
pressure, a more water absorbing masonry. This has economical impact on consolidation
projects and can improve the efficiency of these actions.

Chapter 10. General conclusions and future research

177

The most important achievements can be summarized:

C
The importance of some technological options on the effectiveness of the consolidation
is given: the three way valve, the use of a separate mixing installation in addition to the
collector, the use of an admission system, practical to use.
C
A new test method is developed for testing the stability of the grout. Compared to the
classical methods, the newly developed method provides more detailed information
about the evolution of the stability of the grout in function of time. The test enables to
judge the effect of the composition, for instance the W/C ratio or the amount of
stabilizing admixtures on the stability. The mixing procedure can be adapted in order to
obtain a stable grout. The stability test is relatively easy and economically to do and
should therefore be part of the scientific approach of a grouting project.
C
The development of a reproducible masonry sample creates conditions that are nearer to
the reality of masonry than the sand column test. This way, the injectability test executed
with the reproducible masonry sample, provides information about the water retaining
properties of the grout.
C
The experiments provided a better understanding of the penetration of the grout inside
the masonry. From the observations, we were able to conclude several blocking
mechanisms.
C
The model simulates qualitatively the penetration of the grout inside the masonry. The
influence of the grout properties and the injection pressure on the degree of filling and
the rate of filling is analyzed. General rules can be formulated from the parameter study:
C
Decreasing the critical shear stress of the grout is recommended instead of
increasing the injection pressure
C
Sealing a leakage has to be done as soon as possible. A good preparation of the
masonry reduces the risk for leakages.
C
The injection hole should be in contact with the major cracks.

10.2. Future Research

10.2.1. Information about the masonry
As is was indicated on several occasions in this book, the available non destructive testing
methods are not powerful enough to provide all the information that is required as input data for
the model. Information about the size and position of the major cracks would increase the
reliability of the simulations. Geo-electrical resistivity measurements were successfully used
for the diagnosis of masonry structures.
The same goes for ultrasonic non destructive
testing. Nevertheless, these measuring techniques are not sufficiently accurate to provide
information about the geometry of the major cracks and voids.
Therefore, the further development and a further progress of these techniques or the
development of other non destructive testing methods is a future field of research that could
enhance the possibilities of the model.

178

Grout injection of masonry, scientific approach and modeling

10.2.2. Final goal, online controlled consolidation injection

After the injection is done, the quality of execution can be controlled using non destructive
techniques. If this quality control proves that the injection is poorly done, it can not be undone.
Eventually a second injection campaign can be considered. It is obvious that it would be
interesting to be able to evaluate the degree of filling when the injection is going on. On line
control of the injection makes an immediate reaction possible. If, by any cause, a certain zone is
not injectable through any of the existing drilling holes, on line control enables to observe this
and to drill additional holes in this zone.
By using the simulations of the model, the possible actions can be tried in order to find out about
the best adaptation of process parameters such as injection pressure, grout composition etc...

Figure 10.1: On line control of the injection: a combination of non destructive

monitoring, simulation and adaptation of injection process parameters

10.2.3. Using the model for other purposes

The concept of the model can be used for other fields of research. It is the authors conviction
that it is possible to incorporate capillary forces in the model. This way, the model would
enable to simulate capillary water uptake or the progress of hydrofobic agents in the capillary
pores of building materials. The capillary forces could be incorporated using the same
technique as for incorporating gravity.
A well known example of laminar flow through cylindrical tubes is the flow of blood through
the blood vessels of human being. The contacts with medical researchers in this field, revealed
that there exists no model to simulate the bloodstream in the human body, for instance to quantify
the influence of a constriction of the blood vessels. It is worth to check how the model can be
adapted to this problem.

Chapter 10. General conclusions and future research

179

References
[Acland, 1972]

James H. Acland, Medieval Structure: the Gothic Vault, University

Press, Toronto, 1972

[Ashurst, 1990]

Ashurst J. et Al., Methods of repairing and consolidating stone

buildings, Chapter I, Conservation of building and decorative stone,
Vol2 Butterworth-Heinemann, 1990

[Ashurst, 1989]

Ashurst J, Grouting masonry walls, Stone masonry, Chapter 3, English

Heritage Technical Handbook, Gower Technical Press, 1989

[Atkinson, 1991]

R.H. Atkinson et al., Evaluation of injectable cementitious grouts for

repair and retrofit of masonry, Masonry: design and construction,
problems and repair, 1991

[Bachmann, 1987]

H. Bachmann, W. Ammann, Vibrations in structures induced by man

and machines, Structural Engineering Documents, IABSE, Zrich, 1987

[Baronio, 1992]

G. Baronio et al., Criteria and methods for the optimal choice of

grouts according to the characteristics of masonries, Effectiveness of
injection techniques for retrofitting of stone and brick masonry walls
in seismic areas, International workshop CNR-GNDT, Politecnico di
Milano, march 30,31, Milan Italy 1992

[Bear, 1992]

J. Bear, C.F. Tsang, G. De Marsily, Flow and contaminant transport

in fractured rock, Academic Press, 1992

[Benedetti, 1979]

D. Benedetti, Comparative tests on strengthened stone masonry

buildings, Proceedings of 2nd US National Conference on Earthquake
Engineering, Stanford University, 1979

[Benhamou, 1994]

O. Benhamou, Comportement Rhologique des coulis de liants

hydrauliques ultrafins dstins l'injection, cole nationale
superieure des mines de Paris, 20 may 1994.

[Berlamont, 1989]

J. Berlamont, Hydraulica, uitg. Wouters, 19??

[Berlamont, 1983]

Berlamont J, Van Gemert D., Improvement of soils by injection of

silicates, eight International Harbour Congress, KVIV Antwerp, 1983

180

Grout injection of masonry, scientific approach and modeling

Reference list
[Besson, 1989]

C. Besson, Lvolution des techniques dinjection, Chantiers de

France, Paris, n 222, 1989

[Binda, 1991]L. Binda, et al., Strengthening and durability of decayed brick masonry
repaired by injections, Proceedings 5th North American Masonry conference, vol
VI, pp 839-852, June 3-6, Illinois,1991
[Binda, 1993]L. Binda et Al., Strengthening of masonries by injection technique, 6th North
American Masonry Conference, pp 1-14, Philadelphia, 1993
[Biscontin, 1998]

Biscontin G. et al. Eds, Athens 1998, vol2, pp. 191-200 (ISBN 02578727)

[Blom, 1988]

C. Blom et al., Inleiding in de reologie, Universtieit Twente, Kluwer

technische boeken, 1988

[Bulckens, 1993]

G. Bulckens, P Buyckx, Injectie van metselwerk met hydraulische

grouts, eindwerk VUBrussel, 1993

[Cambefort, 1964]

H. Cambefort, Injections des sols, Editions Eyrolles, Paris, 1964

[Chandra, 1993]

S. Chandra, F. Van Rickstal, D. Van Gemert, Evaluation of cement

grouts for consolidation injection of ancient masonry, Nordic
concrete research 1993

[City of LA] City of Los Angeles, Crack repair of unreinforced masonry walls with grout
injection, rule of general application, 1991
[Colla, 1995]

C. Colla et al.
Investigation of a stone masonry using
electromagnetics, Evaluation and strengthening of existing masonry
structures, Proceedings of the Joint International Workshop RILEM TC
127-MS, pp 163-171,1995

[Collins, 1965]

R.E. Collins, Flow of fluids through porous Materials ,University of

Houston, Reinhold publishing corporation, New York,1965

Chapter 10. General conclusions and future research

181

Reference list
[Constantinides, 1989] G.N. Constantinides, A.C. Payatakes, A three dimensional network
model for consolidated porous media, Basic studies, Chemical
Engineering Communications, 81, pp55-81, 1989
[Coussy, 1995]

O. Coussy, Mechanics of Porous Continua, Wiley & Sons Ltd, 1995

[Descamps, 1997]

F. Descamps, Continuum and discrete modeling of isothermal water

and air transfer in poreus media, PhD thesis KULeuven, 1997

[Domone, 1994]

P.L.J. Domone & S. A. Jefferis, Structural Grouts, Chapman & Hall,

1994

[Dullien, 1979]

F.A.L. Dullien, Porous Media, Fluid transport and pore structure,

Academic Press 1979

[Eurocde 1] Eurocode 1, Basis of design and actions on structuren - Basis of design, Annex A,
Partial factor design, 1994
[Eurode 8[

Eurocode 8, Design provisions for earthquake rsistance of structuren, Part 1-4:

Strengthening and repair of buildings, 1998

[Fatt, 1956]

I. Fatt, The network model of porous media, AIME Petroleum

transactions, 207, pp 144-177, 1956

[Feilden, 1982]

B.N. Feilden, Special techniques of repair and structural

consolidation, Conservation of historic buildings, Chapter 20,
Butterworth-Heinemann Ltd, 1982

[Fitchen, 1961]

John Fitchen, The construction of Gothic cathedrals; a study of

mediaeval vault erection, University Press Chicago, 1961

[Fitchen, 1985]

John Fitchen, Building Construction before mechanization, MIT Press,

Cambridge, Massachusetts - London, 1985

[Gallagher, 1975]

R.H. Gallagher, Finite Element in Fluids, viscous flow and

hydrodynamics, a Wiley - Interscience Publication, 1975

182

Grout injection of masonry, scientific approach and modeling

Reference list
[Ghosh, 1993]

[Gil, 1995]

S.N. Ghosh, S.L. Sarkar, Mineral admixtures in cement and concrete

Vol 4, ABI Books Private Ltd, India, 1993

N. Gil, Grouting as a repair and strengthening technique to historic masonry

buildings, Master Thesis, Lemaire Centre for the conservation of historic towns
and buildings, Leuven, 1995

[Goodwin, 1987]

J. Goodwin, The rheology of colloidal dispersions, Solid/liquid

dispersion, Academic Press, New York, 1987

[Greenkorn, 1983]

R. A. Greenkorn, Flow phenomena in porous media, Purdue University,

West Lafayette, Indiana, Marcel Dekker Inc. New York, 1983.

[Groot, 1993] C. Groot, Effects of water on mortar - brick bond, PhD TU Delft, 1993
[Haelterman, 1993]

K. Haelterman, A Lambrechts, H. Janssen, D. Van Gemert, F. Van

Rickstal, Geoelectrical survey of masonry. RILEM -Materials and
Structures, 1993, S.495-499

[Hinch, 1975]E.J. Hinch, The mechanics of fluid suspensions, Theoretical Rheology, Hutton,
1975
[Janssen, 1993]

H. Janssen, Geo-elektrische controle van consolidatie-injecties in

historisch metselwerk. PhD thesis, K.U. Leuven, 1993

[Klinkenberg, 1941]

L.J. Klinkenberg, Drilling Prod. Proc. ,1941

[Koplik, 1985]

J. Koplik, Two phase flow in random network model of porous media,

Society of Petroleum Engineers Journal, 25, pp 89-100, 1985

[Kubach, 1988]

Erich H. Kubach, Romanesque Architecture, The technical aspect of

church architecture, Faber und Faber, London, 1988

[Lizzi, 1982]

F. Lizzi, The static restoration of a monument. The static restoration

of munuments, Chapter 1, Sagep publisher, 1982

[Mewis, 1976]

J. Mewis, Rheology of concentrated suspensions, Advances in Colloid

and Interface science, 1976

Chapter 10. General conclusions and future research

183

Reference list
[Midoux, 1985]

M. Midoux, Mcanique et rhologie des fluids en genie chimique,

Paris, 1985

[Miltiadou, 1991]

Miltiadou A, tudes des coulis hydrauliques pour la rparation et

renforcement des structuren et des monuments historiques en
maonnerie, LCPC, 1991

[Miltiadou, 1990]

A.E. Miltiadou, Contribution ltude des coulis hydrauliques pour

la rparation et le renforcement des structures et des monuments
historiques en maonnerie, doctorat de lcole Nationale des Ponts et
Chausses, 1990

[Paillre, 1986]

A.M. Paillre, J.J. Serrano, M. Buil, Possibilits offertes par lemploi

dultrafines siliceuses dans les coulis dinjection base de liants
hydrauliques, LCPC 141, 1986

[Paillre, 1993]

A.M. Paillre, J.J. Serrano, A.E. Miltiadou, Formulation de coulis

hydrauliques pour linjection de fines fissures en cavits dans les
structures dgrades en bton et maonnerie, LPCP 186, 1993

[Peirs, 1979]

Giovanni Peirs, Uit klei gebouwd, Baksteenarchitectuur van 1200 tot

1940, Lannoo, Tielt, 1979

[Penelis, 1989]

Penelis G. et al., Grouts for repairing and strengthening old masonry

structures, Proceedings Structural Repair and Maintenance of historical
buildings, pp 179-188, 1989

[Pume, 1989] D. Pume, Strengthening procedures for damaged masonry. Proceedings of 26t h
meeting of CIB-W23, Darmstadt, Germany, 1989
[Reddy, 1992]

J.N. Reddy, The finite element method in heat transfer and fluid
dynamics, CRC press, 1992

[Salvadori,1975]

Mario Salvadori, Structure in architectures, The building of

buildings, Prentice-Hall Inc., Englewood Cliffs, New Jersey, 1975

[Scheidegger, 1974]

A.E. Scheidegger, The Physics of Flow through Porous Media, Univ.

of Toronto Press, Canada, 1974

184

Grout injection of masonry, scientific approach and modeling

Reference list
[Schowalter, 1978]

W.R. Schowalter, Mechanics of non-Newtonian Fluids, Pergamon

Oxford, 1978

[Schueremans, 1999 (2)]

L. Schueremans, R. Hayen, D. Van Gemert, K. Van Balen,

Triaxial Testing of masonry and lime mortars, test setup and
first results, 5th International colloquium MSR99" Materials
Science and Restoration, 30 November - 2 December 1999,
Technische Akademie Esslingen MSR V, Aedificatio Publishers
1999, pp. 527-536

[Schueremans, 1997(2)]

L. Schueremans, D. Van Gemert, J. Van Dijck, A probabilistic

evaluation method for structural masonry, Internationale
Zeitschrift fr Bauinstandsetzen, Vol. 3, No. 6, 1997, pp. 553-567,
Aedificatio Verlag - Fraunhofer IRB Verlag

[Schueremans, 1999 (1)]

L. Schueremans, M. Maes, D. Van Gemert, Evaluating the

reliability of structural masonry elements using the response
surface technique, 8th International Conference on Durability of
Building Materials and Components, May 30 to June 3, 1999,
Vancouver , Canada, Vol. 2, pp. 1331-1342

[Schueremans, 1999(3)]

L. Schueremans, D. Van Gemert, M. Maes, L. Huyse, Stochastic

homogenization technique for unreinforced masonry,
International Congress Numerical modelling of
uncertaintiesValenciennes, France, 18-19 November 1999

[Schueremans, 1997(1)]

L. Schueremans, D. Van Gemert, J. Van Dijck, A probabilistic

evaluation method for structural masonry, International
Conference on Studies in Ancient Structures, Istanbul 14-18 July
1997, (not published)

[Schueremans, 1996]

L. Schueremans, D. Van Gemert, A probabilistic model for reliability

evaluation of historical masonry, 4. Internationales Kolloquium
Werkstoffwissenschaften und Bausanierung, Esslingen, 17-19 Dezember
1996, Proc. Werkstoff-wissenschaften und Bauinstandsetzen, Band II
Mauerwerk, Fachwerk und Tragwerkverhalten, pp. 703-715,
Aedificatio Publ., Fraunhofer IRB Verlag, 1996

Chapter 10. General conclusions and future research

185

Reference list
[Schueremans, 1998]

L. Schueremans, D. Van Gemert, Reliability analysis in structural

masonry engineering, Proceedings IABSE-Colloquium Saving
Buildings in Central and Eastern Europe Berlin, June 4-5, 1998,
IABSE Reports Volume 77, pp. 52-53, CD-ROM ref. nr. 2206

[Schuller, 1995]

M.
Schuller et al., Acoustic tomography for evaluation of
unreinforced masonry, Evaluation and strengthening of existing
masonry structures, Proceedings of the Joint International Workshop
RILEM TC 127-MS, pp 173-182, Padua 1995

[Shing, 1994] Shing P. et al., Evaluation of grout injection techniques for unreinforced
masonry structures, proceeding 5th US International conference on earthquake
engineering, Chicago July 10-14 1994
[Taniguchi, 1996]

T. Taniguchi, E. Fillion, Numerical experiment for 3-dimensional flow

analyses in a fractured rock with poreus matrix, Advances in Water
Resources, Vol 19, pp. 97-107, 1996

[Tomazevic, 1992(2)] M. Tomazevic, The strengthening of stone masonry walls with

grouting, Proceedings Structural Repair and Maintenance of historical
buildings II, pp215-225, Computational mechanics publications, 1992
[Tomazevic, 1992(1)] M. Tomazevic, Laboratory and in situ tests of the efficacy and tying of
stone masonry walls. Effectiveness of injection techniques for
retrofitting of stone and brick masonry walls in seismic areas,
International workshop CNR-GNDT, Politecnico di Milano, march
30,31, Milan Italy 1992
[Tomazevic, 1982]

Tomazevic M, Turnsek V., Verification of the seismic resistance of

masonry buildings, Proceedings of the British Ceramic Society,
Leadbearing Brickwork N 30, Stone on Trent, pp 360-369, 1982

[Toorman, 1992]

E. Toorman, Modelling of fluid mud flow and consolidation, PhD

thesis KULeuven, March 1992

[Toorman, 1995]

E. Toorman, The thixotropic behavior of dense cohesive sediment

suspensions, MAST G8M, Coastal Morphodynamics Research Project,
1995

186

Grout injection of masonry, scientific approach and modeling

Reference list
[Toumbakari, 1999(4)] E. Toumbakari, D. Van Gemert, T.P. Tassios, Mechanical properties
of multi-blend cementitious injection grouts for repair and
strengthening of masonry structures, International Conference
Infrastructure regeneration and rehabilitation. Improving the quality of
life through better construction. A vision for the next millennium.
Sheffield, 28 June - 2 July 1999, Proceedings pp. 767-776, Sheffield
Academic Press
[Toumbakari, 1997]

E. Toumbakari, D. Van Gemert, Lime-pozzolana-cement injection

grouts for the repair and strengthening of three-leaf masonry
structures, Proc. 4th International Symposium on the Conservation of
Monuments in the Mediterranean, Rhodos 6-11 May 1997, Vol. 3, pp.
385-394

[Toumbakari, 1999(1)] E. E Toumbakari et al. Effect of mixing procedure on injectability of

cementitious grouts, Cement and Concrete Research 29, pp 867-872,
1999
[Toumbakari, 1998]

E. Toumbakari, D. Van Gemert, N. Tenoutasse, Injection grouts for

ancient masonry: strength properties and microscopical evidence,
Proceedings of EU DG X Workshop, Athens, December 1998, PACT
56: Compatible Materials for the Protection of European Cultural
Heritage

[Toumbakari, 1999(3)] E. Toumbakari, D. Van Gemert, T.P. Tassios, Multi-blend

cementitious injection grouts for repair and strengthening of masonry
structuren, International Congress Creating with concrete, Dundee 6-10
September 1999, Specialist techniques and materials for Concrete
Construction, Ed. Dhir and Henderson, Thomas Telford, pp. 79-91
[Toumbakari, 1999(2)] E. Toumbakari, D. Van Gemert, T.P. Tassios, N. Tenoutasse, LimePozzolana-Cement Injection Grouts for the Consolidation of threeleaf Masonry Walls, Materials and Structures
[Van Rickstal, 1995(3)]

F. Van Rickstal, D. Van Gemert, Evaluatie van cementgrout en

metselwerk voor consolidatie-injectie, Bouwkroniek, Special
issue Renovation, 20 October 1995, pp. 9-13

Chapter 10. General conclusions and future research

187

Reference list
[Van Rickstal, 1999]

F. Van Rickstal, D. Van Gemert, Modeling grout flow in ancient

masonry, 5th International Colloquium MSR 99" Materials Science
and Restoration, 30 November - 2 December 1999, Technische
Akademie Esslingen, MSR V, Aedificatio Publishers 1999, pp. 10091020

[Van Rickstal, 1995(2)]

[Van Rickstal, 1998]

F. Van Rickstal, D. Van Gemert, S. Chandra, Polymer modified

cement grout for consolidation injection of ancient masonry,
Proceedings VIIIth International ICPIC congress, Ostend 3-5 July
1995, pp. 547-553

F. Van Rickstal, D. Van Gemert, Scientific design of grouting as a

repair and strengthening technique for masonry, Proceedings 8th
ICSCE, Eighth International Colloquium on Structural and geotechnical
Engineering, Ain Shams University, Cairo, 15-18 December 1998, Vol
2, pp.223-233

[Van Rickstal, 1995(1)]

F. Van Rickstal, D. Van Gemert, Evaluation of cement grout and

brick masonry for injection purposes, RILEM TC127MS,
Padova, Italy, 27-28 june 1995

[Van Gemert, 1998]

D. Van Gemert, K. Venderickx, F. Van Rickstal, Geoelektrische

Untersuchungsmethoden zur Hohlraumbestimmung vor und nach der
Anwendung von Injektionsverfahren, Instandsetzen von Mauerwerk, 2.
berarbeitete Auflage, Ed. E. Nil, WTA-Schriftenreihe Heft 6,
Aedificatio Verlag, 1998, pp. 37-50

[Van Gemert, 1997]

D. Van Gemert, L. Schueremans, Structurele injecties van waardevol

metselwerk met minerale of kunststofgrouts, K.VIV Studiedag
Restauratie: Historische materialen en hedendaagse technieken,
Antwerpen, 1997

[Van Gemert, 1998]

D. Van Gemert, E. Toumbakari, L. Schueremans, Strukturelle Injektion

von historischem Mauerwerk mit mineralischen oder polymerischen
Mrteln, Bauhaus Universitt Weimar, Vorlesung Grundlagen der
Bauwerkserhaltung, 1998

188

Grout injection of masonry, scientific approach and modeling

Reference list
[Van Gemert, 1986(3)] D. Van Gemert - M. Vanden Bosch,
Renovatie van
funderingsmassieven in metselwerk d.m.v. injecties ,
K.VIV-Technologisch Instituut - Studiedag Funderingsversterking en
funderings-herstel, Antwerpen, 11 juni 1986
[Van Gemert, 1985]

D. Van Gemert, Materialenkennis - Technologie der bouwmaterialen,

VTK-Leuven, 1985.

[Van Gemert, 1996]

D. Van Gemert, H. Janssens, F. Van Rickstal. Evaluation of electrical

resistivity maps for ancient masonry, RILEM Journal 'Materials and
Structures', vol 29, April 1996, pp 158-163

[Van Gemert, 1984]

D. Van Gemert, Elasticiteitsleer, VTK-Leuven, 1984

[Van Gemert, 1984]

D. Van Gemert, Vernieuwbouw van structuren, VTK-Leuven, 1984

[Van Gemert, 1988(3)] D. Van Gemert, The use of grouting for the consolidation of historic
masonry constructions. Advantages and limitations of the method
Stable-Unstable, Leuven University Press, 1988, S.265-276
[Van Haute, 1969]

Van Haute, Scheikunde van de bouwstoffen, KULeuven, 1969

[Van Gemert, 1986(1)] D. Van Gemert, Consolidation of ancient masonry by means of

injections: data from real scale experiment, Proc. CIB-W23-Wall
Structures, Brussel, 1986
[Van Gemert, 1988(1)] D. Van Gemert, Consolidatie-injekties van historisch metselwerk,
Stadsarcheologie, Bodem en monument in Gent, Jaargang 1988, nr. 1.
[Van Gemert, 1987]

D. Van Gemert, Strengthening by injection. A report and assessment

from Belgium. Building Research and Practice, Journal of CIB, Vol.
15, November-December 1987, pp 359-364.

[Van Gemert, 1989]

D. Van Gemert, Consolidation of tower of St. Mary's Basilica at

Tongeren. Preliminary investigation and Design, STREMA 89 - Int.
Conference on Structural Studies, Repairs and Maintenance of
Historical Buildings, Florence, Computational Mechanics Institute,
Southampton, U.K., pp. 371-383, 1989

Chapter 10. General conclusions and future research

189

Reference list

[Van Gemert, 1990]

D. Van Gemert, K. Van Balen, Long-term creep behavior of grouted

stone masonry. The experience at the church of Our Lady at
Tongeren, Proc. Of the International Technical Conference on Structural
Conservation of Masonry, Rome ICCROM, 1990, pp. 237-245

[Van Gemert, 1988(2)] D. Van Gemert - K. Van Balen, The influence of pointing mortar on the
alteration of brick and masonry, 8th International Brick Masonry
Conference, Dublin, Ireland, 1988
[Van Gemert, 1998]

D. Van Gemert, K. Venderickx, F. Van Rickstal, Geoelektrische

Untersuchungsmethode zur Hohlraumbestimmung vor und nach der
Anwendung von Injektionsverfahren, WTA-Schriftenreihe, Heft 6,
1998

[Van Rickstal, 1996]

F. Van Rickstal, D. Van Gemert, Evaluation of Cement Grout and

Brick Masonry for Injection, Proc. Concrete in the service of mankind,
Dundee 1996, Concrete Repair, Rehabilitation and Protection, Ed. R.K.
Dhir E & FN Spon 1996, pp. 313-322

[Van Rickstal, 1996(2)]

F. Van Rickstal, D. Van Gemert, Flow of cementitious injection

grouts in historical masonry, 4. Internationales Kolloquium
Werkstoffwissenschaften und Bausanierung, Esslingen, 17-19
Dezember 1996, Proc. Werkstoffwissenschaften und
Bauinstandsetzen, Band II Mauerwerk, Fachwerk und
Tragwerkverhalten, pp. 1027-1037, Aedificatio Publ., Fraunhofer
IRB Verlag

[Van Balen, 1998]

K. Van Balen et al.,MDDS, a masonry damage diagnostic system, 1998

[Van Balen, 1991]

K. Van Balen, Karbonatatie van kalkmortel en haar invloed op

historische strukturen, PhD thesis KULeuven, 1991

[Van Gemert, 1986(2)] D. Van Gemert, The use of grouting for the consolidation of historic
masonry constructions - Advantages and limitations of the methode,
International updating course on structural consolidation of ancient
buildings, K.U.Leuven-Icomos, 19-24 may 1986

190

Grout injection of masonry, scientific approach and modeling

Reference list
[Van Gemert, 1988(4)] D. Van Gemert, Van Mechelen, D.: Geophysische kontrole van
konsolidatie-injekties in het Gravensteen te Gent. Tijdschrift der
Openbare Werken van Belgi, 1988, pp. 55-67
[Van der Kooi, 1971] J. Van der Kooi, Moisture transport in cellular concrete roofs, PhD
thesis TU Delft, 1971
[Van Den Berghe, 1994]

N. Van Den Berghe, P. Laga, De Aarde als Fundament, ACCO,

Leuven, 1994

[Venderickx, 1999(1)] K. Venderickx, D. Van Gemert, Calibration of geo-electrical

resistivity maps of historical masonry, 5th International Colloquium
MSR 99" Materials Science and Restoration, 30 November - 2
December 1999, Technische Akademie Esslingen, MSR V, Aedificatio
Publishers 1999, pp. 1095-1110
[Venderickx, 1999(2)] K. Venderickx, D. Van Gemert,
Calibration of Resistivity
Measurements on Stone and Brick Masonry, Proc. Intern. Congress
Structural Faults and Repair, London, 13-15 July 1999
[Venderickx, 1996]

K. Venderickx, D. Van Gemert, Geo-electrical Evaluation of

Historical Masonry for Consolidation Purposes, Proceedings 4. Int.
Colloquium Werkstoff-wissenschaften und Bauinstandsetzen, Band II,
Esslingen 1996, S. 1017-1026.

[Venice Charter]

The Charter of Venice, International Charter for the conservation and

restoration of monument and sites, Venice, may 1964

[Vintzileou, 1995]

E. Vintzileou, P. Tassios, Three leaf masonry strengthened by

injecting cement grout, ASCE, Journal of structural Engineering, Vol
121, no 5, 1995, pp. 848-856

[Vogiatzis, 1989]

S. Vogiatzis, K. Floros, Conservation of the Church of the holy

Apostles in Kalamata, Structural conservation of stone masonry, Athens
1989

[Vonck, 1995]

C. Vonck, J. Pardon, Injectie van metselwerk met hydraulische grouts,

eindwerk KIH De Nayer, Mechelen, 1995

Chapter 10. General conclusions and future research

191

Reference list
[Wilkinson, 1960]

W.L.Wilkinson, Non-Newtonian Fluids, Fluid mechanics, Mixing and

heat transfer, Pergamon Press, 1960

[Zagrocheva, 1988]

E. Zagrocheva , V. Todorov, Restoration of the stone walls of the

Boyar House in Melnik by means of injection, engineering Geology of
Ancient works, pp 969-972, Monuments and Historical Sites, Marinos
& Koukis, 1988 Balkema, Rotterdam

[Zienkeiwics, 1977]

O.C. Zienkiewics, The finite element method in engineering science,

J. Wiley, New York, 1977

192

Grout injection of masonry, scientific approach and modeling