R

Note

Marked set by R

Expansive Soils

BALKEMA - Proceedings and Monographs

in Engineering, Water and Earth Sciences

Expansive Soils

Recent advances in characterizationand treatment

EditorsAmer Ali Al-RawasDepartment of Civil and Architectural Engineering,College of Engineering, Sultan Qaboos University,Sultanate of Oman

Mattheus F.A. GoosenSchool of Science and Technology, University of Turabo,Puerto Rico, USA

LONDON / LEIDEN / NEW YORK / PHILADELPHIA / SINGAPORE

© 2006 Taylor & Francis Group, London, UK

All rights reserved. No part of this publication or theinformation contained herein may be reproduced, stored in a retrieval system, or transmitted in any form or by any means,electronic or mechanical, by photocopying, recording or otherwise,without written prior permission from the publishers.

Although all care is taken to ensure integrity and the quality of this publication and the information herein, no responsibility is assumed by the publishers nor the author for any damage to property or persons as a result of operation or use of this publication and/or the information contained herein.

Published by: Taylor & Francis/BalkemaP.O. Box 447, 2300 AK Leiden,The Netherlandse-mail: Pub.NL@tandf.co.ukwww.balkema.nl, www.tandf.co.uk, www.crcpress.com

British Library Cataloguing in Publication DataA catalogue record for this book is available from the British Library

Library of Congress Cataloging in Publication DataExpansive soils: recent advances in characterization and treatment /

editors: Amer Ali Al-Rawas, Mattheus F. A. Goosen.p. cm.

Includes index.1. Soil consolidation. 2. Swelling soils. I. Al-Rawas, Amer Ali.

II. Goosen, Mattheus F.A.

TE210.4.E96 2006624.1'5136–dc22 2005035532

ISBN10 0–415–39681–6 ISBN13 978–0–415–39681–3

This edition published in the Taylor & Francis e-Library, 2006.

“To purchase your own copy of this or any of Taylor & Francis or Routledge’scollection of thousands of eBooks please go to www.eBookstore.tandf.co.uk.”

Contents

List of contributors ixPreface xi

PART 1Nature, identification, and classification of expansive soils 1

1 Geology, classification, and distribution of expansive soils and rocks: a case study from the Arabian Gulf 3AMER A. AL-RAWAS, MATTHEUS F.A. GOOSEN, AND GHAZI A. AL-RAWAS

2 Identification and classification of expansive soils 15SUDHAKAR M. RAO

3 Prediction and classification of expansive clay soils 25AGUS SETYO MUNTOHAR

4 Overview of mineralogy of bentonites: genesis, physicochemical properties, industrial uses, and world production 37RICHARD PRIKRYL

5 Swelling in non-vertisolic soils: its causes and importance 55MIGUEL ANGEL TABOADA AND RAÚL SILVIO LAVADO

PART 2Volume change characteristics 79

6 ESEM study of structural modifications of argillite during hydration/dehydration cycles 81JOËLLE DUPLAY, GERMAN MONTES-HERNANDEZ, AND LUIS MARTINEZ

7 Large-scale odometer for assessing swelling and consolidation behavior of Al-Qatif clay 85SHAHID AZAM

8 Water sorption and dilatation of bentonites and montmorillonite-rich clays 101RADEK HANUS, IRENA KOLARÍKOVÁ, AND RICHARD PRIKRYL

PART 3Swelling potential measurement 115

9 ESEM–DIA method to estimate swelling–shrinkage of raw and cation-saturated bentonite 117GERMAN MONTES-HERNANDEZ

10 Effect of remolding techniques on soil swelling and shear strength properties 127MOUSA F. ATTOM, MAJED M. ABU-ZREIG, AND

MOHAMMED TALEB OBAIDAT

11 Swelling rate of expansive clay soils 139ROSLAN HASHIM AND AGUS SETYO MUNTOHAR

12 Swelling behavior of Ankara Clay: predictive techniques,damage details, and swelling maps 149ZEYNAL ABIDDIN ERGULER AND RESAT ULUSAY

13 Prediction of swelling characteristics with free swell index 173BHYRAVAJJULA R. PHANIKUMAR

PART 4Advanced techniques for swelling potential assessment 185

14 Remote sensing of expansive soils: use of hyperspectral methodology for clay mapping and hazard assessment 187SABINE CHABRILLAT AND ALEXANDER F.H. GOETZ

15 Spectroscopy as a tool for studying swelling soils 211PATRICK CHEGE KARIUKI, KEITH SHEPHERD, AND

FREEK VAN DER MEER

16 Finite element analysis of piers in expansive soils 231YAHIA E.-A. MOHAMEDZEIN

vi Contents

17 Prediction of swelling pressure of expansive soils using Neural Networks 245YAHIA E.-A. MOHAMEDZEIN, RABAB IBRAHIM, AND ASSIM ALSANOSI

18 Shrinkage strain characterization of expansive soils using digital imaging technology 257ANAND J. PUPPALA, SIVA PATHIVADA, VENKAT BHADRIRAJU, AND

LAUREANO R. HOYOS

PART 5Site characterization 271

19 Swelling behavior of expansive shale: a case study from Saudi Arabia 273ABDULLAH I . AL-MHAIDIB

20 Volume change characteristics of compacted Ankara clay 289ERDAL COKCA AND OZLEM CORA

21 Influence of trees on expansive soils in southern Australia 295DONALD A. CAMERON, MARK B. JAKSA, WAYNE POTTER, AND

AARON O’MALLEY

PART 6Lime stabilization 315

22 Stabilization of expansive Ankara Clay with lime 317MEHMET CELAL TONOZ, CANDAN GOKCEOGLU, AND RESAT ULUSAY

23 Lime stabilization of expansive clay 341ZALIHE NALBANTOGLU

24 Combined lime and polypropylene fiber stabilization for modification of expansive soils 349ANAND J. PUPPALA, EKARIN WATTANASANTICHAROEN, AND ALI PORBAHA

PART 7Cement-stabilization 369

25 Assessment of anisotropic behavior of swelling soils on ground and construction work 371EVANGELOS I . STAVRIDAKIS

Contents vii

viii Contents

26 Stabilization of problematic soils using cement and lime 385EVANGELOS I . STAVRIDAKIS

27 Influence of sand content on strength and durability of cement-acrylic resin treated soil 399COSTAS A. ANAGNOSTOPOULOS

28 Physical and engineering properties of cement stabilized soft soil treated with acrylic resin additive 405COSTAS A. ANAGNOSTOPOULOS

PART 8Other treatment methods 417

29 Pozzolanic stabilization of expansive soils 419P.V. SIVAPULLAIAH

30 Swelling characteristics and improvement of expansive soil with rice husk ash 435AGUS SETYO MUNTOHAR

31 Effects of addition of fly ash on swell potential of an expansive soil 453DEVRIM TURKER AND ERDAL COKCA

32 Dynamic characterization of chemically modified expansive soil 465LAUREANO R. HOYOS, PHONLAWUT CHAINUWAT, AND ANAND J. PUPPALA

33 Assessment of seasonal effects on engineering behavior of chemically treated sulfate-rich expansive clay 483LAUREANO R. HOYOS, ARTHIT LAIKRAM, AND ANAND J. PUPPALA

PART 9Construction techniques and remedial measures 505

34 Granular pile-anchors: an innovative foundation technique for expansive soils 507BHYRAVAJJULA R. PHANIKUMAR AND RADHEY S. SHARMA

Index 523

Contributors

Majed M. Abuzreig, Jordan University of Science and Technology, Irbid, Jordan

Abdullah I. Al-Mhaidib, King Saud University, Riyadh, Saudi Arabia

Amer Ali Al-Rawas, Sultan Qaboos University, Al-Khoud, Sultanate of Oman

Ghazi A. Al-Rawas, Sultan Qaboos University, Sultanate of Oman

Assim Alsanosi, University of Khartoum, Khartoum, Sudan

Costas A. Anagnostopoulos, Aristotle University of Thessaloniki, Thessalonica, Greece

Mousa F. Attom, Jordan University of Science and Technology, Irbid, Jordan

Shahid Azam, University of British Columbia, Vancouver, Canada

Venkat Bhadriraju, University of Texas at Arlington, USA

Donald A. Cameron, University of South Australia, Australia

Sabine Chabrillat, GeoForschungsZentrum (GFZ) Potsdam, Germany

Phonlawut Chainuwat, PSA Engineering, Texas, USA

Erdal Cokca, Middle East Technical University, Ankara, Turkey

Ozlem Cora, Middle East Technical University, Ankara, Turkey

Joelle Duplay, Centre de Géochimie de la Surface, Strasbourg, France

Zeynal Abiddin Erguler, Hacettepe University, Ankara, Turkey

Alexander F.H. Goetz, University of Colorado, USA

Candan Gokceoglu, Hacettepe University, Ankara, Turkey

Mattheus F.A. Goosen, University of Turabo, Gurabo, Puerto Rico

Radek Hanus, Charles University, Prague, Czech Republic

Roslan Hashim, University of Malaya, Kuala Lumpur, Malaysia

Laureano R. Hoyos, University of Texas at Arlington, USA

Rabab Ibrahim, Al-Amin Engineering Company, Khartoum, Sudan

Mark B. Jaksa, University of Adelaide, Australia

x Contributors

Patrick Chege Kariuki, International Livestock Research Institute (ILRI), Kenya

Irena Kolaríková, Charles University, Prague, Czech Republic

Arthit Laikram, University of Texas at Arlington, USA

Raúl Silvio Lavado, Universidad de Buenos Aires, Argentina

Luis Martinez, Universite Henri Poincare, Nancy, France

Freek van der Meer, Delft University of Technology, Delft, The Netherlands

Yahia E.-A. Mohamedzein, Sultan Qaboos University, Al-Khoud, Sultanate of Oman

German Montes-Hernandez, Centre de Géochimie de la Surface, Strasbourg, France

Agus Setyo Muntohar, Muhammadiyah University of Yogyakarta, Indonesia

Zalihe Nalbantoglu, Eastern Mediterranean University, Gazimagusa, Mersin 10, Turkey

Mohammed T. Obaidat, Jordan University of Science and Technology, Irbid, Jordan

Aaron O’Malley, University of South Australia, Australia

Siva Pathivada, University of Texas at Arlington, USA

Bhyravajjula R. Phanikumar, GMR Institute of Technology, India

Ali Porbaha, California Department of Transportation, USA

Wayne Potter, University of South Australia, Australia

Richard Prikryl, Charles University, Prague, Czech Republic

Anand J. Puppala, University of Texas at Arlington, USA

Sudhakar M. Rao, Indian Institute of Science, Bangalore, India

Radhey S. Sharma, Louisiana State University, USA

Keith Shepherd, World Agroforestry Centre (ICRAF), Kenya

P.V. Sivapullaiah, Indian Institute of Science, Bangalore, India

Evangelos I. Stavridakis, Aristotle University of Thessaloniki, Greece

Miguel Angel Taboada, Universidad de Buenos Aires, Argentina

Mehmet Celal Tonoz, Hacettepe University, Ankara, Turkey

Devrim Turker, Middle East Technical University, Ankara, Turkey

Resat Ulusay, Hacettepe University, Ankara, Turkey

Ekarin Wattanasanticharoen, University of Texas at Arlington, USA

Preface

Expansive soils are a worldwide problem. The estimated damage to buildings, roads, andother structures built on expansive soils, for example, exceeds 15 billion dollars in the USannually. Such soils are considered natural hazards that pose challenges to civil engineers,construction firms, and owners. In some underdeveloped countries, buildings wereconstructed without any knowledge of the presence of expansive soils. This was in partdue to a lack of historical evidence. With the rapid development in urban infrastructure,expansive soil problems have become more evident. There is therefore a need to address theproblems associated with these soils.

Expansive soils occur in many parts of the world but particularly in arid and semi-aridregions. In these regions, evaporation rates are higher than the annual rainfall so that thereis almost always a moisture deficiency in the soil. The addition of water will cause groundheave in soils possessing swelling potential. Semi-arid regions are characterized by shortperiods of rainfall followed by long periods of draught causing cyclic swelling and shrinkingphenomena. The ground heave that results from soil swelling potential is a multifactorialphenomenon that involves a combination of the type of material, type and amount of clayminerals, microfabric, initial moisture content, and initial dry density.

Considerable research has been reported on expansive soils over the past three decades.The last international conference on expansive soils was held in Dallas, Texas, USA in1992. The 6th International Conference on Expansive Soils was held in New Delhi inJanuary 1988. Several textbooks on expansive soils are also available: Foundations onExpansive Soils by Chen, F.H., Elsevier 1988; Expansive Soils: Problems and Practice inFoundation and Pavement Engineering by Nelson, J.D. and Miller, D.J., John Wiley &Sons, Inc. 1992; Construction of Buildings on Expansive Soils by Sorochan, E.A.,Aa Balkema January 1991; and Behaviour of Saturated Expansive Soil and ControlMethods – Revised and Enlarged Edition by Katti, R.K./Katti, D.R./Katti, A.R., Routledge2002. Since the most recent comprehensive publication is several years old, a book is nowneeded that updates the state-of-the-art knowledge in this area.

This book provides a broad coverage of recent advances in the characteristics and treat-ment of expansive soils. There are nine parts each with specific chapters. It starts with anoverview section (Part 1) on the nature, identification, and classification of expansive soils.Parts 2 and 3 deal with volume change characteristics and swelling potential measurements,respectively. Part 4 covers advanced techniques for swelling potential assessment. Such testsare important for assessing the actual swelling potential of the soil and estimating groundheave. Part 5 on site characterization presents field measurements of soil swelling potentialand suction. The next three parts deal with lime stabilization, cement stabilization, and

xii Preface

other treatment methods. Chemical stabilization, for example, has gained wide attention asa successful technique for treating expansive soils. In the final section (Part 9), the performanceof engineering structures built on expansive soils such as buildings, houses, embankments,and roads, is evaluated. Remedial measures used to address soil swelling problems are alsodescribed.

The intended audience for this book includes researchers, practicing engineers, contractors,postgraduate and undergraduate students, and others working in expansive soils. The authorshope that the information provided in this book will help to promote a better understandingof expansive soils, contribute toward their treatment, and thereby reducing or minimizingtheir effects. The views expressed in the chapters of this book are those of the authors anddo not necessarily reflect those of their respective institutions. The authors hope that thisbook will contribute to the advancement in research in expansive soils and help engineersin the development of practical solutions to expansive soil problems.

Amer Ali Al-RawasSultan Qaboos University, Al-Khoud, Sultanate of Oman

Mattheus F.A. GoosenUniversity of Turabo, Gurabo, Puerto Rico

2005

Part 1

Nature, identification, andclassification of expansivesoils

Summary

This chapter deals with the prediction and classification of the degree of expansiveness ofclay soil. Statistics analysis was introduced as a simple technique for identifying andpredicting the degree of swelling. There were three properties which were most stronglycorrelated to swelling potential (i.e. plasticity index, liquid limit, and clay fraction). Ingeneral, the models in the current study showed good correlation compared with previousmodels cited in the literature. The multiple linear regression model gave the best-fit for allsoil conditions.

Introduction

Expansive soils are a world wide problem (Seed et al., 1962; Kormonik and David, 1969;Al-Rawas et al., 1998; Alawaji, 1999; Cokca, 2001; Erguler and Ulusay, 2003; Muntoharand Hashim, 2003). Principally, swelling occurs when water infiltrates between the clayparticles, causing them to separate. Several attempts have been made by researchers toobtain time-swell relationships for expansive soils. Some progress has been made towardcharacterizing swelling characteristics, despite the complexity of the behavior. Seed et al.(1962) reported that the time required for completion of swelling is relatively long.

Many tests and methods have been developed for estimating shrink–swell potential. Theseinclude both indirect and direct measurements. Indirect methods involve the use ofsoil properties and classification schemes to estimate shrink–swell potential. Direct methodsprovide actual physical measurements of swelling. Several laboratory methods havebeen developed to directly determine the swelling a soil undergoes as the moisture con-tent changes. These include free swell, expansion index, consolidation-swelling, CaliforniaBearing Ratio (CBR), potential volume change (PVC), and coefficient of linear extensibility(COLE).

Currently, no one method of soil analysis estimates the shrink–swell potential accuratelyfor all soils. Soil scientists recognize that shrink–swell behavior can be best predicted byexamining a combination of physical, chemical, and mineralogical soil properties.Determining these properties and establishing a shrink–swell model that can be extrapolated

Chapter 3

Prediction and classification ofexpansive clay soils

Agus Setyo Muntohar1

1 Department of Civil Engineering, Muhammadiyah University of Yogyakarta, Building F1, 3rd floor. Jl. RingroadSelatan, Taman Tirto, Yogyakarta, Indonesia. 55183. Email: muntohar@umy.ac.id

across the same or similar parent materials is needed. Some researchers consider that thisswelling potential can be linked to a single parameter.

This chapter deals with the prediction and classification of the degree of expansivenessof clay soil. Statistics analysis is introduced as a simple technique for identifying andpredicting the degree of swelling.

Potential of volume change

Holtz and Gibbs (1956), Altmeyer (1955), Seed et al. (1962), and Daksanamurthy andRaman (1973) have evolved different methods to identify expansive soils based on thepercentage of clay content, shrinkage limits (both volumetric and linear), plasticity index,liquid limit, and shrinkage index. Accordingly, they classified soils into low, medium, high,and very high degrees of potential expansiveness (Figures 3.1, 3.2, and 3.3). However, as

26 A.S. Muntohar

0

20

40

60

80

100

0 20 40 60 80 100 120 140

Liquid limits, LL (%)

Plas

ticity

inde

x, P

I (%

)

MidHigh

Low

Very high

Non

Swelling potential

Figure 3.1 Chart for potential expansiveness of soil.

Source: Daksanamurthy and Raman, 1973.

0

20

40

60

80

100

0 10 20 30 40 50 60 70

Clay content (%)

Plas

ticity

inde

x (%

)

Very highHigh

Medium

Low

Low

A = 0.5

A = 2.0Swelling potential

Figure 3.2 Potential expansiveness of expansive soil.

Source: Williams, 1957.

with most soil systems, the activity classification scheme does not accurately estimateshrink–swell potential in mixed mineralogy soils. Parker et al. (1977) found that the activityindex was too imprecise for both mixed and montmorillonitic mineralogy soils to be useful.However, Schreiner (1988) observed consistent trends in soil and bentonite/sand mixturesusing the activity index as an indicator of shrink–swell potential.

The classification of potential expansiveness does not give the same assessment of theswelling potential. It cannot conclude precisely the degree of volume change for particularsoils as presented, for example, in Table 3.1. Seed et al. (1962) have also correlated theswelling potential with the degree of expansion values used by USBR as presented inTable 3.2. The boundaries defining these ranges are plotted in Figure 3.4.

Indirect estimation of swelling parameters

In view of the difference that has been observed between the directly measured values of theswelling parameters and the values output by the earlier models, the first idea was to fit themodels in question. The literature contains a considerable number of empirical techniquesfor assessing the swelling potential of soils, which correlated with consistency limits,moisture content, dry density, and depth of the soil samples (Seed et al., 1962; Chen, 1983;

Classification of expansive clay soils 27

Plasticity index (%)

Colloid content (% of = 0.001 mm)

Shrinkage limit (%)

020 40 0 20 40 0 8 16 24

8

16

24

32

40

Vol

ume

chan

ge (

%)

Air

dry

to

satu

rate

d co

nditi

on u

nder

1 p

si lo

ad

Ver

y hi

ghH

igh

Med

ium

Low

Figure 3.3 Relation of volume change to colloid content, plasticity index, and shrinkage limit.

Source: Holtz and Gibbs, 1956.

Table 3.1 Classifications for degree of expansion (swelling potential)

Degree of Chen (1983) Seed et al. Daksanamurthy USBR (Holtz andexpansion (1962) and Raman (1973) Gibbs, 1956)

Very high LL � 60 PI �35 LL � 70 CC � 28High 40 � LL 60 20 � PI 35 50 � LL 70 20 � CC 31Medium 30 LL 40 10 PI 20 35 � LL 50 13 CC 23Low LL � 30 �10 20 LL 35 CC � 13

28 A.S. Muntohar

Basma et al., 1995; Djedid, 2001). Thomas (1998) proposed an expansive soil rating system,termed as the Expansive Soil Index (ESI). The model was developed as a function of usingthe soil properties most correlated with shrink–swell potential such as the ratio 2:1 betweensmectite and vermiculite minerals, swell index, liquid limit, and cation exchange capacity(CEC). The model gave expansive soil potential ratings (ESI) for each soil series.

Seed’s model (Seed et al., 1962) and Chen’s model (Chen, 1983) are very simple. They usedplasticity index parameters. Their models are given by Equations 3.1 and 3.2, respectively.

SP � 60K (PI)2.44 (3.1)

SP � B eA(PI) (3.2)

Act

ivity

0 10 20 30 5040 60 70 80

Percent Cloy Size (finer than 0.002 mm)

Swelling Potential = 50%

5

4

3

2

1

0

Swelling Potential = 20%Swelling Potential = 10%Swelling Potential = 5%Swelling Potential = 1%

90 100

Figure 3.4 Classification chart for swelling potential.

Source: Seed et al., 1962.

Table 3.2 Classification of degree of expansion

Degree of expansion Swelling potential (%)

Very high �25High 5–25Medium 1.5–5Low 0–1.5

where, SP is swelling potential plasticity index. Figure 3.5 shows the correlations betweenswelling potential and plasticity index that given by some researchers. K � 3.6 � 105,A � 0.0838, B � 0.2558 are constants, and PI is plasticity index.

Data analysis

Data used in this study consisted of 115 pairs and was compiled from many references (Seedet al., 1962; Kormonik and David, 1969; Al-Rawas et al., 1998; Alawaji, 1999; Cokca, 2001;Erguler and Ulusay, 2003; Muntohar and Hashim, 2003). Table 3.3 presents the data source

Classification of expansive clay soils 29

10

9

8

7

6

5

4

3

2

1

0

Swel

ling

pote

ntia

l (%

)

5 10 15 20 25 30 35 40

Plasticity index (%)

Chen (Surcharge pressure 5.94 psi)

Holtz & Gibbs (Surcharge pressure 1 psi)

Seed, Woodward & Lundgren (Surcharge pressure 1 psi)

1

1 2

3 Chen (Surcharge pressure 1 psi)

4

2

3

4

Figure 3.5 Correlations between swelling potential and plasticity index.

Source: Reproduced from Chen, 1983.

Table 3.3 Number of data used for model

Source of data Number of data

Current research data 7Alawaji (1999a) 10Attom et al. (2001) 3Çokça (2001) 1Zeynal and Ulusay (2003) 20Seed et al. (1962) 12USBR (quoted by Seed 28et al., 1962)

Total data 81

Table 3.4 Proposed empirical model for predicting swelling potential

No. Models Regression Regressionstatistics ANOVA

1 Variable: plasticity index (PI) R2 � 0.58 df � 1(a) SP � 1.035(PI)0.816 Ad. R2 � 0.57 F � 109.24

S� � 7.85 Pv � 0.0001(b) SP � 10.106e0.056(PI) R2 � 0.444 df � 1

Ad. R2 � 0.44 F � 63.12S� � 9.04 Pv � 0.0001

(c) SP � 2.231 � 0.453 (PI) R2 � 0.563 df � 1Ad. R2 � 0.56 F � 102.08S� � 9.04 Pv � 0.0001

2 Variable: clay fraction (CF) R2 � 0.226 df � 1(a) SP � 2.919(CF)0.535 Ad. R2 � 0.22 F � 23.08

S� � 10.66 Pv � 0.0001(b) SP � 11.418e0.0135(CF) R2 � 0.152 df � 1

Ad. R2 � 0.15 F � 14.22S� � 11.16 Pv � 0.0001

(c) SP � 7.518 � 0.323(CF) R2 � 0.192 df � 1Ad. R2 � 0.18 F � 18.77S� � 10.89 Pv � 0.0001

3 Variable: liquid limit (LL) R2 � 0.546 df � 1(a) SP � 0.109(LL)1.236 Ad. R2 � 0.54 F � 95.03

S� � 8.16 Pv � 0.0001(b) SP � 6.871e0.0149(LL) R2 � 0.466 df � 1

Ad. R2 � 0.46 F � 69.10S� � 8.85 Pv � 0.0001

(c) SP � 0.393(LL) � 6.298 R2 � 0.56 df � 1Ad. R2 � 0.55 F � 100.53S� � 8.04 Pv � 0.0001

4 Multiple linear regression: R2 � 0.613 df � 3SP (%) � 0.171CF � 0.0012LL R2 Adj. � 0.60 F � 40.608

� 0.409PI � 1.869S� � 7.64 Pv � 0.0001

NotesCoefficient of confidence level (�) � 0.05; SP is swelling potential (%); S�: Standard error.

that was used in the study. Preliminary statistics test was carried out for screening the variablesused in the models. The variables, which only had good correlation with swelling potential,were chosen as independent variables. They were plasticity index (PI), liquid limit (LL), clayfraction (CF), dry density (�d), and water content (w). Due to the large variability of drydensity and water content data, both variables were rejected as independent variables.

Data analysis was considered in two stages (i.e. learning and validating). As much as81 data samples were randomly used for formulating the empirical model in the learningstage. The other data was used for validating. The two most common empirical models,linear and nonlinear, were fitted to the data using a single independent variable. These weredeveloped using SigmaPlot Ver 6.1. Multiple independent variables or multiple linearregression were also established for developing empirical models to indicate reliable assess-ment of swelling potential of a soil. The general models are given as follows:

● Linear (single): y � b0 � b1t● Power: y � b0tb1

● Exponential: y � b0e(b1t)

● Multiple linear: y � b0 � b1t1 � b2t2 � b3t3

The results of the statistical analysis are presented in Table 3.4.

Discussion

Empirical models

The evaluation of swell behavior of a soil using undisturbed samples and specialized swelltests is a difficult and expensive process for practicing engineers and small builders.Therefore, there is a need for simple routine tests that can be performed on disturbedengineered samples to achieve the same purposes. The empirical models appearing in theliterature are primarily related to prediction of swelling and swelling pressure from indexproperties of soils. Sometimes, the empirical models proposed cannot be applied appropriatelyto all soils due to different soil conditions and testing procedures.

The data used, here, was compiled from different determinations of swelling and indexproperties. It was hoped that the models would be acceptable and generalized. Figure 3.6shows the correlation between predicted-swelling and actual (measured) swelling. The fig-ure plotted all the data used (i.e. 115 data samples were used in learning and validating). Thedashed line shows the correlation between measured and predicted swell. It was expectedthat correlation should have lain on the 45 line (1 : 1 line), which refers to the colinearityof model. The figure illustrates that proposed empirical models, given in Table 3.3, givegood correlations. The dash lines in Figures 3.6a, 3.6b, and 3.6c were laid down in the col-inearity range of 0.5–0.8. The empirical model, proposed by Seed et al. (1962), as shown inFigure 3.6d, showed a very weak correlation in which the correlation was below line 0.5.It indicated that the proposed equation by Seed et al. (1962) is only appropriate for ameasured-swelling of less than 30%. In our study, the multiple linear regression method(Equation 3.4, in Table 3.4) indicated a best-fit correlation. In general, the model can beused for all soil conditions.

In the current study, multiple regression analysis was considered to derive an equationthat can be used to predict swelling potential from several index and physical properties. The

Classification of expansive clay soils 31

0

10

20

30

40

50

0 10 20 30 40 50 0 10 20 30 40 50Predicted swelling (%)

0 10 20 30 40 50Predicted swelling (%)

0 10 20 30 40 50Predicted swelling (%)

Act

ual s

wel

ling

(%)

Eq. 1(a)

0.5

0.81:1

0

10

20

30

40

50

Predicted swelling (%)

Act

ual s

wel

ling

(%)

Eq. 3(a)

0.5

0.81:1

0

10

20

30

40

50

Act

ual s

wel

ling

(%)

Eq. 4

0.5

0.8

1:1

0

10

20

30

40

50

Act

ual s

wel

ling

(%)

Seed et al. (1962)1:1

0.5

0.8

(a) (b)

(c) (d)

10

20

30

40

0 10 20 30 40 50Predicted swelling (%)

0 10 20 30 40 50Predicted swelling (%)

0 10 20 30 40 50Predicted swelling (%)

0 10 20 30 40 50Predicted swelling (%)

Eq. 1(c)

0.5

0.81:1

0

10

20

30

40

50 Eq. 1(b)

0.5

0.81:1

0

10

20

30

40

50 Eq. 2(a)

0.5

0.81:1

0

10

20

30

40

50 Eq. 2(b)

0.5

0.81:1

0

50

Act

ual s

wel

ling

(%)

Act

ual s

wel

ling

(%)

Act

ual s

wel

ling

(%)

Act

ual s

wel

ling

(%)

(e) (f)

(g) (h)

Figure 3.6 Correlation of proposed empirical model and actual swelling.

use of multiple regression statistics is very important in reducing the number of variablesthat are considered to be an independent source of information. These variables are reducedto only 3 or 4 which adequately explains the variation in swelling properties of the soils.

Many trials were carried out to correlate the swelling parameters to a combination ofvariables. The test of hypothesis of a linear model involved testing for significance of regres-sion and testing on individual regression coefficients (Montgomery, 2001). For the proposedmodel, since P-value (Pv) is considerably smaller than the confidence level (� � 0.05), thenull hypothesis (H0: �1 � �2) was rejected, indicating a strong correlation between eachvariable. For all statistical models, single or multiple linear regressions, satisfactorilyfulfilled the F-test.

The coefficient of determination (R2) has been used as a global statistic to assess the fitof the model. However, this value will increase when a regressor is added. In this model, theR2 for a single variable is 0.192 (Equation 2c in Table 3.4). When added with two other vari-ables it increases to 0.60. It showed significance. It can be concluded that swelling potentialis linearly related to CF, LL, or PI. Testing of individual regression coefficient requires thatat least one of the variables contributes significantly to the model. It has been observed thatthe CF and PI imply a significant contribution, since the standard error was 0.0661and 0.141respectively, and the P-value was less than � � 0.05 (Table 3.5). The coefficients of vari-ables lie in the range of 95% confidence level. The overall test indicated that the variablesfulfilled the requirements of the t-test and F-test.

Classification degree of swell

The empirical models have indicated that the multiple linear regression model gives thebest-fit correlation for prediction of swelling potential of expansive soil. Furthermore, qual-itative measurement is also needed to classify the degree of swelling. The measurement wasdetermined based on the normal probability plot as shown in Figure 3.7. The measurementwas simply divided into four regions based on the 25% percentile (Quartile), 50% percentile(Mean), and 75% percentile data. The classification of degree of expansiveness (swelling)is presented in Table 3.6. The determination is quite high compared to the category that wasgiven by USBR in Table 3.2.

Classification of expansive clay soils 33

0

10

20

30

40

50

0 25 50 75 100Sample percentile

Swel

ling

pote

ntia

l (%

)

Very high

High

Medium

Low

Figure 3.7 Normal probability plot and qualitative measurement of swelling potential.

It can be noted that the models and classification devised in this study are a simplepredictive tool for assessing swell potential of a given soil, both undisturbed and remouldedspecimens. Using the proposed model (Table 3.5), the swelling potential of the soil used inthis study can be predicted and then classified as presented in Table 3.7.

Conclusions

Index and physical properties of soil are useful indicators to estimate engineering andswelling properties. There are three properties, which are most strongly correlated to swellingpotential, PI, LL, and CF. The proposed models in the current study, showed good correlationcompared with previous models cited in the literature. The multiple linear regression modelgave the best-fit for all soil conditions. The classification of degree of expansiveness(i.e. swelling) has been well devised based on the statistical analysis. The degree of swellingcan be classified into four distinct levels, low, medium, high, and very high.

34 A.S. Muntohar

Table 3.5 Analysis of variance (ANOVA) multiple linear regression

Variables Coefficients Standard error t-stat P-value Lower 95% Upper 95%

Intercept �1.8695 3.1612 �0.5913 0.5559 �8.1644 4.4253CF 0.1707 0.0661 2.5804 0.0117 0.0389 0.3023LL 0.00124 0.1320 0.00947 0.9925 �0.2616 0.2640PI 0.4092 0.1406 2.9105 0.0047 0.1292 0.6890

Table 3.6 Category for swelling potential classification

Proposed model USBR* Expansiveness remarks

SP � 8.68 SP � 1.5 Low8.68 SP 15.1 1.5 SP � 5 Medium15.1 SP 28.8 5 SP 15 HighSP � 28.8 SP � 25 Very high

NotesSP: swell potential (%); * Seed, et al. (1962).

Table 3.7 Predicted swelling potential and classification of soil used in the study

Soil code Clay fraction Liquid limit (LL) Plasticity index (PI) Predicted swelling Remarks(CF)% % % potential (SP)%

KB1 26.9 76.9 37.5 18.2 HighKB2 30.0 89.7 47.5 22.8 HighKB3 32.5 106.8 62.4 29.3 VeryKB4 39.0 121.5 78.4 37.0 VerySB1 4.0 42.9 21.8 7.8 LowSB2 21.7 85.1 57.9 25.6 HighSB3 47.0 138.3 95.1 45.2 Very

Acknowledgments

The author gratefully appreciates the funding provided by the Ministry of Science,Technology, and Environmental (MOSTE) Malaysia through Intensify Research for PriorityArea (IRPA) RM#8 and the University of Malaya through Short-Term IRPA Fund (Vot-F)2002/2003. Sincere thanks go to Ir. Dr Roslan Hashim, Professor of University of Malaya,for his discussion.

References

Alawaji, H.A. (1999), Prediction of swell characteristics of sand – bentonite mixtures, Proceedings of11th Asian Regional Conference on Soil Mechanic and Geotechnical Engineering, Vol. 1, Seoul,pp. 35–37.

Al-Rawas, A.A., Guba, I., and McGown, A. (1998), Geological and engineering characteristics ofexpansive soils and rock in northern Oman, Engineering Geology, Vol. 50, pp. 267–281.

Altmeyer, W.T. (1955), Discussion of engineering properties of expansive clays, Transaction ASCE,Vol. 81 (658), pp. 17–19.

Basma, A.A., Al-Hamoud, A.S., and Malkawi, A.H. (1995), Laboratory assessment of swellingpressure of expansive soils, Applied Clay Sciences, Vol. 9, pp. 355–368.

Chen, F.H. (1983), Foundation on Expansive Soils, Elsevier Scientific Publishing Co., New York, USA.Çokça, E. (2001), Use of class C-Fly Ashes for the stabilization of an expansive soils, Journal of

Geotechnical and Geoenvironmental Engineering, ASCE, Vol. 127 (7), pp. 568–573.Daksanamurthy, V. and Raman, V. (1973), A simple method of identiying an expansive soil, Soil and

Foundations, Japanese Society of Soil Mechanic and Foundation Engineering, Vol. 13 (1),pp. 97–104.

Djedid, A., Bekkouche, A., and Aissa Mamoune, S.M. (2001), Identification et prévision du gonflementde quelques sols de la région de Tlemcen, Algérie, Bulletin des Laboratories des Ponts etChaussées, Vol. 233, pp. 69–77.

Erguler, Z.A. and Ulusay, E. (2003), A simple test and predictive models for assessing swell potentialof Ankara (Turkey) clay, Engineering Geology, Vol. 67, pp. 331–352.

Holtz, W.G. and Gibbs, H.J. (1956), Engineering properties of expansive clays, Transaction ASCEVol. 121, pp. 641–677.

Kormonik, A. and David, D. (1969), Prediction of swelling pressure of clays, Journal of SoilMechanics and Foundation Division ASCE, Vol. 95 (SM1), pp. 209–225.

Montgomery, D.C. (2001), Design and Analysis of Experiments, 5th Edn., John Wiley & Son’s Inc.,New York, USA.

Muntohar, A.S. and Hashim, R. (2003), Swelling Behavior of Engineered Clay Soil, TheSecond International Conference on Advances in Soft Soil Engineering and Technology, 2–4 July,Kuala Lumpur, Malaysia.

Parker, J.C., Amos, D.F., and Kaster, D.L. (1977), An evaluation of several methods of estimating soilvolume change, Soil Science Society American Journal, Vol. 41, pp. 1059–1064.

Schreiner, H.D. (1988), Identification and classification of expansive soils, In Varma, C.V.J. (ed.),Proceeding of 6th International Conference on Expansive Soils, New Delhi, India, pp. 23–29.

Seed, H.B., Woodward, R.J., and Lundgren, R. (1962), Prediction of swelling potential for compactedclays, Journal of Soil Mechanics and Foundation Division ASCE, Vol. 88 (SM3), pp. 53–87.

Thomas, P.J. (1998), Quantifying Properties and Variability of Expansive Soils in Selected Map Units,Ph.D Dissertation, Virginia Polytechnic Institute and State University, Blacksburg, Virginia, p. 79.

Williams, A.A.B. (1957), Discussion on the Prediction of Total Heave from the Double OedometerTest, Symposium on Expansive Clays, South African Instituion of Civil Engineer, p. 57.

Classification of expansive clay soils 35