Bi-Directional Load Testing Method

GeoSS Seminar - PILE LOAD TESTS IN SINGAPORE –

DIFFERENT METHODS AND GOOD PRACTICES
Chandrasegaran, Regional Design Manager, Bachy Soletanche Singapore Pte Ltd
31 October 2014
Bi-Directional Load Test Method

Introduction
Principles
Advantages & Disadvantages
Typical bi-directional test results
Application in Singapore Geology
Case histories
Conclusions
Load Test – Definition

• A loading test on a pile is made for the purpose of finding

the settlement to be expected at the estimated working load
or some multiple thereof, determining the ultimate bearing
capacity or checking the structural soundness of pile – CP4
• Preliminary load tests are conducted to verify design
assumptions and tested to 2.5 to 3 times the working load
• Working load tests tested at 1.5 to 2 x WL are necessary to
verify the performance of the piles installed
• Instrumentation is needed to measure load transfer at shaft
or base and to measure the elastic shortening of the piles
• The basic requirement would be to install and test piles with
construction methods similar to whats expected for the
installation of working piles
Chandrasegaran 3 31 October 2014
Load Tests – Definition

• Ultimate load capacity of a pile is the load at which

resistance of the soil becomes fully mobilized.
• Failure load is defined as load at which the settlement
continues without further load increment
• There are nearly 42 methods available to estimate the
ultimate bearing capacity of piles from load tests
• Sample of piles are also tested to working loads x 1.5 to
verify the performance of the working piles where certain
prescribed settlement values are not to be exceeded
• Nearly 100,000 piles installed in Singapore every year
• Nearly 1000 load tests either sacrificial ultimate capacity
tests or working pile tests are performed every year to
assess capacity and performance

Chandrasegaran 4 31 October 2014

Conventional Load Tests – Kentledge method – 4000 tons

Chandrasegaran 5 31 October 2014

Conventional Load Tests – Kentledge with steel – 4000 tons

Chandrasegaran 6 31 October 2014

Conventional Load Tests – Reaction Test

Chandrasegaran 7 31 October 2014

Conventional Load Tests – Reaction Test – 7600 tons

Chandrasegaran 8 31 October 2014

Bi-Directional Load Tests – 4500 tons

Chandrasegaran 9 31 October 2014

Bi-Directional Testing – Origin

• Pioneering research by Dr Jorj Osterberg from North

Western University in the 1980s
• Initially proposed with 400 tons test load on a 12” pipe
pile to assess the bearing capacity in rock socket
• First test with 1000 tons test load was conducted by
Schemertmann and Crapps in 1989 in Florida on a
drilled shaft
• In early 1990s the product was commercialized
• Load cell capacity gradually increased
• The test load has reached a maximum of 270+ MN by
2005

Chandrasegaran 10 31 October 2014

Bi-Directional Testing – What is it?
• Conventional load testing involves applying loads at the top of the
piles using kentledge or reaction piles
• Bi-directional system of load testing involves placing a load cell or
cells placed in the pile either at the base or at some part of the pile
along the shaft. The load cells are hydraulically expanded so the
upper part of the pile reacts against the lower part of the pile and thus
the name bi-directional
• The cells or the jacks placed in the cages are sacrificial and can’t be
reused
• When the pile has sufficiently set, load is applied to the upper and
lower part of the pile by the expanding jack and the frictional and end
bearing capacity can be measured
• Instrumented piles are able to measure the precise movement of the
piles in both directions and a combined plot of shaft friction and base
movement produces a load vs settlement plot similar to top loaded
test
Chandrasegaran 11 31 October 2014
 Bi-Directional Testing – Advanatges

• No external reaction system so it avoids lots of steel girders and Kentledge

loads. Better safety with decreased material handling and no overhead load
• No reaction piles needed
• Transportation of heavy loads completely avoided
• The concrete is subjected to only maximum half of the load so the stresses
are lower than conventional tests
• Its able to test side shear and end bearing separately
• Rock shear capacity tested as we can focus on the rock socket portion
• More likely to be cost effective with higher loads
• Offshore load tests made possible for larger loads
• Testing made possible in projects with restricted access or less space
• Low concrete cut of piles can be tested without extending the piles
• More environment friendly compared to other load test methods as
transportation, setting up and dismantling is minimized
• If capacity from upper portion is short, it can be increased by kentledge at
the top

Chandrasegaran 12 31 October 2014

Bi-Directional Testing – Disadvanatges

• Test load limited by the weaker of end bearing or skin friction

as they need to be balanced and used to test each other
• The test pile has to be pre-selected as the jacks need to be in
the pile before
• Loads can’t be increased during test as jack capacity is
preselected
• Test results need interpretation unlike more straight forward
interpretation with the conventional test data
• The load cell is sacrificial and can’t be reused.
• This system can’t be used for all types of piles, eg, H-Piles
• Top of the pile is not structurally tested
• Top load movement must be calculated – Final plot

Chandrasegaran 13 31 October 2014

Bi-Directional Testing – Principles

Conventional Top Load Test Bi-directional Load Test Bi-directional Load Test
Top Load

P= Qs + Qb P/2= Qs = Qb P/2= QS1 = QS2 +Qb

QS1

Qs Qs P/2

Expanding load cell

P/2

P/2 QS2
Expanding load cell
Qb Qb Qb
P/2

Chandrasegaran 14 31 October 2014

Bi-Directional Testing – Multilevel Jack Tests

• Multi level jack test

is useful when the
side shear exceeds
end bearing values
• Multi level jacks can
provide engineer
with more
information on skin
friction for different
zones
• Allows for various
combination of load
testing for different
zones by properly
controlling the load
cell closure and
opening

Chandrasegaran 15 31 October 2014

Bi-Directional Testing – Flexiblity with Jack Placement to test different
zones

A B C D E F G

A – Most common with jack B – Jack placed slightly C – Concreted with low cut
at the bottom and end above base to increase the off so no overburden and
bearing and skin friction bottom half load with some side shear and end bearing
are equal base resistance and some tested first and test further
skin friction by concreting the length G – Multi jack tests with
above if necessary loads applied to base first
by pressurizing the lower
D – Base resistance E – Load test on low cut off F – 2 different layers of jack and later upper jack by
increased by providing bell piles stopping at basement skin friction to be tested pressurizing the
at the bottom level with lengths below and depressurizing the jacks in
above the lines concreted certain sequence
separately and tested

Chandrasegaran 16 31 October 2014

 Bi-Directional Testing – Jack arrangement

Jack Schematic
diagram –
Original – Pipe
Modified Jack
in the middle
Schematic
diagram – Hoses
replace pipes to
apply hydraulic
pressure

Chandrasegaran 17 31 October 2014

Bi-Directional Testing – Instrumentation

LVWDTs, Load cell

Instrumentation –
Strain transducers

Chandrasegaran 18 31 October 2014

Bi-Directional Testing – Typical idealised test resutls

Top movement
– Side shear
friction –
Increased
movement with
little change in
load

Bottom
movement –
End bearing
– Large
movement
starting to
happen

Chandrasegaran 19 31 October 2014

Bi-Directional Testing – Interpretation of the results – Construction of
equivalent Top load vs Settlement curves from tests

Top and bottom movements have been combined in to a single top load vs settlement curve

Chandrasegaran 20 31 October 2014

Top load settlement curves - Assumptions

• Pile assumed as a rigid body which means its incompressible.

Corrections will be applied later when these curves are constructed
• Ultimate side shear in upward direction is equal to the side shear in
downward direction
• The end bearing load movement curve in a conventional top loaded pile
has same loads for a given pile movement under the load as the bi-
directional test when the load cell is placed at the bottom (need to
minus off the pile self weight above the load cell)

Chandrasegaran 21 31 October 2014

Equivalent Top load vs Settlement curves – How to make these plots

Top movement – Side

shear friction

Maximum net load from

measured values

Bottom movement – End

bearing

Chandrasegaran 22 31 October 2014

Equivalent Top load vs Settlement curves – How to construct them

• Select loads from bottom and top curves for same magnitude of movement, e.g. point 4
from top curve and point 4 from bottom curve represent same movement
• Plot the summation of loads against the movement to create the above graph, i.e. Point
4 represents 10mm movement and the corresponding load refers to summation of load
from top (skin friction) and bottom (base resistance)
• Apply correction for elastic compression
• Extrapolate results where appropriate

Chandrasegaran 23 31 October 2014

Equivalent Top load vs Settlement curves – Equations for computing
elastic compression for Bi-directional and Top loaded Tests

Chandrasegaran 24 31 October 2014

Equivalent Top load vs Settlement curves – Correction for elastic
compression

Chandrasegaran 25 31 October 2014

Typical Bi-directional Equivalent Load vs Settlement – Typical plots

Top – Virtually no
movement indicating ski
friction not mobilized

Bottom – Ultimate end

bearing fully mobilized

Chandrasegaran 26 31 October 2014

Typical Bi-directional Equivalent Load vs Settlement – Typical plots

Top – Ultimate load occurs

in side shear – Skin friction
fully mobilized

Bottom – Ultimate end

bearing not fully
mobilized

Chandrasegaran 27 31 October 2014

Typical Bi-directional – Typical Equivalent Load vs Settlement plots –
Ideal results

Top –Skin friction fully

mobilized

Bottom – Ultimate end

bearing mobilized

Chandrasegaran 28 31 October 2014

Typical Bi-directional Equivalent Load vs Settlement plots – Initial large
base movement indicating soft toe

Top –movement
indicating skin
friction mobilized

Bottom – Large
movement indicating
soft toe

Chandrasegaran 29 31 October 2014

Typical Bi-directional Equivalent Load vs Settlement plots – Initial large
base movement indicating soft toe

Bottom – Large
movement indicating
soft toe

Chandrasegaran 30 31 October 2014

Typical Bi-directional Equivalent Load vs Settlement plots – Effect of
poor base cleaning in weathered rock

Bottom – Large
movement indicating
soft toe before end
bearing is mobilized

Chandrasegaran 31 31 October 2014

Typical Bi-directional Equivalent Load vs Settlement plots – Effect of
poor base cleaning in sand and gravel

Good end
bearing contact

Bottom – Large
movement indicating
soft toe before end
bearing is mobilized

Chandrasegaran 32 31 October 2014

Typical Bi-directional Tests - % of measured vs estimated capacity

Chandrasegaran 33 31 October 2014

Typical Bi-Directional application in Singapore Geology -
Jurong Formation

Barrette 2.8m x 1.0m @ 7.5 MPa – Working load

= 21000 kN
Test Load = 2.5 x WL = 52000 kN
Skin friction = 2.5 N kPa max 250 kPa
End bearing 60N not exceeding 6000 kPa

Option 1 – Load cell at the base

N=65, fs= 163 End bearing = 16800 kN - Qb
kPa, depth 9m Skin friction = 39650 kN – Qs
Qs >>> Qb – need to raise the Load cell elevation
9m (N63) + 9m (N>100)

Option 2 – Load cell raised above base by 6m

End bearing = 16800 kN - Qb
N=100, fs= 250 Skin friction - 1= 28250 kN – Qs1 (9m (N63)+ 9m (N>100)
kPa, depth 15m Skin friction - 2= 11400 kN – Qs2 (6m (N>100)
Qs1 = Qs2 + Qb – Load cell elevation
6m(N>100)
Load cell Note : Take note the concrete stress is only
N>100, qb= 6000 kPa slightly higher than 7.5 MPa at full test load of
26000 kN in each direction

Chandrasegaran 34 31 October 2014

Typical Bi-Directional application in Singapore Geology –
Bukit Timah Granite Formation
Barrette 2.8m x 1.0m @ 7.5 MPa – Working load = 21000 kN
Test Load = 2.5 x WL = 52000 kN
Skin friction = 2.5 N kPa max 250 kPa
End bearing 60N not exceeding 10000 kPa in Grade III

Load calculation – Load cell at the base

End bearing = 28000 kN - Qb
Skin friction = 30400 kN – Qs (counting only N>100)
Qs ~~ Qb – The values are nearly the same so both
skin friction and end bearing can be fully tested –
N=100, fs= 250 limit the load to 52500 kN (26250 kN in each
kPa, depth 16m direction)
Caution – but the if base in not clean and is with soft
toe, then this test be a failure as we can not test
both end bearing and skin friction

But if measures are taken and QC procedures are

put in place plus base grouting implemented then
Load cell end bearing can be fully mobilized
Grade III
or Grade II N>100, qb= 15000 Note : Take note the concrete stress is only
Granite
slightly higher than 7.5 MPa at full test load of
kPa
26000 kN in each direction
Chandrasegaran 35 31 October 2014
Typical Bi-Directional application in Singapore Geology –
Old Alluvium Formation

Barrette 2.8m x 1.0m @ 7.5 MPa – Working load = 21000 kN

Test Load = 2.5 x WL = 52000 kN
Skin friction = 2.5 N kPa max 250 kPa
End bearing 60N not exceeding 6000 kPa

Option 1 – Load cell at the base

End bearing = 16800 kN - Qb
Skin friction = 28500 kN – Qs
Qs >>> Qb – need to raise the Load cell elevation

Option 2 – Load cell raised above base by 3m

End bearing = 16800 kN - Qb
Skin friction - 1= 26600 kN – Qs1 (14m (N>100)
Skin friction - 2= 9500 kN – Qs2 (5m (N>100)
Qs1 = Qs2 + Qb – Load cell elevation
14m (N>100)
N=100, fs= 250 Note : Take note the concrete stress is only
kPa, depth 15m slightly higher than 7.5 MPa at full test load of
5m(N>100)
Load cell 26000 kN in each direction
N>100, qb= 6000 kPa
Chandrasegaran 36 31 October 2014
Test #01 – 1.0m x 2.4m Barrette - Singapore Jurong Formation, 1997

Barrette 1.0m x 2.4m Ultimate load test WL = 1800 tons

Test load 2 x WL = 3600 tons with 1800 tons applied on
each direction. Load cell placed at the bottom
At 1575 tons, the base showed large movement and thus
not fully testing the end bearing and skin friction

Chandrasegaran 37 31 October 2014

Test #01 – 1.0m x 2.4m Barrette - Singapore Jurong Formation, 1997 –
Barrette details

Chandrasegaran 38 31 October 2014

Test #01 – 1.0m x 2.4m Barrette - Singapore Jurong Formation, 1997
Schematic setup

Chandrasegaran 39 31 October 2014

Test #01 – 1.0m x 2.4m Barrette - Singapore Jurong Formation, 1997
Plots of test results

Chandrasegaran 40 31 October 2014

Test #01 – 1.0m x 2.4m Barrette - Singapore Jurong Formation, 1997
Plots of test results

Chandrasegaran 41 31 October 2014

Test #01 – 1.0m x 2.4m Barrette - Singapore Jurong Formation, 1997
Plots of test results

1575 tons
Large movement
from this point –
Soft toe? Or
inadequate base
resistance?

Chandrasegaran 42 31 October 2014

Test #01 – 1.0m x 2.4m Barrette - Singapore Jurong Formation, 1997
Plots of test results – Skin friction from extrapolated results

Extrapolation from
1575 tons as end
bearing is fully
mobilized at the
point

Chandrasegaran 43 31 October 2014

Test #01 – 1.0m x 2.4m Barrette - Singapore Jurong Formation, 1997
Plots of test results – Base resistance curve

Chandrasegaran 44 31 October 2014

Test #01 – 1.0m x 2.4m Barrette - Singapore Jurong Formation, 1997
Equivalent top load vs settlement curve

Maximum
measured
load = 3150
tons vs test
load of 3600
tons

Chandrasegaran 45 31 October 2014

Test #01 – 1.0m x 2.4m Barrette - Singapore Jurong Formation, 1997
Comments and discussion
• The results show the base resistance from the bottom plate is less
than the skin friction from the top plate, resulting not realising the
ultimate values of both skin friction and base resistance from the test
• Almost 88% of the test load was mobilised and the results were
extrapolated allowing reasonable outcome from the ultimate tests
• Perhaps raising the load cell by few metres by adding another few
metres of rock socket would have resulted in enough capacity
satisfying the test intent
• Is it lack of base resistance capacity or soft toe or bad concrete or
combination of any? It did hold up for almost 43mm movement
Lessons learned
• Improve load cell setup especially at the base to improve load
transfer mechanism
• Base cleaning procedures to be improved
• Explore possibility of other combinations of top and bottom loads by
placing the load cell at appropriate elevation
Chandrasegaran 46 31 October 2014
Test #02 – Hong Kong on a barrette in the late 90s

Chandrasegaran 31 October 2014

Test #02 – Hong Kong on a barrette in the late 90s

Chandrasegaran 31 October 2014

Test #02 – Hong Kong on a barrette in the late 90s

35,000
Bottom Plate

30,000
Nominal Load (KN)

25,000

20,000

15,000

10,000

5,000
Maximum Test Load:
31,226 KN
0
0 10 20 30 40 50 60 70 80
Displacement (mm)
Chandrasegaran 31 October 2014
Test #02 – Hong Kong on a barrette in the late 90s

35,000
Top Plate

30,000
Nominal Load (KN)

25,000

20,000

15,000

10,000
Maximum Test Load:
5,000 31,226 KN

0
0 10 20 30 40 50 60 70 80
Displacement (mm)
Chandrasegaran 31 October 2014
Test #02 – Hong Kong on a barrette in the late 90s

Comments and Observation

• Upper portion (friction only) better than lower portion

(friction + end bearing)
• Is it because of concrete trapped below the load cell?

Lesson learned

• Improve the installation procedure

• Improve load cell arrangement

Chandrasegaran 31 October 2014

Test #03 - May 2014 in Macau – Shaft grouted barrette

• Excavation to Founding Level

85mSL
• Steel cages fabrication
• Install load test’s instruments
(2no. of O-cells, 4 nos. of LVWDTs, 96 nos.
of strain gauges and 10 rod extensometers
in different level) attached to the steel
cages
• Steel cages installation
• Barrette concreting
• Shaft grouting
• Set up and carry out loading test

Chandrasegaran 52 31 October 2014

Bi-Directional Testing , May 2014 in Macau – Shaft grouted
barrette

Chandrasegaran 53 31 October 2014

Bi-Directional Testing , May 2014 in Macau –
Instruementation layout

Chandrasegaran 54 31 October 2014

Bi-Directional Testing , May 2014 in Macau – Top and
Bottom plate movement vsGross Load

Chandrasegaran 55 31 October 2014

Bi-Directional Testing , May 2014 in Macau
– Load vs Time plot

Chandrasegaran 56 31 October 2014

Strain Gauge load distribution

Chandrasegaran 57 31 October 2014

Top Load vs Settlement plot

Chandrasegaran 58 31 October 2014

Top Load vs Settlement plot

Chandrasegaran 59 31 October 2014

Shaft-grouted Barrette Construction & Load Test’s Instrument Installation

Barrette excavation by grab

Barrette excavation by hydrofraise

Chandrasegaran 60 31 October 2014

Shaft-grouted Barrette Construction & Load Test’s Instrument Installation

Steel cages fabrication

Attach Osterberg cells on the steel plates and

steel cage

Chandrasegaran 61 31 October 2014

Shaft-grouted Barrette Construction & Load Test’s Instrument Installation

Install LVWDTs, strain gauges and rod extensometers

Chandrasegaran 62 31 October 2014

Shaft-grouted Barrette Construction & Load Test’s Instrument Installation

Connecting pressure hoses to the O-cells

Tremie pipe funnel at the steel plates and O-

cells area

Chandrasegaran 63 31 October 2014

Shaft-grouted Barrette Construction & Load Test’s Instrument Installation

Installation of steel cages

Chandrasegaran 64 31 October 2014

Shaft-grouted Barrette Construction & Load Test’s Instrument Installation

Connecting load test’s instruments during steel cages installation

Chandrasegaran 65 31 October 2014

Shaft-grouted Barrette Construction & Load Test’s Instrument Installation

Install steel cage with O-cells

Chandrasegaran 66 31 October 2014

Shaft-grouted Barrette Construction & Load Test’s Instrument Installation

Cutting the temp. support between

2 steel plates

Final condition of O-cells before

installation

Chandrasegaran 67 31 October 2014

Shaft-grouted Barrette Construction & Load Test’s Instrument Installation

Barrette concreting

Strain gauges reading final

checking after concreting

Chandrasegaran 68 31 October 2014

Shaft-grouted Barrette Construction & Load Test’s Instrument Installation

Cracking for TAM pipes Shaft grouting for Barrette

Chandrasegaran 69 31 October 2014

Loading Test (using Osterberg Cell Test Method)

Reference beam preparation

Chandrasegaran 70 31 October 2014

C. Loading Test (using Osterberg Cell Test Method)

Shelter installation to prevent

unstable weather

Chandrasegaran 71 31 October 2014

Connecting all the cables

Connecting load test’s

instrument to the computer
system and reference beam

Chandrasegaran 72 31 October 2014

Data logger and system setup

Data logger for strain Electronic Leica system for

gauges monitoring the pile head movement

Chandrasegaran 73 31 October 2014

Pumps and hoses for laoding -

Hydraulic pump using water

and compressed air for loading Pressure releaser for unloading
(O-cells opening)

Chandrasegaran 74 31 October 2014

Loading Test - Set up

General view of the set up

before load test

Instrument orientation and

serial no.

Chandrasegaran 75 31 October 2014

Test in progress

Agreed loading schedule Load test reading shown on

computer system

Chandrasegaran 76 31 October 2014

Plots as the test progresses

Relative graph during load test

Chandrasegaran 77 31 October 2014

Survey Checks

Survey checking the movement of

2 x working load = 50MN
reference beam during loading and
unloading

Chandrasegaran 78 31 October 2014

Final record

Chandrasegaran 79 31 October 2014

Conclusions

• Bi-Directional load test method is a viable alternative

method for ultimate load test in certain geological conditions
• It can be deployed in both bored piles and barrettes
• Testing to higher loads is made possible thanks to
increased load cell capacity over the years
• Quality control in construction and load cell installation
procedures plays a vital role in ensuring its success rate
• It’s a procedure that improves productivity with less reliance
on overhead loads avoiding expensive handling and
transportation
• It gives flexibility to test end bearing, skin friction separately
and allows the rock shaft friction to be tested effectively
• The pile needs to be pre selected and loads can’t be varied
during testing
Chandrasegaran 80 31 October 2014
Email: bss@soletanche-bachy.com (or) chandran@soletanche-bachy.com

Chandrasegaran 81 31 October 2014