TTI-2-8-75-207-1

TEXAS
TRANSPORTATION
INSTITUTE

STATE DEPARTMENT
OF HIGHWAYS AND
PUBLIC TRANSPORTATION

COOPERATIVE
. = RESEARCH

DETERMINING STIFFNESS COEFFICIENTS AND

ELASTIC MODULI OF PAVEMENT MATERIALS
FROM DYNAMIC DEFLECTIONS
in cooperation with the
Department of Transportation
Federal Highway Administration

,=
:=
RESEARCH REPORT 207-1
STUDY 2-8-75-207
FLEXIBLE PAVE.MENT EVALUATION

li
 TECHNICAL REPORT STANDARD TITLE PAGE
1. Report No. 2. Government Accession No. 3. Recipient's Catalog No.

TTl -2-8-75-207-1
4.Title and Su:-bt:--.itl-e---------1'--------------~--~S~.~R-ep-or·-,-Do-te----------~
I
DETERMINING STIFFNESS COEFFICIENtS AND ELASTIC MODULI t-;-:-N:.:-ov-7e:..:..:.:.m:::.;:be:.:.:r~.i~:;..:.g7-=-6~~----l1
OF PAVEMENT MATERIALS FROM DYNAMIC DEFLECTIONS 6. Performing organization code

7. Author's) 8. Performing Organi lotion Report No.

C. H. Michalak, D. Y. Lu, and G. W. Turman Research Report 207-1

9. Performing Organization Name and Address 10. Work Unit No.

Texas Transportation Institute

Texas A&M University 11. ContractorGrontNo.
College Station, Texas 77843 Study 2-8-75-207
~~~~~-~~-~~~-~~---~~~---~~ 13. Type of Report and Period Covered
12. Sponsoring Agency Name and Address
Texas State Department of Highways Intermin- September, 1974
and Public Transportation November, 1976
Transportation Planning Division 14. Sponsorir1g Agency Code
P. 0. Box 5051; Austin, Texas 78763 !
15. Supplementary Notes
Reseanch performed in cooperation with DOT, FHWA. i
Research Study Title: Flexible Pavement Evaluation and Rehabilitation
I
16. Abstract

The several methods of computing stiffness coefficients or elastic moduli

of materia 1s to be used in computerized pavement design p.rocedures and in other
research activities requiring the knowledge of these values are presented in
this report. The methods include computer codes for calculating stiffness
coefficients and elastic moduli of simple two-layer pavement structures based
on the observed surface pavement deflections, a graphical technique of ob-
tainipg elastic moduli of simple two-layer pavement structures and two recent-
ly de~eloped computer codes for calculating stiffness coefficients of multi-
layer pavement structures.
The method of solution and the basic equations of each method are pre-
sented to assist the prospective user in determining which method best meets
his needs.

17. Key Words 18. Distribution Statement

Computer program, dynamic deflection, No Restrictions. This document is
elastic modulus, pavement design, available to the public through the
stiffness coefficients, surface National Technical Information Service,
deflections. Springfield, Virqinia 22161
19. Security Classif. (of this report) 20. Security Clauif. (of this page) 21· No. of Pages 22, Priee ---~
Unclassified
._F_o-rm_D__
Unclassified
O_T_F-17_0_0-.7-,-~.-6-9-,---------'----------------"'-------'--··--,~·-
111
. .-J
DETERMINING STIFFNESS COEFFICIENTS AND ELASTIC MODULI OF
PAVEMENT MATERIALS FROM DYNAMIC DEFLECTIONS

by

c •H. r1; c ha 1ak

D. Y. Lu
G.W. Turman
Research Report Number 207-1

Flexible Pavement Evaluation and Rehabilitation

Research Project 2-8-75-207

con due ted for

The Texas State Department of Highways and

Public Transportation

in cooperation with the

U.S. Department of Transportation
Federal Highway Administration

by the

Texas Transportation Institute

Texas A&M University

November 1976
 PREFACE

This report is the first of a series issued under Research Study 2-8-
75-207, "Flexible Pavement Evaluation and Rehabilitation ... This study is
being conducted by principal investigators and their staffs of the Texas
Transportation Institute as part of the cooperative research program with
the Texas State Department of Highways and Public Transportation and the
Department of Transportation, Federal Highway Administration.

DISCLAIMER

The contents of this report reflect the views of the authors who are
responsible for the facts and the accuracy of the data presented herein.
The contents do not necessarily reflect the official views or policies of
the Federal Highway Administration. This report does not constitute a
standard, specification or regulation.

ACKNOWLEDGMENT

The cooperation and assistance of many individuals in the Texas

Transportation Institute given during the preparation of this report is
appreciated. Dr. R.L. Lytton's assistance was especially helpful in the
final preparation of the report.

ii
 LIST OF REPORTS

Report No. 207-1, "Determining Stiffness Coefficients and Elastic Moduli of

Pavement Materials from Dynamic Deflections", by C.H. Michalak, D.Y. Lu,
and G.W. Turman, is a summary in one document of the various methods of
calculating in situ stiffness coefficients and elastic moduli in simple
two-l4yer and multi-layer pavement structures using surface pavement
deflections.

iii
 ABSTRACT

The several methods of computing stiffness coefficients or elastic

moduli of materials to be used in computerized pavement design procedures
and in other research activities requiring the knowledge of these values
are presented in this report. The methods include computer codes for
calculating stiffness coefficients and elastic moduli of simple two-layer
pavement structures based on the observed surface pavement deflections, a
graphical technique of obtaining elastic moduli of simple two-layer
pavement structures and two recently developed computer codes for
calculating stiffness coefficients of mu1ti-layer pavement structures.
The method of solution and the basic equations of each method are
presented to assist the prospective user in determining which method best
meets his needs.

Key Words: Computer program, dynamic deflection, elastic modulus, pavement

design, stiffness coefficients, surface deflections.

iv
 SUMMARY

Purpose
The principal purpose of this report is to present in one body the
several methods of calculating in situ stiffness coefficients and elastic
moduli of pavement materials in simple two-layer pavement structures from
observed surface deflections. Also presented are two methods recently
developed for calculating stiffness coefficients of multi-layer pavement
structures from observed surface deflections.
All of the methods of calculating stiffness coefficients or elastic
moduli are based on finding the stiffness coefficient or elastic moduli
values which predict the observed surface pavement deflections within the
established accuracy limits.
All of the methods but one, a graphical method, are available to
researchers or other users as computer codes with full documentation of the
theory and method of solution available in the references given.

Findings
There is a significant difference in calculated base course stiffness
coefficients depending upon which method of calculation is used. The
simple two-layer approach (which assumes that the pavement is a base
course with a thin surface course or none at all) usually gives base course
coefficients that are lower than those obtained by considering the pavement
as a multi-layered structure. The values of the subgrade coefficient were
not significantly different regardless of the method of calculation used.
The methods of calculating elastic moduli which use all five of the
observed surface deflections in the solution process are considered to
give more accurate estimates of elastic moduli than those methods which use

v
only two deflection observations in the solution process.

Conclusions
Two new methods of determining stiffness coefficients in multi-
layer pavements using observed surface deflections are recommended to
pavement designers and researchers who have need of these values.

vi
 IMPLEMENTATION STATEMENT

The methods of calculating stiffness coefficients or elastic moduli of

pavement structures from observed surface pavement deflections are available
for immediate use by researchers and other users.
Two recently developed methods for calculating stiffness coefficients
of multi-layer pavement designs are available and are recommended for use
by researchers and other users who need more reliable values of stiffness
coefficients than were previously available.

vii
 TABLE OF CONTENTS

PREFACE . . ii

DISCLAIMER ii

ACKNOWLEDGMENT ii

LIST OF REPORTS . . . . . . . iii

ABSTRACT iv
SUMMARY . v
IMPLEMENTATION STATEMENT vii
LIST OF FIGURES . X

LIST OF TABLES xi
INTRODUCTION . 1
DEVELOPMENT OF THE DYNAMIC DEFLECTION EQUATION 3

METHODS OF COMPUTING MATERIAL STIFFNESS COEFFICIENTS 7

STIF2 . . . 7
STI F5 . . 8
SCIMP . . . 9
SCIMP-PS . 11
Comparison 12
METHODS OF COMPUTING ELASTIC MODULI . 16

ELASTIC MODULUS I . . 16
ELASTIC MODULUS II . 18
GRAPHICAL TECHNIQUE 18
EMP I . . . . . . . . . 23
EMPIRICAL RELATIONS BETWEEN BASE COURSE ELASTIC MODULUS AND
STIFFNESS COEFFICIENT . o • • • • • • o ••• 26

CONCLUSIONS AND RECOMMENDATIONS . 28

REFERENCES . . . . . . . . . . 29

viii
APPENDIX A - DOCUMENTATION OF A METHOD OF COMPUTING STIFFNESS
COEFFICIENTS IN SIMPLE TWO LAYER PAVEMENT STRUCTURES
USING FIVE OBSERVED SURFACE DEFLECTIONS, COMPUTER
PROGRAM STI FS • • • . • • • A-1
INTRODUCTION . . . . . . . A-2

PROGRI,\M IDENTIFICATION . A-3

PROGRAM DESCRIPTION . . . . A-4
FLOWCHART . . . • • • A-6

PROGRAM LISTING . • . A-8

NAME DICTIONARY . . . A-17

INPUT GUIDE .. . • A-21
OUTPUT FORMAT . • A-25

EXAMPLE PROBLEMS • A-26

APPENDIX B - DOCUMENTATION OF A METHOD OF COMPUTING STIFFNESS
COEFFICIENTS IN MULTI-LAYER PAVEMENTS, COMPUTER
PROGRAM SCIMP . . . B-1
INTRODUCTION . . . . . . . 8-2
PROGRAM IDENTIFICATION . . 8-3
PROGRAM DESCRIPTION . . . 8-4

FLOWCHART . . . . · · B-14
PROGRAM LISTING . . . . B-17
NAME DICTIONARY . . • • B- 25

INPUT GUIDE . . • . B-36

OUTPUT FORMAT . . . . . 8-41

EXAMPLE PROBLEr1S . B-42

ix
 LIST OF FIGURES
Figure Page
1 A pavement section of n layers . . . . . . . . . . . 4
2 Position of dynaflect sensors and load wheels during test . 5

3 Typical deflection basin reconstructed from Dynaflect

readings . . . . . . . . . . . . . . . . . . . . . 5

4 Typical plot of w; versus r, derived from measured

deflections . . . . . . . . . . . . . ... 20

5 Two-layer elastic deflection chart . 21

6 Example showing the chart superimposed in 11 best fit 11

position on the deflection plot . . . . . . . . . . 22

X
 LIST OF TABLES
Table
1 Stiffness coefficients of black base calculated by
four different methods . . . . . . . . . . . . . . 14
2 Stiffness coefficients of limestone rock asphalt calculated
by four different methods . . . . . . . . . . . . . . . . • . 15

xi
 INTRODUCTION

The increasing use of computer codes such as FPS and RPS in pavement
design by the State Department of Highways and Public Transportation in
Texas and by pavement designers in other states has created the need for
accur~te measurements of the many variables that are inputs to these
computer codes (l).
One of the variables for which an accurate measurement is ·needed is
the stiffness of the materials that will be considered in selecting the
optim&l pavement design. This structural value of the pavement material
can be either the FPS "stiffness coefficient" or the elastic modulus of
the material.
There have been several methods formulated and converted to computer
codes for computing the stiffness coefficient or the elastic modulus of a
material to be used in the FPS or some other pavement design computer code
{2, 3, 4, 5, 6, 7, 8). All of the methods are based on dynamic
surface deflections of existing pavements as a predictor of the performance
(i.e., life) of the pavement structure, which was shown to be feasible from
the results of the WASHO and AASHTO Road Tests (9, 10).
:Scrivner, et al~, developed an equation· to~. predicting the· svrface
deflections of a pavement subjected to a known load (11). It is this
deflection equation which is the basis for the several computer codes that
compute stiffness coefficients of pavement materials. The development of
this deflection equation is presented in the next section. The equations
in t~e computer codes for computing elastic moduli of pavement materials
are from Burmister's theory of elasticity in layered pavements (14).
This report presents summaries of the several methods of computing
stiffness coefficients or elastic moduli of pavement materials, including

1
the basic equations and a brief explanation of the method of solution. The
user can select the method that best suits his needs and resources and can
obtain detailed descriptions and instructions for using the method selected
from the references.

2
 DEVELOPMENT OF THE DYNAMIC DEFLECTION EQUATION

The sketch in Figure 1 represents a pavement composed of n layers,

including the subgrade. The material in each of these layers is
characterized by a stiffness coefficient, ai' where the subscript, i,
identifies the position of the layer in the structure. Layers are
numbered consecutively from the top downward, thus, i = 1 for the
surfacing layer, and i = n for the foundation layer, which is considered
to be of infinite thickness. The thickness of any layer above the foundation
is represented by the symbol, D1•
An important feature of the flexible pavement design procedure in Texas
is the use of the_ Dynaflect* for measuring deflections on existing highways.
Descriptions of the instrument and examples of its use in pavement research
have been published previously (see reference 1). Suffice it to say here
that a dynamic load of 1000 lbs, oscillating sinusoidally at 8 cps, is
applied through two steel load wheels to the pavement, as indicated in
Figure 2. Five sensors, resting on the pavement at the numbered points
shown in the figure, register the vertical amplitude of the motion at those
points in thousandths of an inch (or mils).
A deflection basin of the type illustrated in Figure 3 results from
the Dynaflect loading. The symbol w1 represents the amplitude--or deflec-
tion--occurring at Point 1, W2 is the deflection at Point 2, etc.
An empirical equation used in Texas for estimating the dynamic
deflection, Wj' from the design variables a1 and D; was developed from
deflection data gathered on the A&M Pavement Test Facility located at
Texas A&M University's Research Annex near Bryan. A description of the

* Registered Trademark, Dresser Industries, Inc., Dallas, Texas

3
 PAVEMENT SURFACE

LAYER I

LAYER 2

LAYER 0·I o,

LAYER n-1

FOUNDATION
(LAYER n--} - =oo
Dn
I

Figure 1: A pavement section of n layers •

4
 SUBGRAOE

Figure 2: Position of Dynaflect sensors and load wheels during

test. Vertical arrows represent load wheels. Points
numbered 1 through 5 indicate location of. sensors •

.A B
ORIGINAL SURFACE
...... .._.._
..............
............
........
'' SURFACE

DEFLECTION (w,)
'
\-==~
\
----
24·--...

SURFACE CURVATURE
\
INDEX (5 =w.--w,)
' c
+
w
Figure 3: Typical deflection basin reconstructed from Dynaflcct
readings.

5
facility is contained in reference (12). The deflection equation, developed
in reference (11) is given below:
n
W. = l: ~jk (1)
J k=l
where
W. = surface deflection in thousandths of an inch (or mils) at
J
geophone j,
n = number of layers, including subgrade, and

~jk c o_
= ___
ak
cl
[
r.
J
.
2
1
. k-1
+ c 2 ( l: a.O.)
i=O 1 1
2 2 ~
r. + c2 ( l: a.D.)
J . 0 1 1
1=
2]
and
co= 0.891087,
cl = 4.50292,
c2 = 6.25,
a 1 -stiffness coefficient of layer i (a 0 = 0),
D; =thickness in inches of layer i (0 0 = 0 and On= oo),
r. = distance in inches from point of application of either load to
J
the jth geophone.
Measured in situ values of the stiffness coefficient, a, range from
about 0.17 for a wet clay to about 1.00 for a strongly stabilized base
material. No way has been found for predicting·these values with suitable
accuracy from laboratory tests. For the present, the stiffness coefficient
of a material proposed for use in a new pavement in a particular locality
must be estimated from deflection measurements made on the same type of
material in an existing pavement located in the same general area.

6
 METHODS OF COMPUTING MATERIAL
STIFFNESS COEFFICIENTS

Several computational procedures and computer codes (STIF2, STIFS,

SCIMP, SCIMP-PS), which employ the deflection equation (Eq. 1) to calculate
stiffness coefficients based on the surface dyn~mic deflections, have been
developed, and are described briefly below.

STIF2
A computer program, STIF2, has been developed to calculate the stiff-
ness coefficient of materials in a special case of a simple structure
consisting of a relatively thin surfacing layer (say less than 2 inches) on
a base with no subbase layers between base and subgrade (2). In this
special case, the surfacing and base layers can be considered one material,
with a thickness of o1 and a stiffness of a1, resting on an infinite
foundation with a stiffness, a2. Consider two deflections: w 1 and w2.
The equations for these deflections are, according to Eq. 1,

w, = c~l [~-
a1 r1
2 1 ~ c~l
r 1 +c 2(a 1D1 ):J
+
a2
[ 2 1 2] (2)
r 1 +c 2(a 1D1 ) ~

By eliminating a2 between these two equations, the following equation in

a1 can be formed:

(4)

7
In Eq. 4 all quantities except a1 are known. The value of a1 can be found
by iteration (trial and error). With a1 known, a2 can be found from either
Eq. 2 or Eq. 3, as a 2 can be isolated in either of those equations.
In theory, any pair of the deflections Wj can be used to find a 1 and
a 2, and the values found from any pair should equal those found from any
other pair. In practice, it has been discovered that this rule does not
generally hold--the values found by using w1 and w2 , for example, are, in
fact, not precisely the same as those found from w1 and w 5 ~ The difference
is ascribed to experimental error, including the error in assuming that a
simple two-layer structure of the type envisioned actually can exist in the
case of a real pavement and its foundation. The difference can also be
ascribed to imperfections in the mathematical model.

STIF5
Usually, the Dynaflect measures the surface deflection at five
geophones. A least square fit of the five deflections would provide
better estimates of a1 and a 2 than those determined by STIF2 which uses two
deflections (3). For a two-layer pavement structure, according to Eq. 1,
W. =F.+ B0 · f. (5)
J J J

where

8
 The constants c0, c1 , c 2 and the variables a1 , o1 , and rj are as
defined previously. The a1 and a 2 values can be found by an iteration process
such as the one described as follows.
Assume an a1 value, then Fj and fj at five geophones can be calculated. B
0
is thus determined by:
5 5 2
B
0
= E (W. - F.)f./ E f .• (6)
j=l J J J j=l J

In turn, a2 can be estimated from B0 ,

1
c,
a2 = (c01B0 ) • (7)

Applying the assumed a1 and the calculated a2 valu~s in the deflection

equation (Eq. 5), the predicted Wj (designated by Wj) at five geophones can
thus be determined. Then the root-mean-square~error can be calculated from

1 5 A 2
RMSE = - E (W.-W.) (8)
5 j=l J J

Now assume another a1 value and repeat the computational procedure to

A
calculate the a2 , Wj and RMSE. A Fibonacci search scheme is utilized in
the STIF5 program to select the optimal a1 and a 2 values yielding the
minimum RMSE which would best represent the material stiffness based on the
dynamic deflection data.

SCIMP
The application of STIF2 and STIF5 programs are restricted to two-layer
pavement structures (or three-layer pavements with a relatively thin
surfacing layer, say, 10%). For more than two layers, the SCIMP (Stiffness
Coefficient in Multi-layer Pavement) ~rogram may be used. Given stiffness
coefficients of n-2 layers above the subgrade in an n-layer pavement
structure, where n ~ 3, the SCIMP calculates the stiffness coefficients of
one of the n-1 pavement layers and the subgrade. In mathematics, the
computational procedures are represented as follows.

According to Eq. 1, the deflection equation can be rewritten as

(5)

where,

~
n-1
co [ 1
F.=
J
l:
k=l ~
ak
2 k-1
r. +c 2 ( E a.D.)
2 2 2]
r. +c 2( l: a.D.) .
J i =0 1 1 J . 0 1 1
1=

c
B = _0_
0 c,
an

f. = - - - - - =1 , . - - - -
J 2 n-1 2
r . +c ( l: a . D. )
J 2 1= 0 1 10

To illustrate the algorithm, let n = 5 and a 2, a3 , a4 be known. Assume an

a1 value, then Fj and fj at five geophones can be calculated. s0 is
determined by
5 s
B0 = l: (W.-F.)f./ l: f. 2 (6)
j=l J J J j=l J

as can thus be estimated from s0

1
c,
as = (co/Bo)
The same Fibonacci search scheme as used in STIFS is utilized here to find
the optimal a1 and as values which yield the minimum RMSE. It should be
pointed out that the unknown pavement coefficient can be any one of the
pavement layers above the subgrade.

10
SCIMP~PS

The SCIMP-PS stands for Stiffness foefficient in ~ulti-layer ~avements,

by ~attern iearch. The original SCIMP program calculates the stiffness

coefficient of one pavement layer and the subgrade, while the stiffness
coefficients of other pavement layers of the multi-layer pavement
structure must be known. The SCIMP by pattern search determines the
stiffness coefficients of all pavement layers pnd the subgrade. The
deflection equation of an n-layer pavement structure (Eq. 1) can be
expanded as follows:

Wj = -a ~-g-1 ~-2 2-+c-2-~-a

[ -r - -rJ-. 1-D-1-)J
+ -a:-g-1 [rj2+c2~a1D1)2-
+ ·- .. . •· •.

2
1
n-1
rJ. +c 2( E
j=l
a.D.)~~
1 1

+ • • • ~ . .

11
 + (-1- - _1_ [
a c1 a Cl 2 n-1
1
2
]1
n n-1 r. +c 2 ( E a.D.) .
J i=l 1 1

=co l(+lH-\-l+
a1 rJ. m=
~2 (+1 - +,)[
am am-1 r. 2+c ( E a .D.)
m~l 2]1. (10)
J 2 i =1 1 1

The pattern search method is a standard library routine (15) which is

programmed to conduct an efficient multi-dimensional search for the set
of n variables which minimizes a prediction error whose form is specified
by the user. In this application, the pavement stiffness coefficients of
all n layers are the variables that are sought. The prediction error (PE)
which is minimized by the pattern search is the squared error between
observed and calculated deflections, and thus by minimizing the following
prediction error a least-squares fit is assured.

PE NoE NdE [ (1- ) (1-)+ E

n ( 1 1 ) '(
=
k= 1 j =1 a c1 r j 2 m= 2
;cl - ;cl ·-2--(m---=-l---2
1 m m-1 r. +c E a.D.)
J 2 i=l 1 1

- wkj] 2.
c0 J ( 11 )

where
N = number of observations
0
Nd = number of geophone measurements per test
Wkj = the kth observation of surface deflection Wj as defined before.

Comparison
STIF2, STIF5 and SCIMP can be used to predict two unknown stiffness
coefficients. One of the two unknowns must be the stiffness coefficient of
the subgrade (or the foundation). For STIF2 and STIF5, the other unknown
is the stiffness coefficient of the composite layer of materials above the

12
subgrade, that is, the program is restricted to 2-layer pavements. The
SCIMP can be used for multi-layer pavements. The SCIMP calculates the
stiffness coefficient of any one selected pavement layer and the subgrade,
when the stiffness coefficients of all other pavement layers are given.
The PS version of the SCIMP, SCIMP-PS, has the versatility to calculate
the stiffness coefficients of all pavement layers, and the subgrade.
STIF2, STIF5 and SCIMP calculate the stiffness coefficients of each
individual observation and average the total observations. The SCIMP-PS
minimizes the prediction error of the total observations.
STIF2, STIF5 and SCIMP programs are very economical in numerical
computation, in comparison with the SCIMP-PS. Due to the cumbersome
computations required in the pattern search routine, the SCIMP-PS program
is suggested for a maximum of S-layer pavements. For more than 5 layers,
other searching techniques which can be found in most optimization and
operations research textbooks are recommended.
A comparison of example solutions of STIF2, STIF5, SCIMP and SCIMP-PS
is shown in Table 1 and Table 2. Table 1 shows the stiffness coefficients
obtained for two designs of a black base pavement using each of the four
methods to calculate stiffness coefficients. Table 2 shows the stiffness
coefficients obtained for two designs of a limestone rock asphalt pavement
using each of the four methods to calculate stiffness coefficients.

13
 TABLE 1: SUMf·1ARY OF STIFFNESS COEFFICIENTS OF A BLACK BASE PAVEMENT
CALCULATED BY FOUR DIFFERENT METHODS
SDHPT
STIFFNESS STIFS SCIMP SCIMP-PS
COEFFICIENTS COEFFICIENTS COEFFICIENTS COEFFICIENTS
STATION AP2 AS2 APS ASS * Al A2 A3 A4 AS Al A2 A3 A4 AS
Sl 0.62 0.24 O.S8 0.2S 0.40 0.49 0.72 0.42 0.2S 0.40 0.43 0.72 0.42 0.24**
S2 0.48 0.28 O.S3 0.26 0.38 0.2S
S3 O.S4 0.26 O.S6 0.2S 0.43 0.2S
S9 o.ss 0.2S 0.60 0.23 0.49 0.23
SlO O.S3 0.24 O.S3 0.24 0.39 0.23
Sll O.S6 0.23 O.S8 0.22 0.46 0.22
AVERAGE 0.55 0.2S O.S6 0.24 0.44 0.24 0.40 0.43 0.72 0.42 0.24
DESIGN: 2-Layers 2-Layers S-Layers S-Layers
PAVEMENT THICKNESSES: 32 Inches 32 Inches Tl=l.O in., T2=4.0 in., T3=19.0 in., Tl=l. 0 in., T2=4 .0 in., T3=19. 0 in.,
T4=8.0 in . T4=8.0 in.
......
.;:::. S4 0. 61 0.24 0.58 0.25 *0.88 0.45 0.69 o. 71 0.24 0.88 0.50 0.69 0.71 0.24
ss 0.84 0.22 0.70 0.24 0.70 0.24
S6 o.so 0.30 0.61 0.2S 0.46 0.2S
57 0.6S 0.22 0.66 0.22 0.58 0.22
sa O.S2 0.27 0.58 0.25 0.43 0.24
AVERAGE 0.62 0.2S 0.63 0.24 O.S2 0.24 0.88 0.50 0.69 0.71 0.24
DESIGN: 2-Layers 2-Layers S-Layers S-Layers
PAVEMENT THICKNESSES: 26 Inches 26 Inches Tl=1.0 in., T2=4.0 in., T3=13.0 in., Tl=l.O in., T2=4.0 in., T3=13.0 in.,
T4=8.0 in. T4=8.0 in.
*Note- The Al, A3, and A4 values were assumed for these problems
**Note - The pattern search method calculates a single set of coefficients that best fits all the data observations
 TABLE 2: Sut·1MARY OF STIFFNESS COEFFICIENTS OF A LIMESTONE ROCK ASPHALT PAVEMENT
CALCULATED BY FOUR DIFFERENT METHODS
SDHPT
STIFFNESS STIFS SCIMP SCIMP-PS
COEFFICIENTS COEFFICIENTS COEFFICIENTS COEFFICIENTS
STATION AP2 AS2 APS ASS * Al A2 A3 A4 AS Al A2 A3 A4 AS
Nl 0.49 0.26 o.so 0.2S 2.77 0.38 O.S4 0.4S 0.24 2.77 0.3S 0.54 0.45 0.24**
N2 0.62 0.23 0.69 0.22 1.04 0.22
N3 0.47 0.26 0.48 0.26 0.3S 0.24
N9 0.43 0.24 0.41 0.24 0.28 0.22
NlO 0.4S 0.2S 0.4S 0.25 0.32 0.23
Nll 0.44 0.29 0.48 0.25 0.36 0.24
AVERAGE 0.48 0.26 0.50 0.25 0.46 0.23 2.77 0.35 0.54 0.45 0.24
DESIGN: 2-Layers 2-Layers S-Layers S-Layers
PAVEMENT THICKNESSES: 32 Inches 32 Inches Tl=l.O in., T2=4.0 in., T3=19.0 in., T1 =1 . 0 i n. • T2=4. 0 in. , T3= 19. 0 in .•
T4=8.0 in. T4=8.0 in.
N4 0.47 0.26 0.47 0.2S *0.67 0.43 0.47 0.60 0.25 0.67 0.38 0.47 0.60 0.26"**
__, N5 0.45 0.27 0.49 0.25 0.47 0.25
0'1 N6 0.48 0.26 0.49 0.26 0.48 0.25
N7 0.41 0.26 0.42 0.26 0. 31 0.24
N8 0.43 0.24 0.43 0.25 0.34 0.23
AVERAGE 0.45 0.26 0.46 0.25 0.41 0.24 0.67 0.38 0.47 0.60 0.26
DESIGN: 2-Layers 2-Layers 5-Layers S.;.Layers
PAVEMENT THICKNESSES: 26 Inches 26 Inches Tl=l.O in., T2=4.0 in., T3=13.0 in. Tl =1 . 0 in. , T2=4. 0 in. , T3= 13. 0 in. ,
T4=8.0 in. T4=8.0 in.
*Note- The Al, A3, and A4 values were assumed for these problems
**Note - The pattern search method calculates a single set of coefficients that best fits all the data observations
 METHODS OF COMPUTING ELASTIC MODULI

The computer codes for computing elastic moduli of two-layer pavements

were developed from Burmister's theory of elasticity in layered pavement (14).
All of the methods for computi·ng the elastic moduli of pavement layers
described herein are limited to simple two-layer designs consisting of a
thin or relatively thin surface layer overlying a base layer, resting on
the subgrade. The rigorous mathematical theory involved for more than two
layers limits the present methods to the simple two-layer design. All of
the methods are based on finding the values of elastic modulus for the
subgrade and the base/surface layer which provide the best agreement between
the observed surface pavement deflections and the calculated pavement
deflections.

ELASTIC MODULUS I

Scrivner, et al. deve1op~d the first method of computfng the in situ

values of elastic modulus of materials for consideration in the FPS
sys tern (4).
If the pavement is assumed to consist of a base or surface layer of
known thickness resting on a homogeneous subgrade of infinite depth and
with Poisson's ratio of l/2, the equation for the surface deflection of a
point according to the theory of elasticity is given by the equation

( 12)

where
= a point load
P
E1 = elastic modulus of the upper layer
E2 = elastic modulus of the subgrade layer

16
 w = the surface deflection of a point on the surface
r = the horizontal distance of the measurement of the deflection w
from the load P
x = mr/h, where m is a parameter

By making certain approximations and assumptions described in reference (4),

Equation 12 can be reduced to

4nEl X=lOr/h
~wr: 1 ~ (V-1) J 0 (X)dx (13)
X=O
Equation 13can then be integrated in the solution process for finding the
elastic moduli of the 2-layer pavement structure. A simplified description
of how Equation 13 is used to calculate the elastic moduli follows.
The surface deflections w1 and w2 of two points located at distances
r 1 and r 2 from the load P, and the thickness of the upper layer h are known.
Let F represent the function on the right side of Equation 13.
The following equations can then be rewritten

(14)

( 15)

A single equation in which E2/E 1 is the only unknown can be obtained

by dividing Equation 14 by Equation 15 as follows:

w1r 1 = F(E 2;E1 , r 1/h)

( 16)
w2r 2 F(E 2;E 1 , r 2/h)

17
By usinq a convergent process, a value of E2;E 1 can be found that satisfies
Equationl6within desired accuracy limits. E1 can then be calculated from
Equation 14 and E2 can be found from the relation

E2
E2 = El (E) ( 17)
1

ELASTIC MODULUS II
Another version of the method by Scrivner for computing in situ values
of elastic moduli was developed primarily to compute the in situ moduli of
rigid pavements (5). The same equations, assumptions, and approximations
described previously hold with the exception that the distance between the
observed surface deflections was increased by one foot. This change was
felt necessary because in the case of rigid pavements, the difference in
the surface deflections of points only one foot apart was not enough to
ensure that accurate values of the elastic modulus of a rigid pavement layer
could be obtained using elasticity theory.
The two methods of computing in situ values of elastic moduli of pave-
ment structures, described above are available as computer codes.

GRAPHICAL TECHNIQUE
A graphical technique of determining in situ elastic moduli of simple
pavement structures from observed surface deflections of the pavement
was developed by Swift (6). From the equation for surface deflections at
various distances from a point load, the following dimensionless relation-
ships can be written if Poisson's ratio is assumed to be l/2 for both layers
of a simple two-layer system:

( 18)

18
The terms w, r, P, E1 , E2 , and hare the same as defined earlier. For any
wr E
value of the ratio E1/E 2 , the quantity P 2 is a different function of
the ratio r/h. The ratio E1/E 2 can be evaluated because of this by finding
the best match between a given set of measured deflection data and the set
of computed deflection data for a certain value of E1/E 2 .
After finding the E1JE 2 value for which the calculated deflections
most closely matches the observed deflections, the E2 value can be
obtained from
wr E2
E2 = p (Computed) + ~r (Obs.) ( 19)

E1 is obtained from

(20)

The procedure for finding the elastic moduli of two layer pavements
using this graphical technique is briefly described as follows. Values of
~r, with w being the observed surface deflections, are plotted versus r using

logarithmic scales, as in Figure 4. This curve is then superimposed on a chart

containing many curves computed for different E1/E2 and r/h ratios (Figure 5).
The location on the observed deflections plot r/h = 1.0 is aligned with a
similar location on the chart. The plotted curve is then moved vertically
up or down until it best matches one of the computed curves on the chart
or a best fit between two of the computed curves is found (Figure 6). The ratio
E1/E 2 and the value of E2 are read directly from the chart. The value of E1
is calculated from Equation 20 as follows:

A more detailed explanation of this graphical method for obtaining elastic

moduli of pavement materials from surface deflections as well as limitations

19
 0.10 IIIIHimmlllllllll!ll!!!i!l!lftlll! Ill! Ul l1T!HI
9
!JH1!ll :tHHI
a ORIGINAL DATA CALCULATED

-• •
7
I
wl = 0.00040 in. rl = 10 in. wlrl = 0.0040 I
6

w2 :: 0.00037 r2 = 15.6 w2r2 = 0.0058

5

w3 = 0.00032 r3 = 26.0 w3r3 = 0.0083

4

w4 = 0.00028 r4 = 37.4 w4r4 = 0.0105 I

i :
3
ws = 0.00024 rs = 49.0 wsrs = 0.0118
h = 19 in. p = 1000 Square inches
1bs. {per 1000 1bs .)
2
:
!lh Lf
i! fi ;;
''

f'
II l ::::li·'
;:!,
;, .:
f.-: il : ;cifi ;; ·:.
I•
· '.,
[[·
·~
.,;, i :: l';i,. .

0.01 '
f-'-
u , I
1:': .::
'I
,: I·. '"j
i

c:;:. : : .,
·'
i
I
I I: I

i
I

II I I
II :I ! i I
t I ! i I

..
I ;
I
I l '' :!

•
!I I . I·
I
,,
§:
llli
.,

'liT
:! :; I' 1::::
T. . I .. :, r:: ..
r:: I· . :1 .. ·
·:tl ·1: ..
0.001 1
! ;1 ::: I'
2 4 5 6 7 e 9 10 2 3 4 5 6
li :! 7 e 9 10

r (inches)

Figure 4: Typical plot of pwr versus r, derived from measured

deflections.

20
 a- =0.5

~IQ.
~ 4,000
U)o
3,000

2,000

0.1 1.0 10
r/h
Figure 5: Two-Layer Elastic Deflection Chart.
____Q'oi~son's Rat~-~~--cr = 0.5)

21
 O.l ~mm~••••••••••••~ammuawm
a CALCULATED
ORIGINAL DATA
7
w1 = 0.00040 in. r1 = 10 in.
6

w2 = 0.00037
!5

0.01

5,000

Q. 4,000

3,000

0.1 .0 10
r/h

Figure 6: Example showing the chart superimposed in "best fit"

Er
position on the deflection plot. In this case, l'2

is found to be equal to 30 and the value of E2 is

founc to be 22,000 psi. Accordingly, E1 is 22, 000 x 30

or 660,000 psi.

22
on its use can be found in reference (6).

EMPI
An empirical equation for predicting surface pavement deflections of
an elastic pavement structure v1as derived by Swift (7). The results
obtained using his equation are in close agreement with results obtained
from solutions based on Burmister's equations. The equation for surface
deflections of an elastic pavement structure derived by Swift is as follows:

(21)

where

L=N
- 3 ~.
+2E2.
X- 2h ~~
2

P, r, and h = same as defined previously.

The advantage of this equation for obtaining predicted surface deflections
is the ease of its use and the speed of solution as opposed to the more
complicated equations from Burmister.
The empirical equation described above was used by Moore to compute the
elastic moduli of simple pavement structures (8). The equation was rewritten
in the· following form:

(22)

where
X = Jr 2
+ a
2

lr-1-.-E-1--
a = 2h ~ 3 + /E2
3

23
 P, r, h, E1 , E2, and~= same as defined previously.
If the surface deflections of the pavement are measured with the Dynaflect,
the only unknowns in the above equation are E1 and E2 • If the set of E1
and E2 values can be found such that the predicted surface deflections
~ calculated from the equation approximate the observed surface deflections
within the accuracy range specified, then these values can be taken to
represent the elastic moduli of the simple pavement system. The 11
best fit ..
between the predicted and observed surface deflections was determined to
be the minimum value of the root mean square error of the predicted and
observed surface deflections as shown below:

1 n A 2
RMSE = -
n .
E (w. - w.)
1 1
(8)
1= 1

In order to use Equation (22) to calculate the values of E1 and E2

which provided the "best fit between the observed and predicted surface
11

deflections, the equation was written in the general form below:

A
w = s0 · f(r, h, s1 ) + £

3P
where B0 -- 4nE
1

Bl = ElfE2

A
£ =a prediction error (w- w).
The minimum RMSE which determined the best fit was obtained by an iterative
process consisting of several steps. A trial value of B1 was selected
first, and then five values of the function f were computed for the values
of r at which surface deflections were measured. A s0 value corresponding
to the trial value of B1 was obtained from the equation below:
5 2 5
s0 = E w.f. I E f. (23)
i=l 1 1 i=l 1

24
The predicted surface deflections for the trial value of s0 and B were
1
calculated and the RMSE computed.
These steps were repeated until RMSE's for 21 logarithmically spaced
trial values of s1 , chosen to cover the entire range of reasonable values
of E1JE 2 , were computed. The smallest RMSE value was selected as the
starting point of a Fibonacci search technique which determined the s0
and B1 values which gave the absolute minimum RMSE. From these two
values, the elastic moduli of the individual pavement layers was computed
from the following equations:

E = 3000 (24)
1 4n s
0

(2 5)

The solution process described above was converted to a computer code

which provides a faster means of obtaining elastic moduli from surface
deflections than the previous method by Scrivner, et al. (4).
A more detailed explanation of the computer code and certain limitations
on its use can be found in reference (8).

25
 EMPIRICAL RELATIONS BETWEEN BASE COURSE ELASTIC
MODULUS AND STIFFNESS COEFFICIENT

A pavement evaluation (13) project, conducted by TTI under a joint

contract with the City of Houston and Harris County, determined the
elastic moduli and stiffness coefficients of a variety of base courses
and subgrades within their jurisdictions. In-service pavements were
tested by the Dynaflect and analyzed by progra~ EMPI to calculate elastic
moduli and program STIF5 to calculate stiffness coefficients. The tests
were run during the cold months of the year on materials such as cement
stabilized shell, black base, and other stabilized materials whose stiff-
ness coefficients were usually much higher than those normally found in
Texas. The stiffness coefficients ranged from 0.68 to 2.14. In addition,
samples of base course and subgrade materials were cored and tested in a
commercial laboratory to determine the elastic modulus.
The linear correlation of the elastic modulus of base course materials
(E) calculated by EMPI and the stiffness coefficient (S) computed by STIF5
of 82 observations can be represented by
E = -276,242 + 799,622 S
(N = 82, R2 = 0.787, C.V. = 81.5%, S.E. = 338,232)
A convex quadratic function shows a better fit:
E = 80,018 - 460,141 s + 526,390 s2
(N = 82, R2 = 0.935, C.V. = 45.3%, S.E. = 187,849)
However, the elastic modulus determined in the laboratory is not correlated
with the elastic modulus determined by EMPI (R2 = 0.046 by quadratic function),
nor correlated with the stiffness coeffieient determined by STIF5 (R2 = 0.034
by quadratic functio~.

26
 These empirical correlations are useful for converting stiffness
coefficients into elastic moduli. Therefore, it is now possible to use
the pattern search method in SCit1P-PS to determine stiffness coefficients
for an n-layer pavement {5 is the practical maximum) and use the above
empirical relation to infer elastic moduli for all of the layers. The
accuracy of the result is, of course, limited by the range of data from
which the equation is drawn. The linear relation will give negative E's
for stiffness coefficients below about 0.34. The quadratic relation gives
negative E's for stiffness coefficients below about 0.22.

27
 CONCLUSIONS AND RECOMMENDATIONS

The several methods for computing stiffness coefficients or elastic

moduli of materials in existing pavement structures from observed surface
deflections have been presented and briefly discussed.
Each method described has been used or is available for use by pavement
designers or researchers who need accurate measurements of pavement stiffness
coefficients or elastic moduli.
Of the four methods for computing material stiffness coefficents (STIF2,
STIF5, SCIMP, SCIMP-PS), STIF5 and SCIMP give the most accurate results.
SCIMP has the capability of finding the stiffness coefficient of any selected
layer for which the stiffness coefficient volume is unknown. SCIMP-PS
is the only method available for calculating all the stiffness coefficients
in an n-layer pavement.
The program EMPI gives the most accurate estimates of elastic modulus
from surface deflections. The theoretical equation in EMPI for surface
deflections agrees closely with results from Burmister•s theory and the
theoretical equation eliminates the approxin~tions and numerical intregra-
tions which limit the accuracy of ELASTIC MODULUS I and ELASTIC MOOULUS II.
A graphical technique for obtaining elastic moduli of simple two layer
permanent structures from surface deflections is available also.
The different methods presented herein may be chosen by the user to
best suit the user•s needs and available data and computational resources.

28
 REFERENCES

1. Scrivner, F.H., W.M. Moore, W.F. McFarland and G.R. Carey, A Systems 11

Approach to the Flexible Pavement Design Problem Research Report 32-11,

11
,

lexas Transportation Institute, Texas A&M University, College Station,

Texas, 1968.
2. "Texas Highway Department Pavement Design System, Part I, Flexible
Pavement Designer's Manual'', Highway Design Division, Texas State
Department of Highways and Public Transportation, Austin, Texas, 1970.
3. Turman, G.W. and Moore, W.M., Appendix A of this report. Developed in
June, 1975.
4. Scrivner, F.H., C.H. Michalak, and W.M. Moore, "Calculation of the
Elastic Moduli of a Two Layer Pavement System from Measured Surface
Deflections'', Research Report 123-6, Texas Transportation Institute,
Texas A&M University, Highway Design Division Research Section, State
Department of Highways and Public Transportation, Center for Highway
Research, University of Texas at Austin, March, 1971.
5. Scrivner, F.H., C.H. Michalak and W.M. Moore, "Calculation of the
Elastic Moduli of a Two Layer Pavement System from Measured Surface
Deflections - Part II", Research Report 123-6A, Texas Transportation
Institute, Texas A&M University, Highway Design Division Research
Section, State Department of Highways and Public Transportation, Center
for Highway Research, University of Texas at Austin, December, 1971.
6. Swift, Gilbert, "A Graphical Technique for Determining the Elastic
Moduli of a Two-Layered Structure from Measured Surface Deflections",
Research Report 136-3, Texas Transportation Institute, Texas A&M
University, College Station, Texas, November, 1972.
7. Swift, Gilbert, "An Empirical Equation for Calculating Deflections on
the Surface of a Two-Layer Elastic System", Research Report 136-4, Texas
Transportation Institute, Texas A&M University, College Station, Texas,
November, 1972.
8. Moore, W.M., "Elastic Moduli Determination for Simple Two-Layer Pavement
Structures Based on Surface Deflections", Research Report 136-5, Texas
Transportation Institute, Texas A&M University, College Station, Texas,
August, 1973.
9. "The WASHO Road Test, Part 2: Test Data, Analyses, and Findings",
Highway Research Board Special Report 22, 1955.
10. The AASHO Road Test Report 5, Pavement Research
11 11
, Highway Research Board
Special Report 61E, 1962.
11. Scrivner, F.H. and W.M. ~1oore, "An Empirical Equation for Predicting
Pavement Deflections, Research Report 32-12, Texas Transportation
Institute, Texas A&M University, College Station, Texas, October, 1968.

29
12. Scrivner, F.H. and W.M. Moore, "Some Recent Findings in Flexible Pavement
Research'', Research Report 32~9, Texas Transportation Institute,
Texas A&M University, College Station, Texas, July, 1967.
13. Edris, E.V., J.A. Epps and R.L. Lytton, "Layer Equivalents for Base
Courses in the City of Houston-Harris County", Texas Transportation
Institute, Texas A&M University, College Station, Texas, August, 1974.
14. Burmister, D.M., "The Theory of Stresses anct Displacements in Layered
Systems and Applications to the Designs of Airport Runways", Proceedings,
Highway Research Board, Vol. 23, p. 130, 1943.
15. Walsh, G.R., "Methods of Optimization", John Wiley and Sons, 1975.
16. Peirre, Donald A., "Optimization Theory with Applications••, John Wiley
and Sons, March 1969, pp. 280-283.

30
 APPENDIX A

DOCUMENTATION OF A METHOD OF COMPUTING STIFFNESS COEFFICIENTS IN SIMPLE TWO

LAYER PAVEMENT STRUCTURES USING FIVE OBSERVED SURFACE DEFLECTIONS, COMPUTER
PROGRAM STIF5.

Page
INTRODUCTION . . . . . . A·2
PROGRAM IDENTIFICATION . . • • A-3

PROGRAM DESCRIPTION . A-4

FLOWCHART .... .... · A-6
PROGRAM LISTING . . .... ..... · A-8
NAME DICTIONARY . . . ,. . · A-17
INPUT GUIDE .. .... .... • A- 21
OUTPUT FORMAT . . ....... · A- 25
EXAMPLE PROBLEMS . . . . . . . . . . . . . . . . . . . . . . . . . . . . A-26

A-1
 INTRODUCTION

In 1968 Scrivner and Moore (11) presented an empirical deflection

equation developed from a mathematical model fitted to Dynaflect data gathered
on a set of experimentally designed pavement test sections. The equation
contained a material property constant for each layer called a layer11

stiffness coefficient... In addition to the equation, they presented a

technique for estimating layer stiffness coefficients for materials in
existing two layer pavement structures from measured surface deflections.
Included herein is the documentation of a modification to the technique
for estimating layer stiffness coefficients in pavement structures that can
be represented as a pavement layer of known thickness resting on an infinitely
thick subgrade. The modified technique calculates the two coefficients that give
a least squares fit to the entire deflection basin rather than an exact
fit of two selected points on the basin. In some cases the latter technique
yields deflection predictions which have rather large errors and the modified
technique overcomes this problem.

A-2
 PROGRAM IDENTIFICATION

Title: Documentation of a Method of Computing Stiffness Coefficients in

Simple Two Layer Pavement Structures Using Five Observed Surface
Deflections, Computer Program STIF5.
Language: FORTRAN IV and IBM 360 Assembly Language
Machine: IBM 360/65
Programmer: G.W. Turman
Availability: Department of Pavement Design
Texas Transportation Institute
Texas A&M University
College Station, Texas 77843
Phone {713) 845-3735
Date: August 1974
Source Deck: Approximately 500 cards
Storage: 110 k bytes
Timing: (1) Compilation time- 0.40 minutes
(FORTRAN G Compiler)
(2) Execution Time - dependent on number of data observations,
0.25 minutes for 21 data observations
Output: (1) Program list- approximately 500 lines
(2) Program Output- one page per problem (30 lines per page}
Documentation: Michalak, C.H., Lu, D.Y. and Turman, G.W., .. Determining
Stiffness Coefficients and Elastic Moduli of Pavement
Materials from Dynamic Deflections", Research Report 207-1,
Texas Transportation Institute, 1976.

A-3
 PROGRAM DESCRIPTION

The data for STIF5 is input to the main program as described in the
input guide. The deflections at each radial distance are calculated from
the geophone deflection readings and multipliers on the appropriate data
cards. The Surface Curvature Index (SCI) is also calculated, SCI = Wl - W2.
If any W (deflection) is equal to zero, or if any Wis greater than its
preceding W, the cases are flagged to denote data errors and are not
used for further calculations. If the W's are valid observations they are
passed to subroutine STIF5 along with the total pavement thickness for the
stiffness coefficients and RMSE calculations.
STIFS returns to the main program the stiffness coefficients for both
subgrade (ASS) and pavement (AP5) along with their corresponding RMSE's.
The counter N (the number of sets of observations) is incremented and the
program reads the next data card and continues the process until all
stations in a section are read.
A loop is set up to print the stations numbers, measured and predicted
deflections, SCI's, stiffness coefficients for the subgrade and pavement
and RMSE's for all data observations. Messages for the following situations
are printed for which a data observation is not used in the calculations:
1. Data observation computes a negative SCI in which case the
message 'NEGATIVE SCI OTHER CALCULATIONS OMITTED' is printed.
2. Data observation where any Wis equal to zero. 'ERROR IN DATA'
is printed.
3. When the minimum value for RMSE is where the stiffness coefficient
of the pavement is less than .1, the message 'MIN. RMSE AT Al LESS THAN
RANGE SEARCHED' is printed.
4. When the minimum value for RMSE is when the stiffness coefficient

A-4
of the pavement is greater than 7.4 the message 'MIN. RMSE AT Al GREATER
THAN RANGE SEARCHED' is printed.
5. If a value for A(l) between .1 and 2.0 does not yield a positive
B value (Equation in 3), other calculations are omitted and the message
'MIN. RMSE INDICATED AT THE VALUE Al= . AT THIS POINT THE VALUE
OF A2 CANNOT BE CALCULATED' is printed.
After all data and any error messages for a section are printed, the
average deflections, SCI's, stiffness coefficients of the subgrade and
pavement, standard deviations and RMSE are calculated. These averages are
then printed along with the number of points used in calculating each
average. Definitions of the heading abbreviations are given next in
footnote form.
The program then returns to its beginning to read data for another
section or terminates execution normally when all data have been read.
For the equations used in STIF5 and an explanation of how they are
used to calculate the stiffness coefficients, see page 8 of the main
report.

A-5
 STIFS
(MAIN) ~

CALCULATE AVERAGES &

STANDARD DEVIATIONS ..,.,
r-
CALCULATE THE RMSE 0

)>
FOR·THE TRIAL
COEFFICIENTS ~
:X:
I )>
0"1 :;:o
--f

CALL FIBO TO FIND THE

PAVEMENT COEFFICIENT AT
CALL STIF5 TO CALC.·_!-=., THE MINIMUM R~1SE
HE Al AND A2 VA.Ll!E3 =~~.:
EACf-1 ~35~=\'.'!-:~~,

CALL ANS TO CALCULATE

CALCULATE :J::FLECT:·~I\S HE SUBGRADE COEFFICIENT
USING CALCULATEG ~·
AND A2 VAL:..'ES ~DC ..,..':
su~s A~c :~~~-~~~
 FIBO ..!ill..!8.

DETERMINE AREA TO
SEARCH FOR MINIMUM RMSE

PERFORM FIBONACCI SEARCH

CALL FUNC TO CALCULATE
AN RMSE FOR EACH A1
VALUE UNTIL MINIMUM
RMSE IS FOUND

)::o
I FUNC
""'-J

.&iS

CALL DELTA TO CALCULATE

THE F2 AND f2 VALUES
FOR THE FINAL SOLUTION
AL VAL

CALCULATE THE RMSE

FOR THE Al AND A2 VALUES
 PROGRAM LISTING

C STIF5 -- STIFFNESS COEFICIENTS -- MAIN P~OGR~~ (FT~E DFFLECTIONS)

c
c
DIMENSION STA(200),W(200,5),0(10),AW(5),AWV(5),AP5(20~),
*AS5(2,n),LA1(5),LA2(5),LA3(5),LA4(5),LA5(5),L~6(5),A(20,,SCIC2~~),
*IXDATE(2)eCOM~(7),~EM(200,4),SSw(5),S0W(5),RMSF.(201),WC(5),R(5),
*SSWC(5),AWC(5),AWCVCS),SOWC(5);tSAVSW(200)
REAL • 8 STAtOAS,OAP,OBLE.~MSE,AVRMS
QOUND( x,EVEN ) -:: AINT( ( X + EVEN * .s ) / EVEN )
* * EVEN
1' CONTINUE
REA0('5,-l,END=1000) NCARD, ( A(l), I = l , 2'))
FORMAT( l3t 19A4,A1 )
C CALL COt::iF' ( A, 80 )
IF(NCARDeEQetOn) GO TO 11
IFCNCAROeEOe200) G~ TO 12
tF(NCARDeEQe300) GO TO 13
14 t =N + t
REA0(5,6) CONT,SECT,OM,OAYtYEAPeSTA(t),(O(Kt,K=l;t,),
*(REM(I,J),J=1e4)tiCK
6 FORMAT(3Xe A4,4A2,A7,3X,5(F2eltF3e2),8Xt4A4ei2t
IF(I eEQe 1 eAND. IC< eGTe 0) GO TO 222
IF(t eEae 1 eANDe PN eEOe leO) GO TO 222
IF(N eGTe ~) GO TO 555
IFCLO eGTe 0) GO TO 555
PRINT 61
61 FORMAT(/,7X, 1 LOCATION Wt W2 W3 W5 set
*ASS AP5 RMS~ 1 e/)
GO TO 555
22? CONTINUE
PRINT 51
OQJNT 52
PRINT 53,DIST
PRINT 54
CALL DATECJXDATE)
PRINT SS,tXOATE
PRINT 56eDISTtCOleC02,COJ,C04
PRINT 57, CONTeSECT,JOB,HWY1,HWY2,XLANE,OM,OAYtYEAR,OYNA
PRINT 58,(COMM(K)tK=1t7),0P
PRINT 61
555 CONTINUE
L =!
D04J=t,5
W(J,J) = O(L) • 0(_ + 1)
L =
L + 2
4 CONTINUE
SCICI) =
W(l,t)- W(l,2t
c
C TEST FOR Wl OR W2 = O, AND WI LESS THAN W2

IF(W( l t l ».Ea. 0 .ORe W( le2t eEOe Ot GO TO 21

IF(W(Iett eLTe W(l,2)) GO TO 22
DO 7 J=2t4
IF(W(I,.Jt, eLT• W([,J+l)) GO TO 23
7 CONTINUE
I SW 1 =
CALL STJF5(0BLE(W( le1) t ,QBLE(W( le2)) tOBLE(W( y, 3) ),OBLE(W( 1 1 4)),
*OBLE(W(l,5),,0BLECOPt,OAS,OAP,RMSE(IteiSW)
AS5(J) =
0-S
APS(l) =
OAP
ISlV~W( I) = ISW

A-8
 N :: N + l
IFCICK eEOe 0) GO TO 10
GO T'J 80
II ~FAD(5,2) OIST,COl,C02tCO~.C04,CONT,SECT,J~R,HWVltHWY2tXLANF,OP,
•OMeOAVeVEAR,OVNA,(COMM(I),t::t,7)
2 FORMAT( 3Xt A2,3A4eA2,A4tA2eA2,A4,A3,A3,F5.2t4A2•7A4
PRTNT 51
51 FORMAT( '1'
P~tNT 52
52 FORMAT(35X,•TEXAS HIGHWAY DEPARTMENT' , ,
PRINT 5'3tOIST
53 FOPMAT(33X,•DtS~RICT •,42e 1 - DFSIGN SECTION' /)
PRINT 54
54 FORMAT(l8Xt 1 0YNAFLECT DEFLECTIONS AND CALCULATED '•
••sTtFFNESS COEFFICIENTS'/)
CALL DATECtXDATE)
PRINT 55t IXDATE
55 FORM~T(24Xe 1 THIS REVISED PROGRAM (MARe 74) ~AS PUN 'e2A4/t
PRINT 56eOIST,C01,:02eC03,C04
PRTNT 57eCONTtSECTeJOBtHWYltHWY2eXLANE.DMe04Y,YEAR,OYNA
PRINT 58,(COMM(K)eK=le7teOP
58 ~ORMAT(tOX,7A4e2Xe*PAVe T~ICKe = 1 tF5.2e 1 INCHES• , ,
N : 0
LO =
f\
DO 15 J=l•5
AW(J) =
OeO
SSW(.J) OeO =
SOWC.Jt OeO =
AWC(,J) = OeO
s s we c J • = o. o
so we ( .J,= o. 0
I 5 CONTINUE
ASCI : ')•«'
SSSCI = o.o
SDSC I = OeO
SSSC IC 0 eO =
ASCIC = OeO
AAS5 =
OeO
SS4S5 = OeO
SDAS5 =
OeO
AAPS = OeO
SSAP5 :: OeO
SOAPS =
o.o
AVRNS =
OeO
RUt = too.
R(2) :: 244e
R(3J = 676.
R( 4) =
1 396e
Q(5) = 2404.
c = .8911
Cl = 4e 5029
CC = 6e25
GO TO l?
C QEAO & PRINT INFORMATION ON DATA CARD 2
12 QE"AO(S,J) (LA1(1),[:t,5t,Tt,(LA2(J),I=t,5),T2,
• (LA 3 ( I ) • I:: t , t; t , TJ
l FOPMAT(3Xt 5A4,F4e2t5A4,F4e2t5A4,F4.2)
. PRIII.IT 59, (LAIC It,f=t,S) ,Tl,(LA2( 1),1=1 1 5) 1 T'2
PRINT 59, (LA3Citet=I 1 5) 1 T3
59 FORMAT(16X, 5A4tlXeF5•2•5X,5A4,1X,F5.2/t
GO TO 10

A-9
C RE~O & PRINT tNFOR~~TION ON DATA CARD 3tiF PRESENT
13 RE~0(5,3) (L44(t),t=1,5),T~,(L~5(1),(=t,5),T5,
*
(L~6(1),t:1,5J, T6
PRINT 59,(LA4(1),t:t,5),T4,(LA5(1)•1=1•5t,T5
PRINT sq, (L46(1),1=1,5),T6
GO TO 10
23 CONTINUE
~55(1) = 55555
~PS(I) 55555 =
N = N + t
IF(ICK eEae OJ GO TO 10
GO TO 80
22 CONTINUE
AS5(1) : 7777777
~P5(l): 7777777
N N + 1=
tFCtCK eEae 0) GO TO 10
GO TO 80
21 CONTINUE
~5!5(1) 888888 =
APS ( t ) 888888 =
N :: N + 1
IF(ICK eEae 0) GO TO 10
80 CONTINUE
LO = 0
NNT 0 =
DO 51) t=t,N
IF(~55(1) .ea. 7777777) GO TO 24
1F(~55(1) .ea. 888888) GO TO 25
IF(A55(1) .ea. 55555) GO TO 26
ISW =
ISAVSW(IJ
GO TO (100e200,300,~00) , ISW
tOO CONTINUE
DO 111 K=l• 5
D4P5 =
(C/AP5(1)**C1)*(1e/R(K)-(1e/(R(K) + (CC*AP5(1)**2*DP**2tlt)
D~S5 =
(C/4S!5(I)**C1)*(le/(R(K) + (CC*~P!5(1)**2*DP**2)))
WC(K) =
OAPS + DAS5
SSWC(K) :: 5SWC(K) + WC(K)**2
~WC(K) = ~WC(K) + WC(K)
111 CONTINUE
SCtC =
WC(l)- WC(2)
SSSCIC =
SSSCIC + SCIC**2
ASCIC =
ASCIC + SCIC
PRINT 63,ST,(I J.(W( t,J) ,J::t,5),5CICI)
63 FORMAT(7X,A7t1Xt6(2XtF5.31)
PRINT68,(WC(K),K::t,5),SCIC,AS5(It,AP5(1),RMSECil
68 FO~MAT(Tl6,6(2X,F5e3),2(2XtF4.2)tT7S,F8e4t/)
GO TO 110
200 CONTINUE
PRINT l63,ST4( I )
163 FO~MAT(7X,A7,3Xt 1 MINe RMSE AT At LESS TH4N RANGE SEARCHED' 1 / )
MNT MNT + 2 =
LO -::: LO + l
GO TO 51)
~00 CONTINUE
PRINT 263tST4(1)
2~3 FQRM~T(7X,A7,JX,'NINe RMSE AT ~~ GRE~TER TH~N R~NGE SE~RCHF0• 1 /)
MNT = NNT + 2
LO = LO +
GO TO 51)
400 CONTINUE

A-10
 WRJTE(6,363) STA(I),RMSE(I),AP5(1)
3~3 FOqMAT(7X,A7,3Xt 1 MINeRMSE(I,F8e4,q INOJCAT.:::O AT TlofE VALUE &.l= 1 ,

*~4e2t/t17X,•AT THIS POINT THE VALUE OF A2 :ANNOT BE CALCULATFD 1 )

MNT = NNT + 2
LO = LO + 1
GO TO !50
110 CONTlNUI!
AASS = AASS + ASS(It
AAP~ =
AAPS + APS(tt
SSASS = SSASS + ASS(I)t$2
SSA05 = SSAP5 + APS(It**2
AVPMS = AVRMS + RMSE(I)
MNT : MNT + 3
00 16 M=t ,S
SSW(M) = SSW(Mt + W(J,M)*$2
16 AW(N) = AW(M) + W( t,M)
ASCI= ASCI+ SCI(!)
SSSCI = SSSCI + SCI(! t••2
IF(MNT eGEe 30) GO TO 31
GO TO 50
31 PRINT 51
PRINT 56tOIST,COt,C02,C03,C04
56 FORMAT( T35,'DtSTe COUNTY 1 /T36tA2e9Xt3A4eA2 /t
PRINT 57, CONTeSECT,JOB,HWY1tHWY2eXLANEeDMtDAYeYEAR,DYNA
57 FORMATCT19e 1 CONTe SECTe JOB HIGHWAY DATE•,
• 1 0YNAFLECT 1 / T19eA4,5XtA2,5X,A2e2X,A4,2A3,2X,A2t'-'•A2,
••-•,A2e7XeA2e/)
PRINT 61
MNT = 0
GO TO 50
2 4 P R I NT 6 5 , S T A( I ) , W( I tl ) , W( I , 2 )
65 FORMAT(7XeA7,tX,2(F7e3),2Xe 1 SCI ZERO OR LESS OTHER CALCULATIONS '•
*'OMMITE0 1 e/t
MNT = MNT + 2
LO = LO + 1
GO TO 50
25 PRINT 66eSTA(I)
~6 FORMAT(7XeA793Xe'ERROR IN OATA 1 , / )
MNT = MNT + 2
LO = LO + 1
GO TO 50
2~ PRINT 36leSTA(I),(W(I,J),J:t,5),SCI(I)
361 FORMAT(7XeA7e1Xt6(2XtF5e3))
PRINT 67,STA(l)
E-7 FORMATC7X,A7e3X, 1 NOTE DISCONTINUITY IN DEFLECTIONS 1 e/t
MNT = MNT + 2
LO =LO + 1
~n CONTINUE
C CALCULATE •vER AGES
N = N - LO
PN = N
DO 17 M=1, 5
AWCV( M) = AWC ( M) , N

,,,
17 AWVOU = AW(M) / N
ASCIV = ASCI N
AAPSV = AAP5 N
AAS5V = AASS
ARMSV = AVRMS ,,N
N
ASCICV = ASCIC N
IF(N eLEe 1 , GO TO 999
00 112 M=l,S

A-ll
 SOWCCMa =
SQRT((SS•C(M) - (N * (AWCV(Mt**2))t I (N- t ))
112 SOW(M) =
SQRTCCSSW(M)- CN * CAWV(M)**2nt I (N- 1))
SDSCIC • SQRT((SSSCIC- (N * ('SCICV**2))) / (N-1))
SOSCI = SQRTCCSSSCI • (N • CASCIV*412))) I (N- Itt
SOAS5 =
SQRTCCSSASS- (N • CAAS5V**2))) / CN- 1))
SOAPS= SQqT((SSAPS- CN • (AAP5V**2))) / (N- 1))
PRINT 8\tCAWV(J),J:t,St,ASCIV
81 FORMAT(/,7X, 1 AVERAGES',6(2X,F5.3))
PRINT 70tCAWCV(J),J:t,S),ASCtCV,AAS5VtAAP5VtA~MSV
70 ~ORMAT(T16e6(2X,F5.3),2(2XtF4.2),T75,FB.4e/)
PRINT 113tCSDW(J),J=lt5)tSOSCI
113 FORMAT(7X, 1 STO DEV '•6(F7.3))
PRINT 114eCSOWCCJ),J=lt5)tSDSCICtSDAS5,SDAP5
114 FORMATCT16,6(2XeF5e3),2(2X,F4e2),/t
PRINT 82tN
82 FORMAT(7X, 1 NUMBER OF POINTS IN AVERAGES •,!3)
999 CONTINUE
PRINT 91
91 FORMAT(/lOX, 1 W1-5 MEASURED DEFLECTIONS AT GEOPHON=s 1t2t3e4 1 ,

*' AND 5 1 )
PRINT t 91
191 FORMATC16X, 1 CALCULATED DEFLECTIONS AT GEOPHONES 1,2,3,4 AND 5 1 )

PRtNT 96
96 FORMAT( IOX, 1 SCI SURFACE CURVATURE INDEX ( Wl MIN'•
• •us w2 • • ,
PRINT 97
97 FORMAT(10Xe•AS5 STIFFNESS COEFFICIENT OF THE SUBGRA0~ 1 )
PRINT 98
98 FORMATC10Xt 1 AP5 STIFFNESS COEFFICIENT OF THE PAVEMENT')
PRINT 198
198 FORMAT(t0Xt 1 RMSE ROOT MEAN SQUARE OF THE EPROQS BETWEFN')
PRIN'T 199
199 FORMAT(t6Xt 1 MEASUREO AND CALCULATED DEFLECTIONS')
DO 333 J=t ,s
AWCCJ) = OeO
SSWCCJt o.o =
SDWCCJ) t.)eO=
AWCJ) : 0.0
SSW( J) =
0.1)
SOW( J) = OeO
333 CONTINUE
SSSCIC =
OeO
ASCJC = Oe"t
ASCI : OeO
SSSCI OeO=
SDSC I OeO=
AAS5 = OeO
SSAS5 = OeO
SDAS5 OeO=
AAP5 = OeO
SSAP5 = OeO
SOAPS 0.0=
1\VRMS = Oe~
LO = "'
N =0
GO TO 10
'OOn CONTINUE
STOP
END
SUBROUTINE STIF5(Wl,W2tW3,w4,WS,Ot,AZ,AB,QMS,ISW)
IMPLICIT REAL*8(A-H 1 0-Z)

A-12
 0 t NE NS tON A ( 3 0) • X ( 5) • ~( 5 t • R ( 5 t , E ( 5 t , A 0( 5 ) • RMSE ( 30 t
C OMMONI' DI'R • W• N
N =
5
W(l) = Wl
W(2) = W2
we 3t = W3
W(4) = W4
W( 5t = W5
C2 =
6e25
co =
• 8911
Ct =
4. 5029
Rl =
e2709
RUt = too.
R(2) 244e=
R(3) 676e=
R(4) = l396e
R(5) =
2404e
A(1) = el
OL'tA el =
KNT = n
.JNT = 0
11 CONTINUE
00 20 J=2 ,20
20 A(J) =
A(J-l) + DLTA
00 30 .1=1.20
At= 1\(J)
00 40 l=t.N
Tt =
CO I' (Al**Cl)
T2 = 1• I' R( t )
Xl =(Al*DlJ * (At•Ot)
X(tt = l e I ' (~(t)+(C2 *XI))
AO(t) = Tl *
(T2- X(t))
40 CONTINUE
SMXW ()eO=
SIIIIXSQ o.o=
SMESQ = Oet)
DO 50 I=t,N
SMXW = SMXW + X(l) (W(IJ - AOCI)) *
50 SIIIIXSQ = SMXSQ + X(l) X(lt *
B = SMX II I' SMX SO
DO 60 t:t,N
E(tt =
W(l)- ((8 X(l)t + AO(I)) *
60 SMESQ =
SMESQ + ECI) ECI) *
AMSE =
SMESQ I' N
RMSE(J) = OSQRT(AMSE)
30 CONTINUE
K = 20
TEMPt = RMSE(t)
!SUBI = 1
00 St L.:t,K
IF(RMSE(L) eGTe TEMPI) GO TO 51
TEMPt =
RMSE(L)
I SUBt = L
51 CONTINUE
t F ( I SUB 1 • NE • 1 t GO T 0 41
ISW =2
GO TCI 10
*l CONTINUF
I F ( t SUB l • NE • 20 t :; 0 T 0 52
A(l) = 1\(ISUBl - t)
KNT = KNT + t
A-13
 IF(KNT eNEe 3) GO TO 11
ISW J =
GO TO 10
52 CONTINUE
Zt : A(tSUBl - 1)
Z2 =
•(ISUBl + 1)
NOt 14 =
CALL FIBO(NOI,zt,z2,AB,RMS,Ol)
CALL ANS(AB.Ot.A2)
IF(.2 eNEe 2~222e) GO TO 10
tsw =4
t~ CONTINUE
RETURN
ENO
SUBROUTINE FIBO(N,XleX2,X,VeDt)
IMPLICIT REAL*8CA-H,O-Zt
DIMENSION FIBC20)
FIB( t ) =
t •
FIB(2) : 2e
FIBC 3) = 3.
Fl~(4) = 5.
Ft8(5) =
8e
Fl8(6):: 13•
Fl8(7) : 21•
FI8(8) :: 34.e
FIR(9) :: 55e
FIB(10J = A9e
FIB(lt):: t4t\.e
FI8Ct2) :: 2l3e
F 18( 13) : 377 •
FIB(t4) = 610.
FtR(t5) :: 987e
FIB(l6) =
J597e
FIB(t7t =
2584e
FIB(18) : 4.181•
FIB(19) :: 6769e
Fl8(20) =
t0946e
DX=(X2-Xl)/FIB(N)
XL=Xt
X~=X2
N=N-1
X:::XL +FIB ( N) iiDX
CALL FU~C(X,VReDl)
1 N=N-1
X::XL+FIB(N)*DX
CALL FUNC(X,VLeDl)
2 IFCNeEOel) GO TO 4
IFCVLeGTeVR) GO TO 3
XR=XR-FIB(N)•Dx
V~=VL
GO TO 1
3 XL=XL+FIBCN)*DX
VL=VR
N=N-1
X=XR-FIB(N)*DX
CALL FUNC(X,VRe01)
GO TO 2
~IFCVL.GTeVR) GO TO 7
tF(XL.EOeXl) GO TO 6
5 X=XL+DX
Y=VL

A-14
 RETURN
6 CALL FUNCCXteVeDl)
IFCVeGTeVLt GO TO 5
X=Xt
Y=V
RETURN
7 IF(XPeEOeX2) GO TO 9
8 X:XR-OX
Y=VR
RETURN
9 CALL FUNCCX2tVe01)
IF(VeGTeVR) GO TO 8
X=X2
Y=V
RETURN
END
SUBROUTINE ANS(AleD1eA2)
IMPLICIT REAL*8CA-H,O•Zl
DIMENSION ~(30)eX(5)eW(5)eR(5),f(5),~0(5)
COMMON.ID/R eWeN
co = •891 t
Ct = 4e !029
C2 = 6e25
81 = .2709
oo I:t,N
Tl = co .I c~t••ctt
T2 = t • .I PCtt
xt = ( Al*Dl t • CA1*01)
X( I) =
•• •
~0( I) = Tl
•
.I ( R (J) + CC2
(T2 ~ X( I) )
Xl))

t CONTINUE
SMXW = OeO
SMXSQ = OeO
SMF.SQ = 0.0
DO 2 I =1tN
SMXW = SMXW + X(t) * (W(I)- ~0(1))
2 SMXSQ = SMXSQ + XCI) * X(l)
1:t = SMXW .I SMXSO
IFCB eGTe OeO) GO TO 91
A2 = 22222·
GO TO 4
91 CONTINUE
CA = le .I Cl
~2 = ceo .1 B)**CA
~ CONTINUE
RETURN
END
SUBROUTINE FUNCCAteRMS,Dl)
IMPLICIT REAL*8(A-He0-Z)
OIMENSION ~(30)eX(5)eWC5)eR(5),E(5),~0(5)
COMMON/0/P eW tN
co = .8911
C1 = •• 5029
C2 = 6e25
81 = • 2709
DO t I= l, N
Tl = co .I c "t • •ct ,
T2 = t. .I P(l)
XI = C~l*Dl) * (Al*Dlt
X ( t ) = t • .I ( R ( I ) + ( C2 * X 1 ) }
"0( t t = Tl • ( T2 - X( 1) )

A-15
 t CONT tNUE
SMXW =
OeO
SMXSQ =
OeO
SMESQ =
0•0
00 2 t=teN
SMXW = s•xw
+ X( I) * (W( I) - AO( I) t
2 SMXSQ =
s•XSQ + X(l) * X(I)
B =
SMXW / SMXSQ
DO 60 l=l,N
E ( t ) = W( I ) - ( (B • X( I t ) + ~0 ( I))
6~ SMESQ =
SMESQ + E(l) * ECtt
AMSE = SMESQ / N
RMS =
OSORT(AMSEt
RETURN
END

A-16
 NAME DICTIONARY

A Dunmy array used with subroutine CORE to select the correct

input fonnat for each card read
AAP5 Sum of pavement stiffness coefficients
AASS Sum of subgrade stiffness coefficients
AAPSV Average pavement stiffness coefficient
AASSV Average subgrade stiffness coefficient
APS Stiffness coefficient of the pavement
ASS Stiffness coefficient of the subgrade
ASCI Sum of (Wl - W2); Wl - ~12 = Surface Curvature Index for
measured deflections
ASCIC Sum of {Wl ~ H2); Wl - ~12 = Surface Curvature Index for
predicted deflections
ASCIV Average Surface Curvature Index for measured deflections
ASCICV Average Surface Curvature Index for predicted deflections
AW Array of the sums of geophone deflection
AVRMS Average of RMSE's
AWCV Average of the predicted deflections
AWV Average of measured deflections
c Assigned value of .8911
Cl Assigned value of 4.5029
cc Assigned value of £.25
COMM Comments related to the project
CONT SDHPT control number
CORE Subroutine to re-read a card under format control
C0l,C02,C03,C04 County name
D Array of measured geophone readings and accompanying multipliers

A-17
DAP Pavement stiffness coefficient as calculated in subroutine
STIF5
DAS Subgrade stiffness coefficients as calculated in subroutine
STIF5
DATE An IBM subroutine that returns the current month, day,
and year
DAY Day the deflections were measured
DIST District number
DP Total pavement thickness
DYNA Dynaflect number
HWY 1 ' HWY 2 Highway name and number
I Pointer for data read into storage
ICK Switch to indicate last data card in each section
ISW Denotes type of message to be printed for each set of
deflections not considered valid
ISAVSW Array for I SW
IXDATE Return argum~nts for subroutine DATE
(month, day, year)
LAl Description of material in layer 1
LA2 Description of material in layer 2
LA3 Description of material in layer 3
LA4 Description of material in layer 4
LAS Description of material in layer 5
LA6 Description of material in layer 6
LO Counter for data cards not considered
M Month the deflections were taken
MNT Counter to control printing of 20 lines per page
N Counter for number of data cards read

A-18
NCARD Denotes card type
100 = Project identification card
200 = Existing pavement description card {layers 1, 2, 3)
300 = Existing pavement description card {layers 4, 5, 6)
400 =Data card (geophone readings and multipliers)
R Array of distanc~s from load wheels
REM Dummy array to allow reading of standard data cards
SECT SDHPT section number for the highway
STIF5 Subroutine that takes 5 deflections and pavement thickness
and returns with values for Al, A2, RMSE and ISW
SCI Surface Curvature Index {Wl - W2), in mils, for measured
deflections
A A

SCIC Surface Curvature Index (Wl - W2), in mils, for predicted

deflections
SOAPS Standard deviation of the pavement stiffness coefficients
SDAS5 Standard deviation of subgrade stiffness coefficients
SDSCI Standard deviation of the~ SCI's of measured deflections
SDSCI5 Standard deviation of the SCI's of predicted deflections
sow Standard deviation of the measured deflections
SDWC Standard deviation of the predicted deflections
SSAP5 Sum of the APS's squared
SSAS5 Sum· of the ASS's squared.
SSSCI Sum of measured deflections SCI~s squared
SSSCIC Sum of predicted deflections SCI's squared
ssw Sum of measured deflections squared
sswc Sum of predicted deflections squared
STA Station number
Tl Layer 1 thickness
T2 Layer 2 thickness
T3 Layer 3 thickness

A-19
T4 Layer 4 thickness
T5 Layer 5 thickness
T6 Layer 6 thickness
w Array of measured geophone deflections
we Array of predicted deflections
XLANE Traffic lane and direction
YEAR Year the deflections were taken

A-20
 INPUT GUIDE

The data input format for the main computer program is the same as
that used by several previously written computer programs that compute
pavement strength properties from Dynaflect data, namely the Texas State
Department of Highways and Public Transportation stiffness coefficient
program, ELASTIC MODULUS I, and ELASTIC MODULUS II. Each input data card
is read into a storage area and the subroutine CORE is used to select the
read statement and data format to read each data card. Subroutine CORE
allows a FORTRAN program to read under format control from a storage area
which contains alphabetic character codes of a card image. Each data card
has a code punched in the first three columns that designate the card type.
100 - Card that indicates the beginning of data cards for each job
and contains control information about the job, location, date and total
pavement thickness.
200 - Card contains word descriptions and thicknesses of the first
three la.vers of the pavement.
300 - Card contains word descriptions and thicknesses of layers 4, 5
and 6 (if present}.
400 or blank - Card contains station number and geophone deflection
readings and multipliers for each observation. Two digit numbers in
columns 75 and 76 of this card denotes end of data.

A-21
 CARD TYPE 1: PROBLEM IDENTIFICATION

Variable
Name Variable Definition Format Column

DIST SDHPT district number A2 4-5

COl Four characters of county name A4 6-9
C02 Four characters of county name A4 10-13
C03 Four characters of county name A4 14--17
C04 Two characters of county name A2 18-19
CONT SDHPT control number A4 20-23
SECT SDHPT section number A2 24-25
JOB SDHPT job number A2 26-27
HWYl Four characters of highway number A4 28-31
HWY2 Three characters of highway number A3 32-34
XLANE Lane identification A3 35-37
DP Thicknesses of pav_~_ment layers 1-4 4F5.3 38-42
OM Month deflections were measured A2 43-44
DAY Day deflections were measured A2 45-46
YEAR Year deflections were measured A2 47-48
DYNA SDHPT dynalfect number A2 49-50
COMM 1/\11 ',1 /H

A-22
 CARD TYPE 2: LAYER IDENTIFICATION

Variable
Name Variable Definition Format Column

LAl Name of material in pavement layer one 5A4 4-23

Tl Thickness of layer one F4.2 24-27
LA2 Name of material in pavement layer two 5A4 28-47
T2 Thickness of layer two F4.2 48-51
LA3 Name of material in pavement layer three 5A4 52-72
T3 Thickness of layer three F4.2 73-76

CARD TYPE 3: LAYER IDENTIFICATION

Variable
Name Variable Definition Format Column

LA4 Name of material in pavement layer 5A4 4-23

four
T4 Thickness of layer four F4.2 24-27
LA5 Name of material in pavement layer t)A4 ?A-47
fivP.
T5 Thickness of layer five F4.2 48-51
LA6 Name of material in pavement layer 5A4 52-72
six
T6 Thickness of layer six F4 .2 73-76
Note- for a five layer problem, LAS will be the name of the subgrade
material and T5, LA6, and T6 will be blank.

A-23
 CARD TYPE 4: DEFLECTION DATA

Variable
Name Variable Definition Format Column

CONT SDHPT control number A4 4-7

SECT SDHPT section number A2 8-9
DM Month deflections were measured A2 10-11
DAY Day deflections were measured A2 12-13
YEAR Year deflections were measured A2 14-15
STA Station (location on highway where A7 16-22
deflections were measured)
D Dynaflect readings and multipliers 5(F2.1, F3.2) 26-50
R~1K Remarks of data recorder 4A4 59-74
ICK Check field for last data card 12 75-76

A-24
 OUTPUT FORMAT

The output of STIF5 consists of two parts which may include one or
more pages per pavement section (depending on the number of deflection
measurements per section).
The first part of the output is the input data of the identification
information of the pavement section where the surface deflections were
measured. This information includes the SDHPT job, control and section
number of the pavement section, the highway name and number, date the
deflections were measured, total pavement thickness and the materials and
thicknesses of the various layers of the pavement.
The second part of the output is a list of the locations (stations)
where surface deflections were measured on the pavement section, the
observed deflections, followed by the calculated deflections and the
subgrade and pavement stiffness coefficients calculated from the observed
deflections. Averages and standard deviations of the observed deflections,
calculated deflections, and the stiffness coefficients are printed as the
final lines of the printed output.

A-25
 EXAMPLE PROBLEMS

On the following page is a computer printout of the STIFS program

showing the deflection data used and the stiffness coefficients calculated,
to illustrate the utilization of STIF5~

A-26
 TFXAS HIGHWAY DEPAPTMENT
DISTRICT 21 - D~SIGN SECTION
DYNAFLFCT DEFLECTIONS AND CALCULATED STIFFNESS COEFFICIENTS
THIS REVISED PROGRA~ (MARo 74) WAS ~UN 101 J76
DlSTo COUNTY
21 PHAR~

CONTo SECT. JOB HIGHWAY OYN~FLECT

1234 51 l UNKNOWNUNK qq

PAVo THICK. = 32o00 INCHFS

HMAC SUPFACE 1.00 BLACK BASE 4.00
FLEX. BASF 19.00
LIME STAB. SUBGRADE 8e00 SUBGRADE o.o
o.o
LOCATION WI w2 W3 W4 w5 SCI ASS AP~

Sl Oe28S Oo231 Ool56 Oei2J OolOS 0.054

Oo289 Oo2l9 Oo1n6 0.128 0.098 Oo070 0.25 Oo58
52 Oo380 Oo234 Ool92 Oo135 Oo117 Ool46
Oe369 Oo260 Oo184 Ool35 Oo10l Ool09 0.26 o.s~ Oo0150
53 Oo330 Oo234 Oo180 Oo123 Ooll4 Oo096
Oo327 Oo240 Ool76 Ool33 OolOl ~.OA7 0.25 o.~~ o.oo~2

59 Oo350 Oo26l Oo219 Ool71 Ool4l OoCA9

Oo343 Oo275 Oo217 0.170 Ool33 0.067 Oo23 Oo60
510 Oe420 Co310 Oo207 Ool65 Ool50 0~110
Oo417 0.308 Oo225 Oo168 Ool26 0.109 Oo24 o.s~ 0.0137
511 Oo450 0.360 0.~64 Oo225 Oo2Cl 0.090
Oo442 Oo363 Oo287 Oo225 Oel74 Oo080 Co?? 0.5~ OoC164

AVERAGES Oo369 Oo272 Oo203 Oo157 Ool38 Oo097

Oo365 Oo277 Oo209 Ool60 Oo122 Co087 Oo24 0.56
STO DEV Oo060 Oo053 0.037 Oo039 0.035 Oo030
o.os7 o.osz o.o4s o.o37 o.0?9 o.ot9 o.o1 o.o3
NUMBER OF POINTS IN AVERAGES 6
Wl-5 MEASURED DEFLECTIONS AT GEOPHONES 1 0 2.3 0 4 AND 5
c•LCULATEO DF.FLtCTIONS AT GEOPHONES 1o2o3•4 A~D 5
SCI SURFACF CURVATURE IND~X ( Wl MINUS W2)
ASS STIFFNESS COEFFICIENT OF THE SUBGRAOF
AP5 STIFFNESS COEFFICieNT OF THE PAVtMENT
~MSE ROOT MtAN SQUARE OF THE ERROPS RETWE~N
MEASURED AND CALCULATED DEFLECTIONS

A-27
 APPENDIX B

DOCUMENTATION OF A METHOD OF COMPUTING STIFFNESS COEFFICIENTS IN MULTI-

LAYER PAVE~1ENTS, COMPUTER PROGRAM SCIMP.

Page

INTRODUCTION . . • . • . B-2
PROGRAM IDENTIFICATION • . B-3
PROGRAM DESCRIPTION . B-4

FLOWCHART . . . . . B-14

PROGRAf~ LISTING o B-17

NAME DICTIONARY 0 • B-25

INPUT GUIDE . . . B-36

OUTPUT FORMAT . • B- 41

EXAMPLE PROBLEMS . . B-42

B-1
 INTRODUCTION

The Flexible Pavement Design System (FPS) developed in Research Study

2-8-62-32 and currently in use in Texas requires a stiffness parameter of
the materials to be considered as an input to the computer program (1).
Several methods for determining this stiffness parameter (either elastic
modulus or stiffness coefficient) of pavement materials have been reported
previously {2, 4, 5). This report describes a procedure for finding two
unknown stiffness coefficients in an existing pavement consisting of up
to five layers.

8-2
 PROGRAM IDENTIFICATION

Title: Stiffness Coefficients in Multi-layer Pavements (SCIMP)

Language: Fortran IV
Machine: IBM 360/65
Programmer: C.H. Michalak
Availability: Texas Transportation Institute
Pavement Design Program
Texas A&M University
College Station, Texas 77843
Phone (713)845-3735
Date: April, 1975
Source Deck: Approximately 500 cards
Computer Storage: llOk bytes
Timing: (1) Compilation time- 0.40 minutes (Fortran G compiler)
(2) Execution time - 0.25 minutes for 21 data observations
Output: (1) Program list- about 500 lines
(2) Program output- one page per problem (30 lines per page)

OOCU~·1ENTATION OF A ~1ETHOD OF COMPUTING STIFFNESS COEFFICIENTS IN SIMPLE TWO

LAYER PAVEMENT STRUCTURES USING FIVE OBSERVED SURFACE DEFLECTIONS, COMPUTER
PROGRAr·1 ST I F5.

B-3
 PROGRAM DESCRIPTION

For convenience, the computer program used to calculate two unknown

stiffness coefficients in a multi-layer pavement is called SCIMP, for
itiffness foefficients in ~ulti-layer Pavements.
The computer code for SCIMP is written in FORTRAN for an IBM 360/65
computer. It uses a re-read routine to permit transfer of the input data
from a buffer area in core after the appropriate data format has been
determined. Any convenient routine that will "read" data from a buffer
area can be used or the user can bypass this feature. The data input
and the "re-read" feature will be explained more fully on pg. B-10 where
the READIT subroutine is described.
The computer code is written to allow for several problems in succession
to be run by initially zeroing all counters and internal data storage areas
at the start of each new problem. The input data format is common to the
Texas State Department of Hiqhways and Public Transportation stiffness coefficient
program, ELASTIC ~10DULUS I, and ELASTIC r10DULUS II (2,4,5) computer programs.
Subroutine READIT is called from the main program to read each data
card and to compute the observed surface deflections. The deflections
are checked for accuracy (W(I) > W(I+l), W(I) > 0) and are passed to
subroutine STIF5 where they are used to calculate the two unknown stiff-
ness coefficients. Subroutine STIF5 returns an indicator switch which
indicates a valid solution or specifies the condition for which a
solution was not found. If a valid solution is indicated, the predicted
surface deflections are calculated and the appropriate counters are
incremented for determining average values of the observed and predicted
deflection data, surface curvature index, stiffness coefficients, RMsE•s
and the variances of these values. The computer code is repeated until

B-4
all the data cards for a particular problem have been read and the stiff-
ness coefficients for each set of observed surface deflections have been
calculated.
The averages and variances of the observed and predicted surface
deflections, surface curvature index, stiffness coefficients, and RMSE 1 S
are computed and printed and the program either returns to its beginning
to work the next problem or terminates normally.
As many problems as desired can be worked in one run time, with the
only limitation being the amount of computer time specified by the user.

B-5
 SUBROUTINE STIF5

This subroutine creates an array of values of A1 , the unknown pave-

ment stiffness coefficient, from which the Fibonacci search routine
determines the A1 value which gives the minimum RMSE within the accuracy
limits specified.
The code is written so that a range of A1 values from 0.1 to 7.4 in
0.1 increments will be used to calculate a root mean square error, RMSE,
between the observed and the predicted surface deflections.
Subroutine DELTA is called to calculate the fi and Fi terms from
equations (5) and (6) for any A1 value being used and the RMSE is calculated
and stored. After the predetermined number of A1 values have been used to
calculate RMSE's, the smallest RMSE value is found. The A1 values of the
RMSE's which bracket this smallest RMSE are used as the range of values
for a Fibonacci search to find the A1 value that gives the minimum RMSE
within the accuracy limits specified.
Subroutine ANS is then called to calculate the stiffness coefficient
of the subgrade from equation (9). Subroutine STIF5 then returns to the
main program with an indicator switch which tells if the solution is a
valid answer or indicates the conditions for which a solution was not
obtained.

B-6
 SUBROUTINL FIBO

This subroutine utilizes a Fibonacci search technique to determine

the A1 value that gives the minimum RMSE.
Since the absolute minimum RMSE will likely be different from the
smallest RMSE selected from the array of RMSE•s calculated in subroutine
STIF5, the A1 values that bracket this smallest RMSE value are used as the
range for a Fibonacci search to find the true minimum RMSE.
The Fibonacci search technique is a convergence scheme based on the
Fibonacci number series to search for a minimum between two points (16).
The two points that define the range of the search are the A1 values of
the RMsE•s that bracket the smallest RMSE found in subroutine STIF5.

8-7
 SUBROUTINE ANS

The subgrade coefficient, A5 , is calculat~d in this subroutine.

After the A1 value of the pavement coefficient giving the minimum RMSE
is found by the Fibonacci search routine, this value is used in suhroutine
DELTA to calculate the fi and F; terms from equations (5) and (6).
Equation (5) is transformed to solve for 80 as shown below:

5
l:
B0 = i=l 5
l: F.
. 1 1
1=

From equation (9), the subgrade coefficient is calculated after the

following transformation has been made:

(4)

B-8
 SUBROUTINE FUNC

This subroutine is used with the Fibonacci search to calculate t~e

RMSE value for each trial value of A1 .

Subroutine DELTA is called to calculate the f. and F. terms used in
A 1 1
calculating the predicted deflections, wi~ from equations (5) and (6).
80 is calculated from equation (6), except the observed surface deflections
A .
ware used instead of w. With 80 , and fi and Fi terms known, the predicted
deflections are calculated from equation (5). The RMSE between the
observed surface deflections and the predicted surface deflections is
calculated and is tested in the Fibonacci search routine to determine if
the RMSE value obtained with the current A1 value is within the accuracy
limits specifie~.

B-9
 SUBROUTINE READIT

The input data cards are read and the program output headings are
printed in this subroutine.
The main program calls subroutine READIT to read each data card.
The data cards are of four types, according to the code numbers punched
in the first three columns of the data card. The code numbers and card
types are defined below:
100 - the card contains identification information such as county
name, SDHPT control, section, and job number, highway name and
location, date deflections were taken, the individual layer
thicknesses, and the known coefficients
200 - the names of the materials in the fi;rst three pavement layers
are given on this card
300 - the names of the materials in layers four, five, and six (if
present) are given on this card
400- the card contains the dynaflect sensor readings and multipliers
used to calculate the pav~ment surface deflections at an
individual location (the first three columns can also be left
blank on these cards).
Each data card is read into an array or buffer storage area as eighty
individual character variables. The card code is tested and the particular
data input format corresponding to the card code is selected. A subroutine
called CORE is then called to transmit the data from the buffer area
according to the variable names and the data format. Any convenient
routine that will 11
re-read data cards from a buffer area according to a
11

data format can be used instead of subroutine CORE. It is also possible

to avoid using subroutine CORE by having a blank card with the same card

B-10
code preceding the actual data card and reading the data cards directly
into the program. This is more desirable for small amounts of data as
the use of subroutine CORE or any. other re-read routine adds greatly to
the computer time needed to work a problem.
As the header and layer identification cards are read the information
on these cards is printed on the program output. The Dynaflect sensor
readings are converted to the pavement surface deflections and these
values are returned to the main program.

B-11
 SUBROUTINE DELTA

The subroutine ca 1cul ates the fi and Fi terms of equations (5) and

(6) for each trial value of A used.

1

B-12
 FLOWCHART
.illll.
SCIMP
(MAIN)

TIIICREMENT SUHS AND

~UNTERS TO CALCULATE CALCULATE THE RMSE
~w~RAGES AND STANDARD
OJ DEVIATIONS FOR THE TRIAL
__,
I COEFFICIENTS
,f::::.

NO

CALL FIBO TO FIND THE

PAVEMENT COEFFICIENT AT
THE MINIMUM R~1SE

CALL ANS TO CALCULATE

HE SUBGRADE COEFFICIENT
 £.!Jill .Q.UB.

DETERMINE AREA TO
SEARCH FOR MINIMUM RMSE

PERFORM FIBONACCI SEARCH

CALL FUNC TO CALCULATE
AN RMSE FOR EACH Al
VALUE UNTIL MINH4UM
RMSE IS FOUND

OJ
I ~
01

.1lliS

CALL DELTA TO CALCULATE

THE F2 AND f2 VALUES
FOR THE FINAL SOLUTION
Al VALUE

CALCULATE THE RMSE

FOR THE Al AND A2 VALUES
 READ IT

SELECT THE CORRECT

READ STATEMENT AND FORMAT
FOR EACH CARD TYPE

READ AND PRINT THE

PROBLEM IDENTIFICATION
LAYER DESCRIPTION
AND DEFLECTION DATA
FROM EACH CARD TYPE

t:D CALCULATE THE

I
__, DEFLECTIONS FOR EACH
OBSERVATION
0'\

PRINT HEADER INFORMATION

LAYER INFORMATION
AND COLUMN HEADINGS FOR
FOR PROBLEMS REQUIRING
MULTIPLE PAGES OF OUTPUT
 PROGRAM LISTING

c
c
IMPLICIT RE~*8(l-H,~-Z)
c
DIMENSION AW(S), AWC(S), AVGCOF(S), COEF(5), DCOF(3), 0P(4),
* SA0(5), SOW(S), Si>WC(S), SDCOF(5), SMCOF(5)e S$W(5), SSWC(5),
+ SSCOFCS), W(5), WCCS)
DIMENSION AWV(5), AWCV( 5), COF(5)
COMMON /FIBI/ FtBC20t
COMMON /CONS/ c, Cit CC
COMMON /RAO/ R(5)
COMMON /POiNT/ IUNK
c
ROUND( X ,EVEN ) =I liNT( ( X + EVEN * e5 ) / EVF.:N
* * I':VEN
c
RU • = too-.
R( 2) = 244e
RC3) : 1',76e
R ( 4) : t 396e
R ( 5) : 2404•
c = .8911
Cl =4e5029
CC =6e25
F18(1) = leO
FIB( 2) = 2•0
DO l J = 3, 20
FIB(J) = FtB(J-1) • FlB(J-2)
CONTINUE
c
~ CONTINUE
c
N =0
LINES =0
00 l 0 J = t. 5
AW(.J) = OeO
SSW(J) = OeO
AWC( J) = 0 eO
SSWCCJI = o.o
SMCDF(.J, = o.n
SSCOF( J) = OeO
tn CONTINUE
ASC t = o.o
SSSCI = o.o
ssc:;ctc = o • .,
ASCit = Oe')
lVPMS = OeO
c
15 CONTINUE
ClLL RE lDIT ( w. OP, COEF, STA, ICK )
c
c CHECK FOR ERRORS IN THE DATA
c
DO t 7 J = t • 4
IF( W( J) .Ea. n.o ) GO TO 30
[F( W( J) eLTe
l ., CONTINUE
W( J+l ) GO TO 32 ,
IF( W( 5) .ea. n.~ ) GO TO 30
t sw = l
CALL STIF5C w. OP, COEF, RMSE, Isw )
GO TO (100.200,300,400)
• ISW

B-17
1 C!l CONT INU'=
N : N + 1
SAO( t t : 0 eO
00 1 q .J = 2. 5
S~D(J) = CDSQRT(S-DCJ-1))+eOEFCJ-1)*DP(J-1tt••2
lQ CONTINUE
DO 20 K = l , 5
DCOF(KI =
(C/COEF(<) ** Clt
we< K t= o .o
2" CONTINUF.
D 0 26 K = 1 , 5
COF(t) =
DCOF(l) • Clt/R(K) - (t,/(R(Kt + CC * SAJ(l))))
DO 22 .J =
2, 4
eOF(J) =
DeOF(J) * (le/(P(K) + ec * SAD(J)) - (lei(R(K) +
= CC • SAO(J+l ))))
22 C0114TINUI=
COFC5t =
OCOF(5) * (le/(R(K) + CC * s•0(5)))
00 24 J =
t. 5
We(K) = WC(K) + COF(J)
24 CONTINUE
25 CONTINUE
DO 28 I =
l, 5
SMeOF(I) = SMCOF(I) + COEF(I)
SSCOF(I) =
SSCOF(I) + COEF(l) ** 2
'WC(t) = AWCCtt +We([)
SSWe(t) =
SS~CCI) + WC(l)**2
SSW( I t = SS.W ( I t + W( I) * 2 *
AW( I) = AW( I) + W( I)
2B CONTINUF.
'VRMS = AVPMS + RMSE
set = wctt - wc2t
ASCI = ASCI + SC l
SSSCI =
SSSei + SCI ** 2
SeiC = ~CClt - WC(2)
SSSCIC = SSSCIC + SCIC**2
•sere = ~SCIC + SCie
PRJ NT 63, STA, ( W( J), J = 1, 5), SCI
63 FQRMAT(7X 9 A7,tX,6(2X,F5e3tt
PRINT 68, (WC(K)tK=le5)e SCIC, CCOEF(K),K=1,5), RMSE
68 F~PMATCT16,6(F7e3),5(=6.2),Fl3e4,/)
COEF(IUNK) = OeO
C OE F ( 5 ) = 0 • n
L I NE S = L I NE S + 2
GO TO 34
30 CONTINU.E
PRINT 65, STA, W(l)t W(2)
65 FORMAT(7X,A7t1Xt2CF7e3)e2Xt 1 SCl ZERO QP LESS OTHER CALCULATIONS '•
*'OMMITE0 1 , / )
LINES = LINES + 2
GO TO 34
32 CONTINUE
PRINT 66, STA
66 FORMAT(7XeA7t3Xe•ERROR IN DATA 1 e/)
LINES= LINeS+ 2
3. CONTINUE
IF( LINES eLTe 30 ) GO TO 36
LINES = 0
IPRT = 2
C'LL HEADNG( IPRT )
36 CONTINUE
IF( tCK eEOe 0 ) GO TO 15

B-18
C': CALCUL ~TF. ~VE~ ~GES
PN : N
DO ~A M = 1, 5
AVGCOF(Mt : SMCOF(~t/N
AWV(Mt : AW(M)/N
~WCV(MJ : ~WC(Mt/N
38 CONTINUE
ASCtV ASCI / N=
~PMSV : AVRMS / N
ASCtCV = ASCtC / N
I F ( N • LE • 1 J G 0 T 0 4 2
00 4-~ M = l• 5
SDCOF1M) :OSORTCCSSCOF(M)- CN * (~VGCOF(Mt**2)))/(N-1))
SOWC(M) :OSQRT((SSWC(Mt - (N * (AWCV(Mt••~t)t / (N- t ))
SDW(MJ =DSORT((SSW(M)- (N • (AWV(M)**2)JJ / (N- l)t
4') CONTINUE
SOSCIC =OSQRT((SSSCIC- (N * CASCICV**2))) / (N-1))
SDSCt =DSQRT((SSSCI - (N * (ASCIV**2J)J / (N- 1))
PRINT 81tC~WV(J)tJ=lt5t.~SCtV
et FORMAT(/,7X, 1 AVERAGES 1 t6C2X.F5e3))
PRTNT 70, (AWCV(J),J:l,St, ASCICVt (COEF(J),J:t,SJ,ARMSV
70 FQRM~T( T16t 6(2XtF5e3), 5(2XtF4-e2tt 5X, F8.4 / t
PRINT 113tCSDWCJ),J=1t5) eSDSCI
It~ FORMAT(7X, 1 STD DEV '•6(F7e~)J
P R I NT 1 1 4 , ( S 0 WC ( J ) • J =l t 5 ) • S 0 S C 1 C t ( S DC 0 F ( J ) t J = 1 , 5 )
114 FOR~AT( T16t 6(2X,F5.3), 5(2X•F4-e2), 5Xt F8•4 / J
PRINT 82,N
82 FORMAT(7Xt 1 NUMBER OF POINTS IN AVERAGES •,I3)
,.
...
o\ 2 CO.NT! NUE
c
PRINT 91
91 FORM~TC/tOx,•wt-5 ME~SURED DEFLECTIONS AT GEOPHONES 1t2t3t4 1 t
*• ~NO 5')
PRINT 191
191 FQRMAT(16Xt'CALCULATE0 DEFLECTIONS AT GEOPHONES lt2e3t4 ~NO 5 1 )

PRINT 96
96 FORMAT( tOX, 1 SCI SURFACE CURVATURE INDEX ( Wt MIN•,
• •us w2 t • t
PRINT 97
97 FORMATC10X, 1 AS5 STIFFNESS COEFFICIENT OF THE SUBGRADE•t
PQtNT 98
98 FORMAT{10X,•~P5 STIFFNESS COEFFICIENT OF THE P'VEMENT')
PRINT 298
298 FORMAT( lOX, 1 ASS5 STIFFNESS COEFFICIENT OF STABe SUBGRADE•t
PRINT 398
398 FORMAT( lOX, 1 ASB5 STIFFNESS COEFFICIENT OF SUB-BASE')
PRINT 498
4-98 FORMAT( lOX, 1 AB5 STIFFNESS COEFFICIENT OF BASE' )
PRINT 198
198 FOQMAT(tOx.•RMSE ROOT MEAN SQUARE OF THE ERRORS BETWEEN 1 t
PRINT 199
199 FORMAT(16Xt 1 MEASURED AND CALCULATED DEFLECTIONS')
GOTO!§
200 CONTINUE
P R I NT 1 63 , ST ~
163 FORMAT(7X.~7.3X,•MINe RMSE AT AI LESS THAN R~NGE SEARCHED' 1 / )
LINES =
LINES + 2
GO TO 34
300 CONTINUE
PRINT 263, STA

B-1 9
263 FORNAT(7XeA7e3Xe 1 NINe RMSE AT AI GREATER THAN PANGE SEARCHF.D 1 , / )
LINES =
LINES + 2
GO TO 34
400 CONTINUE
PRtNT 3E3e STAe RNSEe COEF(l t
363 FOR~AT(7XeA7,3Xe 1 NINeRMSE( 1 eF8e4e 1 ) INDICATED AT T~E VALUE At= '•
*F4e2•/et7Xe'AT THIS POINT THE VALUE OF A2 CANNOT BE CALCULATED' t
LINF.S =
LINES + 2
GO TO 36
END
SUBROUTINE STIF5( w. DP, COEF, ftMSt ISW )
I NPL I CIT REAL•a (A-He o-z)
DIMENSION A(JOt, AJ(5)• E(5,, RMSEC30), W(5), XCSt • COEF(~)
I
0 t MENS I ON DP ( 4 )
:oMMON/POtNT/IUNK
N = 5
•<t) = .t
DLTA = el
KNT = 0
.INT = I)
00 4 t = 1. 4
IF( COEF(I) eEOe 0.0 ) GO TO 5
4 CONTINUE
5 CONTINUE
IUNK =I
11 CONTINUE
DO 20 .1=2.20
20 A( J) = A(J-1) + OLTA
00 30 J=i. 21)
COEF( JUNK ) = A(J)
CALL DELTA( AO, COEFe OP, X )
SMXW = ')eO
SMXSQ = OeO
SMESQ =
OeO
00 22 t = le N
SMXW: SMXW + X(l) * (W(I)- AO(I))
SMXSQ =
SMXSQ + X(l) * X(l)
22 CONTINUE
B = SMXW / SMXSQ
00 25 t t. N=
E (I) = W( I t - ( ( B *
X (I ) ) + AO( I) )
SNESQ = SMESO + E(l) *
E(t)
25 CONTINUE
AMSE =
SMESQ / N
RNSEC.It : OSORT(AMSEt
30 CONTINUE
K =20
TEMPI= RMSE(t)
t suet t =
DO 35 L = le K
IF(RMSE(Lt eGTe TEMPt) GO TO 35
TFMPJ =
RMSE(L)
1 suet = L
'35 CONTINUE
IF(JSUBt eNEe 1) GO TO 41
tsw = 2
GO TO 80
41 CONTINUE
IF(lSUBl eNEe 20t ~0 TO 52
A ( l ) = A( I SUB 1 - 1 )
KNT = KNT + 1

8~20
 IFCKNT eNEe 3) GO TO 11
tSW :a 3
GO TO 80
52 CONTINUI:
Zt = A( ISUBl - 1 t
Z2: A(ISUB1 + 1)
NOt = 14
CALL FIBO( NOt. zt. Z2, lB• RM$, DP, COEF ,w )
CALL ANS( ABe OPe COEF, A2 , W )
IF( l2 eNEe 22222• ) GO TO 70
tSW :: 4
70 COEF ( 5 t :: A2
80 CONTINUE
COF.F ( IUNK) = ~B
RETURN
END
SUBROUTINE FI~O ( N, X1e X2, X, Vt o,s • W t
IMPLICIT RElL*BCA-HeO-Z)
D I ME NS I ON D( 4 ) • S C5 ) , W( 5)
COMMON /FIBl/ FIA(20t
COMMON/POJNT/IUNK
OX=CX2-X1)/FIB(N)
XL=X1
XR=X2
N=N-1
X=XL+FJB(N)*DX
S(tUNK)=X
CALL FUNC(X,VR,O, S e W)
1 N=N-1
X=XL+FIB(N)*DX
S (I UNK) =X
CALL FUNCCXe VLe O, S • W )
2 IFCNeEOe1) GO TO 4
tFCVLeGTeVR) GO TO 3
XR=XR-FIBCN)*DX
VR=VL
GO TO 1
3 XL=XL+FIB(N)*DX
VL=VR
N=N-1
X=XR-FIB(Nt*OX
S(IUNK):X
CALL FUNCCXe VR • o. s • w )
GO TO 2
• tFCVLeGTeVR) GO TO 7
IFCXLeEOeXl) GO TO 6
5 X=XL+OX
V=VL
RETURN
6 CONTINUE
S( JUNK) =XI
CALL FUNC(Xl, Ve o. s • w )
tF(VeGTeVL) GO TO 5
X=Xt
Y·=V
RETURN
7 IF( Xl:1eEOeX2) GO TO 9
8 X=XR-OX
V:VR
RETURN
9 CONTINUE

B-21
 S( JUNK) =X2
CALL FUNC(X2e V, D, S • W )
IF(VeGTeVR) GO TO 8
)(:)(2
Y:V
RETURN
END
SUBPOUTINE t\NS( t\le DP, COEF, A2t W )
IMPLICIT REAL*8(·-~,0-l)
DIMENSION A(3''H• AJ(SJ, C0EF(5), 0P(4t, W(5J, X(5)
COMMON /CONS/ COt Cl, C2
N =5
CALL DELT~( t\0, COEF, DP, X )
SMXW : Oe 0
SM>tSQ = o.n
SME:~Q = OeO
DO 21l t t, N =
SMX~ = SMXW +XCI) *
(W(I)- AO(JJ)
20 SMXSQ =
SMXSQ + X(l) X(l) *
B =
SMXW / SMXSQ
IF(B .GT. OeOJ GO TO 30
A2 =
22222.
GO TO 40
3~ CONTINUE
C t\ =
t • / Ct
A2 =
cco / at••c•
4!) CONTINUE
RETURN
END
SUBROUTINE FUNC( Alt RMS, DP, COEF, W )
IMPLICIT REAL*6CA-H,Q-Z)
OtMENSION A(30), AO(S)t COEFCS),\ OP(4), E(St, W(5), X(5)
N : 5
CALL DELTt\( t\0, COEF, DP, X )
sMx.w = o. o
SMXSQ o. 0 =
SMFSQ = Oe 0
DO 20 I It N =
SMXW: SMXW + X(IJ * (W(I) - AO(I)t
20 SMXSQ = SMXSQ + X(l) X(l) *
B : SMXW / SMXSQ
DO 61) J-:t,N
E(l) = WCIJ- ( ( 8 *
X(J)) + AO(I))
6n SMESQ = SMESQ + E(l) E(IJ *
AMSE = SMESQ / N
RMS·= DSQRT(AMSE)
RETURN
END
SUBROUTINE READIT( We OPe COEF, STA, ICK )
I MPL IC IT REt\L*6( A-H,O-Z)
DIMENSION A(20t,COEF(5t.COMM(7),D(10J,DP(4),LAt(5),LA2(5)eLA~(5),
1LA4(5)tLA5(5)eLA6(5),REM(4), W(5)
COMMON/HEAD/ COleC02,C03eC04tCONT,DAY, OIST,OM, DYNA,HWYt,
* HWY2e SECT, XLANE, YEAR, JOB
c
t CONTINUE

~EADC5e1234tEND=l000) NCAR~, ( A(l), I =1 , 20 J

1234 FORMAT( 13, l9A4,Al )
C CALL CORE ( A, 90 )
tFCNCARDeEOelOOJ GO TO It

B-22
 _tFCNC~ROeEOe2~0) GO TO 12
IF(NCAROeF.Qe3~0t GO TO 13
14 CONT TNUE
RE~0(5,6)
CONT,SECT,OM,DAY,YEAP,ST~ ,(O(K),K:t,t,),
* (RE'M(.Jt,.J:t,4), ICk
6 FORMAT(3Xe ~4,4A2tA7t3X,5(F2eltF3e2),8X,4A4,12)
L =t
D04J=t,5
W(J) = D(L)*D(L+tt
L = L + 2
4 CONTINUE:
RETUPN
11 REt\0(5,2) OtST,CQt,C02tC03tC04,CONTtSECT,.JOB,HWY1,HWY2,XLANE,
•o~,DAY,YFAR,DYNA,(COMM(t),t=1t7), (OP(J)t J:l,4),(COEF(J),J=t,5)
2 FOPMAT(3X, t\2, 3A4tA2tA4tA2tA2eA4tA3tA3t SX, 4A2,7A4/35Xt qF5t2 )
tP~T =1
C~LL HE~DNG(IPRT )
PRINT58,(0P(.J),COEF(.J),.J=1,4),COEF(5)
58 FORMAT(8X, 1 PAVe'tllX, 1 BASE 1 ,8X,•SUBBASE 1 ,6X,•STAB SUB'e8X,•SUBGRAO
*E'/ 7Xt 4( 1 THICK :o~F, 1 ), ' COF.Fe 1 /4X,4(FRe2tF6e2), F8e2/)
61 FORMAT(/,7X, LOC~TION 1
Wl W2 W3 W4 W5 SCI
•AS5 ASS5 ASB5 ABS AP5 RMSE • / )
GO TO 1
: RE~O & PRINT INFORMATION ON DATA CARD 2
12 READ(5,3) (LA1(1),t:t,S),Tt,(LA2(IJ,t=l,5),T2,
* (LA3(1)tl=t,5t.T3
3 FORMAT(3X, 5A4,F4e2t5A4,F4e2t5A4tF4e2t 5A4t F4e2
PRINT 59,(Lt\1(t),t=t,5),T1,(LA2CI)tl=l•5),T2,(LA3(1),t=l•5), T3
5~ FORMAT(16X, 5A4,1X,F5e2t5X,5A4,1X,F5e2t SX, 5A4, IX, F5e2 / )
GI'J TO 1
C READ & PRINT INFORMATION ON DATA CARD 3 9 IF PRESENT
13 READ(5,J) (Lt\4(1),1=1,5),T4,(LA5(1)ti=1,5),T5,
* (LA6(l)tl=1•5)t T6
P R I NT 59 , CL A4 ( I ) , I = 1 , 5) , T 4 , ( LA 5 ( I ) , I = 1 , 5 ) , T 5 , ( LA 6 ( I ) , I = \ , 5 ) , T 6
PRINT 61
GO TO 1
1000 CONTINUE
PRINT 250
250 FORMAT( 1 1 1
STOP
END
SUBROUTINE DELTA(AO,COEF,DP,X)
IMPLICIT REAL *8 (A- H, 0- Z t
DIMENSION ~0(5), COEF(S), DELCS,S), OP(4), SUM(5), X(5)
COMMON /CONS/ C?t Cl, C2
COMMON /RAD/ R(5)
N = 5
no 2 K = 1 • ...
SUM(Kt = o.,
DO I L = 1 • K
SUM(K) = SUM(K) + COEF(L) * DP(L)
CONTINUE
SUM(K) :;:: SUM(K)
2 CONTINUE
•
SUM(I<)

00 4 I :: 1 • N
AO( I ) o. 0 =
DELCitl) =
CO/CCOEF(l)**Ct) *
Cle/R.Cl)- 'ei'(R(I) • C2 * SUM(t•H
A0 ( I ) = AO ( I ) + DEL ( J, t )
DO 3 K = 2t 4
OEL(J,K) =
(CO/(COEF(K))**Clt * Cle/(R(J) + C2 * SJM(K-1))-
= le/(R(I) + C2 * SUM(K)))

B-23
 AO( I t =
AQ(I J + DEL ( t, K )
3 C'lNTINUE
X(l) =
le/(R(IJ + C2 * SUM(A))
4 CONTINUE
QF.TURN
E"'D
SUB~OUTINE HEADNG( IPQT
IMPLICI,. RtAL*8 (A-H,O-Z)
COMMON/HEAD/ COleC02eC03eC04eCONT,OAY• DIST,OM, DYNA,HWY1,
* HWY2e SE=CT, XL,NE, 'lEAP, JOB
DIMENSION IXOATE(2)
GO TO ( 50, 70), JPRT
5') CONTINUE
J:)~JNT 51
5t F QO M .T ( •1 I
D~tNT 52
52 FORMAT(35Xe 1 TEX~S HIGHWAY DEPARTMFNT 1 /}

PRTNT 5~eOIST
53 FQRM~T(33Xe'01STRICT 1 eA2,• - DESIGN SECTIO~' /)
PRINT 54
54 FOPMAT ( 18X, 1 0YNAFLECT DEFLECT IONS AND CALCULATED •,
*'STIFFNESS COEFFICIENTS'/)
CALL DATECIXD~TE)
: PRINT 55eiXOATE
c 55 FQP~AT(24Xt 1 THIS Q!VISED PPOGRAM (APR. 75) WAS PUN 1 t2A4/)
PQTNT 56eOJSTtCOte:02,C03,C04
56 FORMAT( T35e 1 DIST• COUNTY 1 /T36eA2e9Xe3A4,A2 I)
PRINT 57, CONTeSECT~JOBeHWYleHWY2eXLANE,DM,DAV,YE~R,DVNA
57 FOPMAT(T19e 1 CONTe SECT. JOB HIGHWAY DATE•,
* ' DYNAFLECT• / Tl9eA4e5XeA2e5XeA2e2XeA4e2A3e2XtA2t 1 ••,A2,
• • - 1 ,A2e7XeA2,/.
RETURN
7("! CONTINUE
PQTNT 5t
PPINT 56eOISTeC01eC02eC03,C04
PRINT 57,CONTeSECT,JOBeHWY1eHWY2eXLANEeOM,DAY,YEAQ,OYNA
Pt:H NT l;t
61 FO~MAT(/,7X, 1 LOCATION ill W2 W3 WA ws SCI
*ASS AS$5 ~585 ~85 APS RMSE ' / )
RF.TURN
END

B-24
 NAME DICTIONARY
MAIN PROGRAM VARIABLES

ARMSV - Average RMSE value

ASCI - Sum of SCI (surface curvature index, Wl-W2) of the observed
surface deflections in mils
ASCIC - Sum of SCI of the calculated surface deflections, mils
ASCICV - Average SCI of calculated surface deflections, mils
ASCIV - Average SCI of measured surface deflections, mils
AVGCOF - Array of the average stiffness coefficient values
AVRMS - Sum of RMSE values
AW - Array of sums of the observed surface deflections, mils
AWC - Array of sums of the calculated surface deflections, mils
AWCV - Array of average values of the calculated surface deflections, mils
AWV - Array of average values of the observed surface deflections, mils
C A constant in the deflection equation, value 0.8911
CC - A constant in the deflection equation, value 6.25
Cl - A constant in the deflection equation, value 4.5029
COEF - Array of the known and unknown stiffness coefficients
COF - Array containing the individual terms of the deflection equation
used to calculate the pavement surface deflections
c
DCOF - A part of the deflection equation defined as A·gl
1
DP - Array of the pavement layer thicknesses, inches
FIB - Array of the Fibonacci numbers used in the Fibonacci search routine
ICK - A non-zero number punched on the last data card to signal the end
of the data for a problem
IPRT - Indicates which output headings are to be printed by subroutine HEADING
ISW - Indicates a valid solution for each set of surface deflections, or is
used to select a message to be printed to indicate why a solution was
not obtained
 IUNK - The subscript in the array COEF of the unknown pavement coefficient
LINES - Counter to control number of lines printed on a page
N - The total number of data points for which valid solutions were
obtained
NLAY - The total number of layers in the pavement design (subgrade counts
as one layer)
NL 1 - Number of pavement layers above the subgrade
R - Array of the squared distances of the dynaflect sensors from
the dynaflect load, inches
RMSE - Root mean square error between the observed surface deflections,
Wi and the calculated surface deflections ~i
SAD -Array of the squared terms (Ai*Di) used in the deflection equation
to obtain the calculated surface deflections
SCI - Surface curvature index, W1-W2, in mils
SOC OF - Array of the standard deviations of the stiffness coefficients
SDSCI - Standard deviation of the surface curvature index of the observed
surface deflections
SDSCIC - Standard deviati~on of the surface curvature index of the calculated
surface deflections
SOW - Standard deviation of the observed surface deflections
SDWC - Standard deviation of the calculated surface deflections
SMCOF - Array of the sums of the stiffness coefficients
SSCOF - Array of the sums of the squared values of the stiffness coefficients
SSSCI - Sum of the squared values of surface curvature index of the observed
surface deflections
SSSCIC - Sum of the squared values of surface curvature index of the calculated
surface deflections
STA - Identifies location of each set of surface pavement deflection
observations
SSW - Array of the squared values of the observed surface deflection<;
SSWC - Array of the squared values of the calculated surface deflections
W - Array of the values of the observed surface deflections, mils
WC - Array of the values to the calculated surface deflections, mils

B-26
 SUBROUTINE STIFS VARIABLES

A - Array of initial values of the unknown pavement stiffness coefficient

AB - The value of the unknown pavement stiffness coefficient that gives
the minimum RMSE within the accuracy limits specified
AO - Array of the fi terms for each trial value of the unknown stiffness
coefficient, (equation 5)
A2 - The value of the unknown subgrade stiffness coefficient calculated
from the final solution value for the unknown pavement stiffness
coefficient, AB
AMSE - The root mean square er.ror (RMSE) for a trial value of the unknown
pavement stiffness coefficient
B - A trial value of B0 (equation 6)

COEF - Array of the known and unknown stiffness coefficients

DLTA - Incremental value used in selecting the initial values of the unknown
pavement stiffness coefficient
DP - Array containing values of the pavement layer thicknesses, in inches
A ,
E -The error (Wi-Wi) between the observed surface deflections and the
calculated surface deflections
ISUBl - Pointer to the smallest RMSE of the initial set of unknown pavement
stiffness coefficient values tried
ISW Indicates a valid solution for each set of surface deflections or is
used to select a message to be printed to indicate why a solution
was not obtained
IUNK - The subscript in the COEF array of the unknown pavement coefficient
K - Counter to control the number of trial values of the unknown stiffness
coefficient to be tried
KNT - Counter to control the number of times the Fibonacci search is used
before the search is abandoned
N - Number of surface pavement deflections observed at each measurement
location
NLAY - The number of layers (including the subgrade) in the pavement section
Nll - The number of pavement layers above the subgrade in the pavement
section

B-27
NOl - Determines the level of precision of the answer found by the
Fibonacci search
RMSE - Root mean square error between the observed and calculated surface
deflections for a trial value of the unknown pavement stiffness
coefficient
A
SMESQ -Sum of the squared values of (Wi-W;)
SMXSQ - Sum of the f.*(\~.- F.1 ) va 1ues squared
1 1

SMXW - Sum of the f.*(W.-F.)

1 1 1
values
TEMPl - Contains the value of the smallest RMSE of the initial values of the
unknown stiffness coefficient tried
w - Array of the observed surface deflections, mils
X - Array of the F; values for each trial value of the unknown stiffness
coefficient, (equation 5)
Zl - Value of the RMSE just preceeding the smallest RMSE
Z2 - Value of the RMSE just following the smallest RMSE

B-28
 SUBROUTINE FIBO VARIASLES

D - Array of the pavement layer depths, in inches

DX - The incremental amount to change the unknown stiffness coefficient
in the Fibonacci search to find the minimum RMSE
FIB - An array of Fibonacci numbers used in the Fibonacci search routine
IUNK - Pointer to the unknown pavement stiffness coefficient in the COEF
array
N - Determines the level of precision of the answer found by the Fibonacci
search
NLAY Total number of layers in the pavement section, including the subgrade
Nll - Number of pavement layers above the subgrade
S - Array of the known and unknown stiffness coefficients
V - An RMSE calculated from a value of the unknown stiffness coefficient
selected in the Fibonacci search
VL - The RMSE value from the Fibonacci search when a value of the unknown
stiffness coefficient just smaller than the current value of the
unknown stiffness coefficient is tried
VR - The RMSE value from the Fibonacci search when a value of the unknown
stiffness coefficient just larger than the current value of the
unknown stiffness coefficient is tried
W - Array of the observed surface deflections, in mils
X - The value of the unknown stiffness coefficient selected by the
Fibonacci search
XL - The lower limit of values of the unknown stiffness coefficient to
be searched
XR - The upper limit of values of the unknown stiffness coefficient to
be searched
Xl - The lower limit of values of the unknown stiffness coefficient to
be searched
X2 The upper limit of values of the unknown stiffness coefficient to
be searched
Y - An RMSE calculated from a value of the unknown stiffness coefficient
selected in the Fibonacci search

B-29
 SUBROUTINE ANS VARIABLES

AO -Array of the Fi terms (equation [5]) for the solution value of the
unknown pavement stiffness coefficient
A2 - The value of the unknown subgrade stiffness coefficient calculated
from the final solution value for the unknown pavement stiffness
coefficient
B - The final solution value of B0 after the unknown pavement stiffness
coefficient has been found
CA -A constant, value is 1./4.5029
CO A constant, value is 0.8911
C2 - A constant, value is 6.25
COEF - Array of the known stiffness coefficients, including the final
solution value of the unknown pavement stiffness coefficient
DP - Array of the pavement layer thicknesses, in inches
N - Number of pavement surface deflections observed at each measurement
point
A
SMESQ - Sum of the squared va 1ues of ~i ~Wi)

SMXW - Sum of fi*(Wi-Fi) values

W - Array of the observed surface deflections, in mils
X - Array of the Fi values, (equation 5)

B-30
 SUBROUTINE FUNC VARIABLES

AO -Array of the ~ terms (equation [5]) for any value of the unknown
stiffness coefficient
AMSE - The root mean square error (RMSE) for a trial value of the unknown
pavement stiffness coefficient
B -A trial value of B0 (equation [6])
COEF - Array of the known and unknown stiffness coefficients
DP - Array of the pavement layer thicknesses, in inches
A
E -The error (Wi-Wi) between the observed surface deflections and the
calculated surface deflections
N - Number of surface pavement deflections observed at each measurement
point
RMS - The root mean square error (RMSE) for a trial value of the unknown
stiffness coefficient
A
SMESQ - Sum of the squared values of (Wi-W;)
SMXSQ -Sum of the squared values of Fi*(Wi-Ki)
SMXW - Sum of the f;*(W;-Fi) values
W - Array of the observed surface deflections, mils
X - Array of the fi terms, (equation 5)

B-31
 SUBROUTINE READIT VARIABLES

A - Array for storing each card image as individual character variables

for the in core transfer of data under format control routine
(Subroutine CORE)
COl, C02, C03, C04 - A fourteen column field for the county name
COEF - Array of the known stiffness coefficients with the unknown stiffness
coefficients left blank or entered as zeroes
COMM - A twenty-eight column field for any comments concerning the deflections
CONT - The SDHPT control number
D - Array of the dynaflect sensor readings and multipliers for calculating
the surface deflections, in mils
DAY - Day of month the deflections were measured
DIST - The SDHPT district number
OM - Month the deflections were measured
DP - Array of the pavement layer thicknesses, in inches
DYNA - The SDHPT dynaflect number
HWYl, HWY2 - A seven column field for the highway name
ICK - A non zero value to indicate the last data card for each problem
IPRT - Indicates which output headings are to be printed in subroutine
HEADING
IUNK Pointer to the unknown pavement stiffness coefficient in the COEF array
JOB - The SDHPT job number
LAl - Name of material in first pavement layer
LA2 Name of material in second pavement 1ayer
LA3 Name of material in third pavement layer
LA4 Name of material in fourth pavement layer
LAS Name of material in fifth pavement layer
LA6 Name of material in sixth pavement layer
NCARD - Code indicating type of card, header, layer identification, or
deflection data

B-32
NLAY Total number of pavement layers (including the subgrade) in the
pavement section
Nll - Number of pavement layers above the subgrade
REM - A sixteen column field on each deflection data card for appropriate
remarks
SECT - The SDHPT section number
STA - Identifies location of each set of surface pavement deflection
observations
Tl - Thickness of pavement layer one, in inches
T2 Thickness of.pavement layer two, in inches
T3 - Thickness of pavement layer three, in inches
T4 - Thickness of pavement layer four, in inches
T5 - Thickness of pavement layer five, in inches
T6 - Thickness of pavement layer six, in inches
W - Array of the pavement surface deflection~, in mils
XLANE Identifies the lane the deflections were measured in
YEAR - Year the deflections were measured

B-33
 SUBROUTINE DELTA VARIABLES

AO -Array of the F; terms (equation [5]) for any value of the unknown
stiffness coefficient
COEF - Array of the known and unknown stiffness coefficients
CO - A constant, value of 0.8911
Cl - A constant, value of 4.5029
C2 - A constant, value of 6.25
DEL - Array of the terms used in the calculation of the Fi terms,
(equation 5)
DP - Array of the pavement layer thicknesses, in inches
IUNK Pointer to the unknown pavement stiffness coefficient in the COEF
array
N - Number of surface pavement deflections observed at each measurement
locatio·n
NLAY - Total number of layers (including the subgrade) in the pavement section
NLl - Total number of pavement layers above the subgrade
R - Array of the squared distances of the dynaflect sensors from the
dynaflect load, inches
SUM - Array of Ai*Di terms used in calculating the Fi terms, (equation 5)
X -Array of the fi terms (equation [5]) for any value of the unknown
stiffness coefficient used

B- 34
 SUBROUTINE HEADNG VARIABLES

COl, C02, C03, C04 - A fourteen column field containing the county name
CONT - The SDHPT control number
DAY - Day of the month the deflections were measured
DIST - The SDHPT district number
DM - The month the deflections were measured
DYNA - The SDHPT dynaflect number
HWYl, HWY2 - A seven column field for the highway name
IPRT - Indicates which output headings are to be printed
IXDATE - The date the problem was run
JOB - The SDHPT job number
SECT - The SDHPT section number
XLANE - Identifies the lane the deflections were measured in
YEAR - The year the deflections were measured

B-35
 INPUT GUIDE

The SCIMP computer program is written to solve one or more problems in

each run of the program. The input data for each problem is a set of data
cards containing information pertinent to each problem. Each data set is
made up of the card types shown in the input guide tables. Each card type
is identified by the card code punched in the first three columns of the card.
The card types and the information contained on them are listed below:
Card type 1 contains problem identification information and has card
code 100 punched in the first three columns. The identification information
consists of the SDHPT district number, county name, SDHPT control t section and
job number, the highway number, the lane direction, the month, day and year
the deflections were measured, the SDHPT dynaflect number, and any comments by
the data recorder. Additional information contained on a (;econd card is thP
pavement layer thicknesses and the known values of stiffness coefficients.
Although card type 1 consists of two cards, the information is read as though
it were contained on one card.
The second card type has code number 200 punched in the first three
columns. It contains the names and thicknesses of the materials in the first
three pavement layers.
Card type 3 has code 300 punched in the first three columns. It contains
the names and thicknesses of the materials in pavement layers four, five, and
six (if these are present).
Card type 4 has code 400 punched in the first three columns, or the card
code field can be left blank if desired. This card contains the SDHPT control
and section numbers, the month, day, and the year the deflections were
measured, the station number (location on the highway) where the deflections
were measured, the dynaflect geophone readings and multipliers, and any

B-36
remarks listed by the ~ata recorder. The card containing the last data
observation for each problem has a check field which is coded with any
non-blank characters or digits to signal the end of data for each problem.
The normal data input deck for each problem will be one card type 1,
one card type 2, one card type 3 and N card types 4, where N is the number
of surface deflection observations for the problem. The Nth or last card
type 4 will have the check field coded -to signal the end of data for the
problem.

B-37
 CARD TYPE 1: PROBLEM IDENTitiCATION

Variable
Name Variable Definition Format Column

DIST SDHPT district number A2 4-5

COl Four characters of county name A4 6-9
C02 Four characters of county name A4 10-13
C03 Four characters of county name A4 14-17
C04 Two characters of county name A2 18-19
CONT SDHPT control number A4 20-23
SECT SDHPT section number A2 24-25
JOB SDHPT job number A2 26-27
HWYl Four characters of highway number A4 28-31
H\~Y2 Three characters of highway number A3 32-34
XLANE Lane identification A3 35-37
DM Month deflections were measured A2 43-44
DAY Day deflections were measured A2 45-46
YEAR Year deflections were measured A2 47-48
DYNA SDHPT dynalfect number A2 49-50
COMM Conments of data recorder 7A4 51-78

SECOND CARD

DP Thicknesses of pavement layers 1-4 4F5.2 36-55

COEF Values of the known stiffness 5F5.2 56-75
coefficients (zeroes or blanks
indicate unknown values)

B-38
 CARD TYPE 2: LAYER IDENTIFICATION

Variable
Name Variable Definition Format Column

LAl Name of material in pavement layer one SA4 4-23

Tl Thickness of layer one F4.2 24-27
LA2 Name of material in pavement layer two SA4 28-47
T2 Thickness of layer two F4.2 48-Sl
LA3 Name of material in pavement layer three SA4 52-72
T3 Thickness of layer three F4.2 73-76

CARD TYPE 3: LAYER IDENTIFICATION

Variable
Name Variable Definition Format Column

LA4 Name of material in pavement layer 5A4 4-23

four
T4 Thickness of layer four F4.2 24-27
LAS Name of material in pavement layer 5A4 28-47
five
TS Thickness of layer five F4.2 48-51
LA6 Name of material in pavement layer SA4 S2-72
six
T6 Thickness of layer six F4.2 73-76
Note- for a five layer problem, LAS will be the name of the subgrade
material and TS, LA6, and T6 will be blank.

B- 39
 CARD TYPE 4: DEFLECTION DATA

Variable
Name Variable Definition Format Column

CO~T SDHPT control number A4 4-7

SECT SDHPT section number A2 8-9
DM Month deflections were measured A2 10-11
DAY Day deflections were measured A2 12-13
YEAR Year deflections were measured A2 14-15
STA Station (location on highway where A7 16-22
deflections were measured)
D Dynaflect readings and multipliers 5( F2. 1 , F3. 2) 26-50
RMK Remarks of data recorder 4A4 59-74
ICK Check field for last data card 12 75--76

B- 40
 OUTPUT FORMAT

The printed output of program SCIMP consists of two parts, the

identification information of the pavement section and the deflection
data (both observed and calculated values).
The identification information includes the SDHPT job, control and
section numbers, the highway name and number, county and district number,
the thic~ness of all pavement layers, the stiffness coefficient values,
and the materials in each pavement layer.
lhe deflection data is listed by location or station where each
deflection measurement was made and consists of the observed surface
deflections followed by the calculated deflections and the stiffness
coefficients (the known and calculated values).
The averages and standard deviations of the deflections (observed
and calculated) and the stiffness coefficients are printed as the last
line of the output.

B-41
 EXAMPLE PROBLEMS

On the following page is a computer printout of the SCIMP program showing

the deflection data used and the stiffness coefficients calculated to
illustrate the utilization of SCIMP.

B-42
 TEXAS HIGHWAY DEPARTMFNT
DISTRICT 21 • DESIGN SECTION
DYNAFLECT DEFLECTI~NS AND CALCULATED STIFFNESS COEFFICIENTS
Disr. CO UN TV
21 PHARR
CONTe SECT. JOB HIGHWAY DVNAFLECT
1234 51 1 UNKNOWNUNK 99
PAY. BASE. SUBBASE STAB SUB SUBGRADE
THICK COEF. THICK COEF. THICK COEF. THICK COEF. COEF.
s.oo 0.75 4.oo o.o 19.00 o.•o a. oo o.3o o.o
H MAC SURF ACE t.oo BLACK BASE 4.00 FLEXe BASE 19.00
LIME STAA. SUBGAADE e.oo SUBGAADE o.o o.o

LOCATION WI W2 Wl W4 W5 SCI AP5 AB5 ASB5 ASS5 AS5 RMSE

51 Oe285 Oe23l o.t56 Oel23 Oe105 Oe054
Oe277 0.236 Oe171 Oel2l Oe087 0.041 0.75 1.11 0~40 0,30 Oe27 o.ottt
52 Oe380 Oe234 Oe192 Oe135 Oel17 Oe146
0.354 0.282 Oe188 Oel27 Oe088 0.012 Oe75 le30 Oe40 0.30 0.21 0.0277
s3 o.330 o.234 o.teo o.t23 o.~t4 o.o96
Oe313 0.260 Oel81 Oe126 0.089 Oe05J Oe75 le5l 0.40 0.30 0.27 0.0175
59 0.350 0.261 0.219 0.171 0.141 0.089
o.J27 o.291 0·224 o.t67 o.t25 o.o36 o.75 t.92 o.4o o.Jo 0.24 0.0183
s1o o.420 Oe310 0.207 o.165 o.t5o o.tto
Oe403 0.330 0·229 o.t58 Oell2 0•073 0.75 lell 0.40 Oe30 0.26 0.0228
sa1 o.450 o.36o o.2&4 0.225 0.201 o.o9o
Oe426 Oe379 Oe295 Oe220 Oel65 Oe047 Oe75 1e81 Oe40 Oe30 Oe23 Oe025?

AVERAGES Oe369 Oe272 Oe203 Oe157 Oel3R 0.09~

0.3~0 Oe296 Oe215 Oel53 0.111 Oe054 Oe75 o.o 0.40 0.30 0,0 0.0204
STD DEY 0.060 Oe053 Oe037 o. 039 o. 035 o.o.Jo
0.056 Oe052 0.046 0.038 OeOJO o.Ot6 o.o Oe26 o. 00 o.oo
NUNBER OF POINTS IN AVFPAGES 6
WI -5 MEASURED DEFLECTIONS AT GEOPHONES 1e2e3t4 AND 5
CALCULATED DEFLECTIONS AT GEOPHONES lt2elo4 AND 5
SCI SURFACE CURVATURE INDEX ( WI MINUS W2) .
ASS STIFFNESS COEFFICIENT OF TH~ SUBGRADE
AP5 STIFFNESS COEFFICIENT OF THE PAVEMENT
ASS5 STIFFNESS COEFFICIENT OF STAB. SUBGRADE
ASB5 STIFFNESS COEFFICIENT OF SUB-BASE
AB5 STIFFNESS COEFFICIENT OF BASt
IUIISE ROOT MEAN SQUARE OF THE ERRORS BETWEEN
MEASURED AND CALCULATED DEFLECTIONS

B-43