Citation
A modified soil suction heave prediction protocol

Material Information

Title:
A modified soil suction heave prediction protocol with new data from Denver area expansive soil sites
Creator:
Diewald, Gary Anthony
Publication Date:
Language:
English
Physical Description:
194 leaves : ; 28 cm

Subjects

Subjects / Keywords:
Swelling soils -- Forecasting ( lcsh )
Soil mechanics ( lcsh )
Soils -- Testing ( lcsh )
Genre:
bibliography ( marcgt )
theses ( marcgt )
non-fiction ( marcgt )

Notes

Bibliography:
Includes bibliographical references (leaves 192-194).
General Note:
Department of Civil Engineering
Statement of Responsibility:
by Gary Anthony Diewald.

Record Information

Source Institution:
|University of Colorado Denver
Holding Location:
Auraria Library
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
54663416 ( OCLC )
ocm54663416
Classification:
LD1190.E54 2003m D53 ( lcc )

Full Text
A MODIFIED SOIL SUCTION HEAVE PREDICTION
PROTOCOL: WITH NEW DATA FROM DENVER AREA
EXPANSIVE SOIL SITES
by
Gary Anthony Diewald
B.S., Brigham Young University, 1997
M.S., University of Colorado at Denver, 2003
A thesis submitted to the
University of Colorado at Denver
in partial fulfillment
of the requirements for the degree of
Master of Science
Civil Engineering
2003


This thesis for the Master of Science
degree by
Gary Anthony Diewald
has been approved
by


Diewald, Gary Anthony (M.S., Civil Engineering)
A Modified Soil Suction Heave Prediction Protocol:
with New Data from Denver Area Expansive Soil Sites
Thesis directed by Professor Jonathan T. H. Wu
ABSTRACT
An important parameter used in the design and repair of residential structures
built over expansive soils in the Denver area is the estimation of potential heave.
In recent years, the suction method for estimation of heave has become more
widely used in engineering practice. An essential component for determining
potential heave by this method is the calculation of suction change, which
requires the comparison of soil suction measurements to a final suction profile.
The purpose of this study is to utilize new pre- and post-construction soil suction
data, from Denver area expansive soil sites, to modify a soil suction heave
prediction protocol developed by Thompson (1997). The new soil suction data
were employed to determine the influence of soil type and age of structure on
soil suction trends and the depth to constant suction. This study has shown the
importance of separating sand and sandstone from clay and claystone when
using suction profiles to determine the depth to constant suction. A depth to
constant suction of around 28 to 34 feet below the ground surface, which is
deeper than indicated in previous studies, is supported by the updated soil
suction data. The depth to constant suction appears to increase overtime. A


suggested adjustment to the final suction profile has been made to better
represent the increase in the depth of wetting.
This abstract accurately represents the content of the candidates thesis. I
recommend its publication.
Signed
IV


DEDICATION
I dedicate this thesis to my loving wife and children for their patience and
support while I was working on this thesis.


ACKNOWLEDGMENT
I would like to thank Ron McOmber for his insight and advice and
CTL/Thompson, Inc. for the use of soil suction data which made this thesis
possible.


CONTENTS
Figures.................................................................ix
Tables............................................................... xii
Chapter
1. Introduction......................................................... 1
1.1 Background .......................................................2
1.2 Research Need ....................................................4
1.3 Research Objective and Tasks ....................... 5
1.4 Thesis Organization...............................................8
2. Literature Review.................................................... 9
2.1 Soil Suction .................................................... 9
2.1.1 Factors Influencing Soil Suction ............................... 12
2.1.2 Relation of Suction to Physical Properties of Soil............. 19
2.2 Soil Suction Profile ............................................23
2.2.1 Depth of Wetting Trends in the Denver, Colorado Area ............27
2.3 Methods of Determining Soil Suction..............................28
2.3.1 Filter Paper Method..............................................28
2.3.2 Testing Procedure............................................... 30
2.3.3 Advantages of the Filter Paper Method................... 32
2.4 Prediction of Potential Heave ...................... 35
2.4.1 Percent Swell Method ........................................... 36
2.4.2 McKeen Soil Suction Method...................................... 37
2.4.3 Suction Compression Index....................................... 39
vii


2.4.4 Comparison of Heave Predictions using Swell Method vs. Suction
Method for Denver Area Expansive Soil Sites...................44
3. Updated Denver Area Soil Suction Data.............................45
3.1 Pre-Construction Soil Suction Data.............................. 45
3.2 Post-Construction Soil Suction Data..............................46
4. Influence of Soil Type on Denver Area Suction Trends .............48
4.1 Pre-Construction Suction Profiles by Soil Type...................48
4.2 Post-Construction Suction Profiles by Soil Type................. 58
4.3 Comparison of Pre- and Post-Construction Suction Profiles......67
4.3.1 Variation in Depth to Constant Suction ......................... 71
4.3.2 Suggested Adjustment of Final Suction Profile................... 73
5. Influence of Age of Structure on Soil Suction Trends..............76
5.1 Age of Structure Suction Profiles............................... 76
5.2 Variation in Depth to Constant Suction
According to Age of Structure ................................ 80
6. Summary and Conclusions ........................................... 82
6.1 Summary......................................................... 82
6.2 Conclusions..................................................... 84
6.3 Suggestions for Future Study ................................... 85
Appendix
A. Raw Pre-Construction Suction Data..................................87
B. Raw Post-Construction Suction Data .............................. 117
C. Raw Age of Structure Data........................................ 176
References........................................................... 192
viii


FIGURES
Figure
2.1 Measured Suction Profile of Summer vs. Winter Seasons ...... 13
2.2 Measured Suction Profiles under Well Ventilated Floors ..... 14
2.3 Measured Suction Profiles below Well Watered Lawns ......... 15
2.4 Measured Suction Profiles of Treed Area vs. Sealed Carpark .... 16
2.5 Measured Suction Profile below a Perched Water Table ....... 17
2.6 Measured Suction Profile within Well Fissured Soils ........ 18
2.7 Suction vs. Water Content Comparison..........................20
2.8 Denver Area, Before and After Construction Suction Data......22
2.9 Suction Profile Indicating a Depth to Constant Suction
of about 2 Meters............................................24
2.10 Seasonal Influence on Suction (Before Construction)...........25
2.11 Seasonallnfluence on Suction (After Construction) ............26
2.12 Calibration Curves for Two Most Common Filter Paper Types .... 29
2.13 Total Suction Profiles as Determined Using Filter Paper
Method and Thermocouple Psychrometers....................... 33
2.14 Suction Profiles Obtained Using Filter Paper Method and
Thermocouple Psychrometers...................................34
2.15 Chart for Estimation of Suction Compression Index (C,,).......40
2.16 Comparison of Suction Compression Index to the Slope of the
Suction vs. Moisture Content Curve and Empirical Relationship
of Ch........................................................41
2.17 Influence of Denver Area Bedrock Types on the Suction
Compression Index ...........................................43
ix


Figure
4.1 Pre-Construction Suction Measurements of All Soil and
Bedrock with Average Fit Line................................... 50
4.2 Pre-Construction Suction Measurements for Clay Soil
with Logarithmic Fit Line ...................................... 51
4.3 Pre-Construction Suction Measurements for Claystone
Bedrock with Logarithmic Fit Line .............................. 52
4.4 Pre-Construction Suction Measurements for Sand and
Sandstone with Logarithmic Fit Line ............................ 53
4.5 Pre-Construction Suction Measurements for Clay and
Claystone with Logarithmic Fit Line.............................56
4.6 Pre-Construction Suction Measurements for Clay and Claystone
with Plus and Minus One Standard Deviation......................57
4.7 Post-Construction Suction Measurements of All Soil
and Bedrock with Average Fit Line ..............................59
4.8 Post-Construction Suction Measurements for Clay Soil
with Logarithmic Fit Line ......................................60
4.9 Post-Construction Suction Measurements for Claystone
Bedrock with Logarithmic Fit Line ..............................61
4.10 Post-Construction Suction Measurements for Sand and
Sandstone with Logarithmic Fit Line ............................62
4.11 Post-Construction Suction Measurements for Clay and
Claystone with Logarithmic Fit Line.............................65
4.12 Post-Construction Suction Measurements for Clay and Claystone
with Plus and Minus One Standard Deviation......................66
4.13 Comparison of Pre- and Post-Construction Suction
Profiles for Clay Soil .........................................68
4.14 Comparison of Pre- and Post-Construction Suction
Profiles for Claystone Bedrock .................................69
4.15 Comparison of Pre- and Post-Construction Suction
Profiles for Clay and Claystone.................................70
x


Figure
4.16 Comparison of Plus and Minus One Standard Deviation of Pre-
and Post-Construction Profiles for Clay and Claystone..........72
4.17 Comparison of Pre- and Post-Construction Suction Profiles and
Original Final Suction Profile...................................74
4.18 Comparison of Pre- and Post-Construction Suction Profiles and
Adjusted Final Suction Profile ..................................75
5.1 Age Distribution of Suction Data ...............................78
5.2 Suction Profiles for Clay and Claystone According
to Age of Structure............................................. 79
5.3 Suction Profiles for Clay and Claystone According to Age of
Structure Compared to Pre-Construction Suction Profile ........81


TABLES
Table
2.1 Conversions of Various Suction Units........................... 11
2.2 Correlation of Suction with Physical Properties of Soil ....... 19
xii


1. Introduction
Residential development in the Greater Denver Metro Area has been
forced into areas which are underlain by expansive soils as most of the
good (non-expansive soil) building sites have been developed. These
expansive soils can cause damage to floor slabs and structures upon
wetting. One of the parameters used to design and repair residential
structures built over expansive soils is the estimation of potential heave.
Presently there are two methods commonly used in consulting geotechnical
engineering practice to estimate potential heave. These are the
consolidation swell test method (ONeil, 1980) and the soil suction method
(McKeen, 1992).
An essential component in the determination of potential heave by the
soil suction method is the suction change for each soil layer. Soil suction is a
parameter used to describe the state of the soil in relation to moisture. Once
the soil suction of a particular soil sample at depth has been determined,
usually through the filter paper test method (Nelson and Miller, 1992), that
value must then be compared to a final suction value which is estimated
based on past post-construction soil suction trends and the depth of wetting
for a particular region. As new post-construction soil suction data becomes
available, usually through warranty-type investigations where the structure
has experienced some kind of heave distress, it should be evaluated to
determine if any change in the soil suction trends and depth to constant
suction, i.e. depth of wetting, warrant an adjustment to the estimated final
1


suction values.
This study focuses on the practical use of soil suction data in
estimating potential heave. The procedures used to determine soil suction
and methods used to estimate potential heave will be addressed. Updated
post-construction soil suction data from Denver area expansive soil sites has
been evaluated together with some pre-construction soil suction data to
determine new trends in soil suction profiles and depth to constant suction.
Where the suction profiles show a change from former trends or depth to
constant suction, the final suction profile has been adjusted.
1.1 Background
Soil suction theory was developed in the early 1900's and for
approximately forty years was applied only in the agricultural field to study
the soil-water relation to plant systems. The concept of soil suction for the
purpose of studying the behavior of unsaturated soils was first proposed by
the Road Research Laboratory in England in the late 1940's and early 1950's
(Fredlund and Rahardjo, 1993). However the use of soil suction research for
the purpose of predicting heave did not come about until the late 1970's and
early 1980's. Soil suction in geotechnical consulting practice has only been
utilized within the past ten years.
In 1997, a protocol for evaluating the repair of residences damaged by
expansive soils was published by Thompson (1997). This protocol was
brought about by a series of Denver area class action lawsuits filed by
homeowners against developer and home builder organizations. The
2


residences involved in the lawsuits ranged from 6 to 10 years in age. An
important outcome of these lawsuits was the establishment of a protocol for
evaluation and mitigation of basement floor slabs damaged by expansive
soils.
An essential component of the evaluation protocol is the determination
of remaining potential heave within the expansive soils below the structure.
The protocol allowed for two alternative procedures of estimating remaining
potential heave. These procedures are the percent swell method, developed
by ONeil (1980), and the soil suction method, developed by McKeen (1992).
In order to calculate potential heave using the soil suction method, the
suction profile of the soil must be compared to a final suction profile
(McKeen, 1992). The final suction profile established in the protocol was
estimated by Robert W. Thompson, of CTL/Thompson, Inc. in Denver,
Colorado based upon post-construction soil suction data of Denver area sites
from 1989 to 1995, and a depth of wetting and suction change of 20 feet
below the ground surface. This depth of wetting was established through
negotiation between engineers representing both sides of the lawsuits.
After applying the protocol to distressed residence investigations for
about one year, suction data was reviewed to evaluate the original protocol
(McOmber and Thompson, 2000). This data included suction measurements
from 51 residential sites prior to construction and 41 sites after construction
where some form of damage to the structure had been reported. A better
understanding of the variability of soil suction trends was obtained, and the
data generally indicated the protocol assumptions were reasonable.
3


1.2 Research Need
In the years that have followed since 1998, soil suction measurements
have been obtained from numerous warranty-type investigations at
residences where damage to the basement floor slab or structure has
occurred. A total of 133 residences have been investigated through
September 2002. There is a need to collect and compile the new soil suction
data from these investigations. Analysis of soil suction measurements from
these investigations combined with data from previous investigations would
provide opportunity to update and improve the understanding of potential
heave caused by expansive soils, through consideration of the following
questions:
1. How does measured suction vary according to soil type?
2. Does the new suction data support the present depth of wetting
or depth to constant suction?
3. What influence does the age of the structure have on the
measured suction profile and depth to constant suction?
4. Is the present final suction profile still valid, or does it need to
be adjusted based on new suction data?
4


1.3 Research Objective and Tasks
The objective of this study was to propose a modified protocol for soil
suction heave prediction based upon updated pre- and post-construction soil
suction data, from Denver area expansive soil sites. The new soil suction
data were employed to determine the influence of soil type and age of
structure on soil suction trends and the depth to constant suction.
Five main research tasks were conducted to achieve the research objective,
including:
Task 1. Summarize the advantages and methods of using soil suction to estimate potential heave;
Task 2. Collect and compile updated pre- and post-construction suction measurements obtained from warranty-type and design level investigations during the past four years;
Task 3. Collect the date of construction and date of investigation for each of the warranty-type investigations;
Task 4. Determine the influence of soil type and age of the structure on the soil suction trends (profiles) and depth to constant suction; and
Task 5. Adjust the final suction profile to better represent any changes in the updated suction profiles and depth to constant suction.
To carry out Task 1, the following were undertaken:
An extensive literature review of past studies related to soil
suction heave prediction was conducted;
The methods of obtaining soil suction measurements were
5


researched for their applicability to consulting engineering
practice; and
The soil suction method of estimating potential heave was
compared to the consolidation swell method by reviewing the
results of a Denver area case study.
To carry out Task 2, the following were undertaken:
Pre-construction soil suction measurements were collected from
16 updated (since 1998) Greater Denver Area geotechnical
investigations conducted by CTL/Thompson, Inc.;
Post-construction soil suction measurements were collected
from 133 updated warranty-type geotechnical investigations
conducted by CTL/Thompson, Inc.; and
The sample depth, moisture content and soil type
corresponding to each suction measurement for both pre- and
post-construction investigations were compiled in Microsoft
Excel spreadsheets and added to the previous data (since
1996) for use in the analysis.
To carry out Task 3, the following were undertaken:
The date of investigation was collected from CTL/Thompson
records, and the date of construction for each residence was
obtained from on-line county assessor records;
The age data was compiled with suction measurements in a
Microsoft Excel spreadsheet;
6


To carry out Task 4, the following were undertaken:
Both pre- and post-construction soil suction data were sorted
according soil type (clay, claystone, sand, sandstone) and
plotted according to depth;
Logarithmic regression fit lines (profiles) were generated for
both pre- and post-construction data. These log fit lines were
plotted together for determining the depth to constant suction;
and
Soil suction data was sorted according to the age of structure at
the time of investigation, then the data and log fit lines were
plotted in three year intervals and compared to the pre-
construction data.
To carry out Task 5, the following were undertaken:
The intersection of the pre- and post-construction soil suction
profiles for clay and claystone soil types was used to estimate
the depth to constant suction;
The depth to constant suction represented by the updated data
was compared to the original final suction profile; and
The updated depth to constant suction and percentage of
suction data points to the left of the final suction profile was
used to estimate an adjusted final suction profile.
CTL/Thompson, Inc. has pioneered the incorporation of soil suction to
engineering practice in the Denver area and has been obtaining soil suction
measurements from geotechnical investigations since about 1982. Other
engineering firms in the Denver area have only been obtaining suction
measurements for the past few years. Recent sharing of soil suction data
among the few Denver area geotechnical firms which collect soil suction data
7


has shown that suction measurements tend to vary slightly according to the
lab where suction tests were performed. Therefore, in order to collect a
large amount of suction data and limit the variability in suction
measurements, suction data used in this study was obtained only from
CTL/Thompson, Inc.
1.4 Thesis Organization
Chapter 1 presents the introduction, background, research needs,
research objectives and methods, and thesis organization. Chapter 2
describes the theory of soil suction, methods of determining soil suction and
potential heave, and the advantages of soil suction vs. other methods of
heave prediction as described in the published literature. Chapter 3
describes the methods used to collect, compile and analyze the updated soil
suction data. Chapter 4 presents the influence of soil type on Denver area
suction trends and a determination of the depth to constant suction
represented by the updated data. Chapter 5 presents the influence of age of
structure on suction trends and the depth to constant suction. Chapter 6
describes the summary, conclusions and suggestions for future research.
Appendix A contains the raw pre-construction soil suction data. Appendix B
contains the raw post-construction soil suction data. Appendix C contains
the raw age of structure data.
8


2. Literature Review
Soil suction theory was developed in the early 1900's to study the soil-
water relation to plant systems in the agricultural field Soil suction was first
applied to study the behavior of unsaturated soils in the late 1940's and early
1950's (Fredlund and Rahardjo, 1993). Soil suction has only been applied to
the prediction of potential heave in expansive soil since the late 1970's and
early 1980's. Soil suction in geotechnical consulting practice has only been
utilized within the past ten to fifteen years. During this time numerous
studies have been conducted to compare the soil suction method of heave
prediction to other methods of predicting potential heave.
In this study, an extensive literature review was conducted to
summarize soil suction theory and its usefulness in geotechnical engineering
practice. The various methods of determining soil suction and potential
heave are presented. The results of a Denver area case study which
compare the consolidation swell method to the soil suction method are
summarized.
2.1 Soil Suction
Soil Suction is defined as the free energy state of soil water (Fredlund
and Rahardjo, 1993). Chen (1988) has described soil suction in laymans
terms as a measure of the soils affinity for water. Soil suction is a parameter
used to describe the state of the soil in relation to moisture. This parameter
can be used to characterize the effect of moisture on the volume and strength
9


properties of the soil (Snethen, 1980). Soil suction is compose of two parts,
matric suction and osmotic suction. Matric suction is composed of the
surface tension forces of the water in the soil and the absorptive forces
exerted on the water molecules by the cations in the soil (Chen, 1988). The
matric suction is therefore both water content and surcharge pressure
dependent (Snethen, 1980). The osmotic suction in soil is due to the
presence of soluble salt solutions in the soil water and is identical in context
with the osmotic attractive forces. Osmotic suction is independent of the
water content and surcharge pressure. In order for osmotic suction to occur
there must be a chemical differential between the concentration of salts in the
soil and in the free water. This chemical differential is very small to non-
existent in Denver area soils. Therefore, osmotic suction is not a significant
factor in the total suction of these soils.
The sum of the matric and osmotic suctions is the total suction of the
soil. The total suction is defined by the Review Panel of the 1960 Moisture
Equilibria Symposium as, The negative gauge pressure relative to the
external gas pressure on the soil water to which a pool of pure water must be
subjected in order to be in equilibrium through a semi-permeable membrane
with the soil water (Wray, 1984). In order to avoid the use of negative
characteristics of logarithms, the soil suction can be expressed more
conveniently as the log of the height of the water column in centimeters, or
pF (McQueen and Miller, 1968; McKeen, 1980). Soil suction is also
commonly reported in units of kPa, bars and psi as shown in Table 2.1.
10


Table 2.1
Conversions of Various Suction Units
Atmospheri c Pressure (cm) ' V PF v t1 r' > kPa r ? / / ' nci * psi T i' * ^ ; -
1 0 0.0981 0.00098 0.1422
10 1 0.981 0.0098 1.422
102 2 9.81 0.098 14.22
103 3 98.1 0.98 142.2
104 4 981 9.81 1422
10s 5 9810 98.1 14220
106 6 98100 981 142200
107 7 981000 9810 1422000
(after Mckeen, 1980; and Chen, 1988)
11


2.1.1 Factors Influencing Soil Suction
Total soil suction is influenced by temperature, pressure, solute
concentration, gravity, absorption, soil composition (amount and type of clay
mineral), cation environment of the soil, soil structure (i.e. fissures, bedding
planes), soil moisture and soil density (Snethen, 1980; Wray, 1984). The
factors most important to expansive soils are soil moisture, soil composition,
cation environment, which is related to the type and amount of clay mineral
present, and soil structure. The soil structure is critically important in
determining the response of soils to climatic/moisture changes (McKeen,
1992). Dense non-fissured clays may have large suction potential but lack
sufficient permeability to permit water to penetrate the clay. On the other
hand, highly weathered and fractured clays may allow water to penetrate
deep into the clay layer.
The soil suction quantitatively describes the interaction between the
soil particles and water (Snethen, 1980). This interaction determines the
physical behavior of the soil mass. Therefore, the moisture change within a
soil mass is an important characteristic influencing soil suction variations in
expansive clay soils. Moisture flows through the soil from regions of low
suction to regions of high suction. This is termed the suction gradient. A
suction profile (soil suction versus depth curve) will tend towards an
equilibrium with a moisture source at the boundaries or within the soil mass
(Mitchell and Avalle, 1984). Boundary conditions of moisture change include
climatic and seasonal changes, structures, irrigation, vegetation,
12


water table and soil fabric. Measurements of suction profiles for several of
these boundary conditions were conducted by Mitchell and Avalle on sites in
Adelaide, South Australia. Figures 2.1 through 2.6 illustrate variations in soil
suction according to various boundary conditions.
Total Soil Suction (pF)
Q.
a>
n 3 4 5
- I

l - o A
A
A
2 a A
9 A
3 a J
a Si te at end of
dry summer
4 9 A m Site at end of
wet winter
A
A
5 A
A
6
Open Sites, Extreme
Seasonal Conditions
Figure 2.1
Measured Suction Profiles of Summer vs. Winter Seasons
(from Mitchell and Avalle, 1984)
13


Total Suction (pF)
4.0 4.5 5.0 5.5
f l 1 , a-,
0.5
1-0 O 9
1.5 9 O 9 0
2.0 O o Mitchell (1979) oAitchison &
Woodburn (1969)
2.5 o 0 OSchumann (1967)

Measured. Suction Pro-
files under well
Ventilated Floors
Figure lb
Figure 2.2
Measured Suction Profiles under Weil Ventilated Floors
(from Mitchell and Avalle, 1984)
14


Total Suction (pF)
n 3 4 5
7 O' 1 i
_ Mitchell (1979), lawnJ adjacent to 4 year old
house
Values reported by
i ^ o Woodburn (1972)
o Authors' own tests
wo

2
O 0

J. 3 - o
-C
(D Q - o
4 o
o v\
5 - O
o
6 , ri
Well Watered Lawns
Figure Ic
Figure 2.3
Measured Suction Profiles below Well Watered Lawns
(from Mitchell and Avalle, 1984)
15


Total Soil Suction (pF)
u 1 : | r*
A
1 - A
A
2 A.
A
'S 3 A Treed area
(large
CL eucalypts)
o A Bitumen Sealed
4 Carpark
5
6
Measured Suction Profiles
in an Area of Trees and a
nearby Sealed Carpark
Figure.Id
Figure 2.4
Measured Suction Profile of Treed Area vs. Sealed Carpark
(from Mitchell and Avalle, 1984)
16


Total Suction (pF)
3.4 3.6 3.8 4.0 4.2
containing, a Perched Water
Table (Aitchison et al, 1973)
Figure le
Figure 2.5
Measured Suction Profile below a Perched Water Table
(from Mitchell and Avalle, 1984)
17


Total Suction (pF)
. j 1 T
A
A
ii
c A A. '
9
A 9
A A Under
A Wetted
'f 10 Surface
A. A
x: 9 a Under
Q_ Q 15 A ' Field
9
A
a
20
Moisture Penetration
into Fissured Soil
(Williams, 1980)
Figure If
Figure 2.6
Measured Suction Profiles within Well Fissured Soils
(from Mitchell and Avalle, 1984)
18


2.1.2 Relation of Suction to Physical Properties
of Soil
The most valuable attribute of using soil suction for interpretation of
the behavior of expansive soils is that the suction level of soil water
correlates with the physical behavior of the soil (McKeen, 1992). These
correlations are presented in Table 2.2. The behavior at each suction level
will be consistent regardless of the clay mineral composition, although the
water content of the soil at each suction level will be affected by the type
and amount of clay mineralogy.
Table 2.2
Correlation of Suction with Physical Properties of Soil
Soil Behavior Suction Level, pF (kPa)
Saturation 0.0
Liquid Limit 1.0 (0.98)
Field Capacity 2.0-2.5 (9.8-31)
Plastic Limit 3.2-3.5(155-310)
Plant Wilting Point 4.2-4.5 (1,554-3,101)
Tensile Strength of Water 5.3(19,566)
Shrinkage Limit/Air Dry 5.5 (31,000)
Oven Dry 7.0 (980,638)
(after McKeen, 1992)
19


McKeen (1992) has shown that the relation between suction levels
and water content tends to be a straight line between the levels of around 6
pF and 3 pF which covers the suction range of most real soils. At suction
levels between 2 and 2.5 pF water will begin to run off as the soil nears field
capacity. At field capacity the additional water will drain away rather than be
absorbed into the soil. This relation is shown in Figure 2.7.
Figure 2.7
Suction vs. Water Content Comparison
(from McKeen, 1992)


Soil suction data collected since the early 1980's by CTL/Thompson,
Inc. has shown that suction values for expansive clays in the Denver,
Colorado area generally vary between 3.5 to 4.5 pF, with few data points
outside these limits (McOmber and Thompson, 2000). A portion of this data
collected since 1996 is shown on Figure 2.8. The data include suction values
from 51 residential sites prior to construction and 41 sites after construction.
Suction values less than about 3.0 pF have not been measured on expansive
soils in the Denver area, even under extreme wetting conditions. The data
indicates that according to typical suction values the physical property of
Denver area soil falls between the plastic limit and the plant wilting point.
Field capacity of the soil is never reached in expansive soils above a depth
of 25 feet.
21


Suction (pF)
Figure 2.8
Denver Area, Before and After Construction Suction Data
(after McOmber and Thompson, 2000)
22


2.2 Soil Suction Profile
An essential factor in calculating the heave of expansive soils is the
determination of the depth of wetting, depth to constant suction or active
zone. The active zone depth is the depth to which moisture changes will
occur due to climate and man-made influences in the soil moisture regime. If
there are no changes in the soil moisture environment then there is no
tendency for a change in soil volume regardless of the suction potential. The
suction profile (plot of suction vs. depth) combined with the moisture content
profile are effective tools in determining the depth of the active zone for
heave calculations. The suction profile shows the wet and dry areas with
respect to the soil moisture condition below the ground surface. This is
apparent in Figure 2.9 where the soil suction varies to a depth of 2 meters
(6.5 feet) below the surface, after which the suction remains fairly constant.
The suction values plotted to the left of the constant line below 2 m represent
sand seams in the expansive clay. The magnitude of suction change will
vary from a maximum value at the surface to zero at the depth of the active
zone (McKeen, 1992).
23


o F
;
9 b
4,
; P
I > | r!i
p
: rrrrt
7 11 11------' ' ' ' 1 '----' .
2 3 4 5 6
Suction (pF)
Figure 2.9
Suction Profile Indicating a Depth to Constant Suction of about 2 Meters
(from McKeen, 1992)
A combination of suction values from the suction profile and
consideration of the climate that produced the suction profile, in comparison
to the normal weather cycle is a good basis for estimating the depth of the
active zone (Mckeen, 1992). McOmber and Thompson (2000) have shown
the effect of climate on suction profiles at pre- and post-construction sites in
the Denver, Colorado area, Figures 2.10 and 2.11. These figures show that
seasonal moisture variations in the Denver area have little effect on the post-
construction suction profile. McOmber and Thompson (2000) have
concluded that once soils and bedrock in the Denver area become wetted,
they tend not to dry out significantly.
24


Depth (ft)
2.5
3.0
4.0
4.5
5.0
Suction (pF)
2.0
3.5
Seasonal lnlfuence on Suction
(Before Construction)
Figure 2.10
Seasonal Influence on Suction (Before Construction)
(from McOmber and Thompson, 2000)
25


Suction (pF)
2.0 2.5 3.0 3.5 4.0 4.5 5.0
Seasonal Influence on Suction
(After Construction)
Figure 2.11
Seasonal Influence on Suction (After Construction)
(from McOmber and Thompson, 2000)
26


2.2.1 Depth of Wetting Trends in the
Denver, Colorado Area
Suction data from the Denver, Colorado area collected since the early
1980's using a number of different laboratory techniques indicate a depth of
wetting of 8 to 10 feet for non-irrigated profiles influenced by climate only
(Thompson, 1997). The suction and moisture profiles of expansive clays and
claystone are significantly altered by construction and irrigation. These
man-made processes often cause a greater variation in the suction profile
than natural climatic variations. Data from remediation investigations, on
sites underlain by flat-lying bedrock throughout the Denver area, indicate the
depth of wetting is increased to between 16 and 22 feet for residential and
commercial sites where there is considerable landscaping and irrigation
(Thompson, 1997).
Pre-construction and pre-irrigation total suction measurements in the
Denver area generally range from about 4.2 to 4.5 pF for the upper 20 feet of
soil (Thompson, 1997). After wetting and heaving, total suction profiles
typically decrease to between 3.6 to 3.8 pF. The largest changes in total
suction are usually observed in the upper 10 to 15 feet of soil. Figure 2.8
(presented previously) shows the effect of construction on the log fit of
suction data compiled by CTL/Thompson since 1996. The suction profile has
shifted to the left indicating a decrease in suction values, i.e. an increase in
soil moisture, in the top 15 to 20 feet of soil. This would indicate an active
zone of around 15 to 20 feet below the surface is typical for post-construction
27


sites in the Denver Area.
2.3 Methods of Determining Soil Suction
Numerous methods have been used for the measurement of soil
suction. These methods include thermocouple psychrometers, tensiometers,
pressure plates, filter paper and thermal conductivity sensors. Of these, the
filter paper method is gradually becoming more accepted in the geotechnical
engineering field due to several operational advantages. These advantages
include: simplicity of operations, low cost (about the same as water content
test), and the wide range of suction values which can be measured (McKeen,
1980).
2.3.1 Filter Paper Method
The filter paper method was first developed for the soil science and
agronomy fields. The use of filter paper for the in-direct measurement of soil
suction in the geotechnical engineering field was first attempted in the late
1970's and early 1980's. The theory behind the filter paper method stems
from thermodynamic principles of liquid or vapor moisture diffusion (Fredlund
and Rahardjo, 1993). The basis of the method is that the filter paper will
come to equilibrium with the relative humidity in the test canister. The
relative humidity is controlled by the moisture content and suction potential of
the soil. The filter paper method is able to measure both matric and total
suction. When the paper is in good contact with the soil, moisture is
transferred in the liquid state, and the matric suction is measured. When the
filter paper is suspended above the soil sample, moisture is transferred in the
28


vapor state, and the total suction is measured.
The filter paper method utilizes ash-free filter papers which exhibit a
consistent and predictable relationship between suction and water content
(Nelson and Miller, 1992). The most common of these paper types are
Whatman No. 42, and Schleicher and Schuell No. 589. A calibration curve
must be generated relating suction to water content for each brand of filter
paper. The calibration of suction to water content is accomplished by placing
the filter paper over a known solution of salt in an airtight container until
moisture equilibrium is reached (ASTM D 5298-94). Typical calibration
curves for the two common filter papers are shown in Figure 2.12.
Figure 2.12
Calibration Curves for Two Most Common Filter Paper Types
(from Fredlund and Rahardjo, 1993)
29


2.3.2 Testing Procedure
The procedure for conducting filter paper tests was first proposed by
McQueen and Miller in 1968. It was not until 1992 that the filter paper
method was standardized by ASTM as Method D 5298. The ASTM standard
test method indicates filter papers are to be dried in a drying oven for at least
16 hours or overnight and then placed in a desiccator container. A sample of
soil between 200 to 400 grams is placed inside a corrosion resistant
container. A spacer or ring is placed around the soil to hold the filter paper
above the soil without allowing the soil and filter paper to come into contact.
Two filter papers are then removed from the desiccator container and placed
on top of the spacer or ring. The lid is then placed on the container and
sealed with at least one wrapping of electrical tape. The sealed container is
then placed in an insulated chest or environmentally controlled chamber
where the temperature variations are less than 3. A typical temperature
should be around 20 C (68 F). The soil specimen and filter paper should be
allowed to come to moisture equilibrium for a minimum of seven days.
The most important aspect of the filter paper test is to obtain a precise
measurement of the moisture content of the filter paper after moisture
equilibrium is reached. In order to accomplish this care must be taken to
minimize the moisture loss while the filter paper is transferred to the weighing
container and during determination of the mass. Therefore, before the
specimen container is removed from the insulated chest, the mass of the filter
paper drying tins should be determined. Once cold tares for the tins are
30


determined to within 0.0001 grams, the containers can be removed from the
insulated chest and the filter papers removed with tweezers and placed in the
dying tins. The operation of removing the filter papers and placing them in
the tins should be done quickly (3 to 5 seconds) to minimize the moisture
loss. Then immediately determine the mass of the tins with filter papers to
the nearest 0.0001 gram.
Next place the tins with the lids ajar or unsealed in a oven at 1105 C
(2305 F) for at least 2 hours to allow drying of the filter papers. At the end
of at least 2 hours the tins should be sealed and remain in the oven for at
least 15 min to allow for temperature equilibration. Determine the dry total
mass of the tin and filter paper to within 0.0001 gram. Then remove the filter
paper and place the tin on a metal pad to provide a heat sink. Next
determine the hot tare of the tin. Now the moisture content of the filter paper
can be determined and compared to the appropriate calibration curve to
determine the total suction. Filter papers should be thrown away and not
used again. The total suction can be taken as the average of the suction
values determined from the two filter papers. If the difference in suction
between the two filter papers is greater than 0.5 log kPa the results should
be discarded and a second filter paper test should be performed (ASTM D
5298-94).
31


2.3.3 Advantages of the Filter Paper Method
Suction values obtained from the filter paper method have been
compared to suction values obtained through thermocouple psychrometer
measurements by Mckeen (1981) and van der Raadt et al. (1987).
Comparisons of the suction profiles generated by the two methods are
presented in Figures 2.13 and 2.14. A comparison of the profiles indicates
that the suction profiles generated from the filter paper test are similar to
those profiles generated from the more traditional thermocouple
psychrometer test. McKeen (1980) used the filter paper method for laboratory
measurements and thermocouple methods for in-situ measurements of soil
suction during a study of airport pavements on expansive clays. McKeen
determined that the filter paper measurements from the laboratory were a
good indicator of the in-situ suction values measured with thermocouple
psychrometers.
32


Suction (kPa)
TO1 102 103 104 105
Figure 2.13
Total Suction Profiles as Determined Using Filter Paper Method and
Thermocouple Psychrometers
(from McKeen, 1981)
33


Suction (kPa)
S
JZ
o
a
10 100 1000 10000
Figure 2.14
Suction Profiles Obtained Using Filter Paper Method
and Thermocouple Psychrometers
(from van der Raadt et al., 1987)
During a study of the effects of heaving soils on residences in a
subdivision southwest of Denver, Colorado, CTL/Thompson, Inc. performed
numerous soil suction measurements using both thermocouple psychrometer
and filter paper methods. After only a few months of testing the
thermocouple psychrometer method was abandoned in favor of the filter
paper method due to a number of problems experienced in the consistency of
laboratory results (Thompson and McKeen, 1995). Thompson indicated that
34


for normal commercial laboratory soil testing, where the work is usually
performed by engineering technicians, the filter paper method yields more
reliable and consistent test data. Another advantage of the filter paper
method over the thermocouple psychrometer method is the range of suction
that can be measured. The filter paper method is effective in measuring
suction values over the entire range of suction potential from 0 to 7 pF,
where as the psychrometer method has a limited range (McKeen, 1980).
One disadvantage to the filter paper method is that is has not yet been
developed for in-situ measurements of soil suction. However as shown by
McKeen (1980), laboratory filter paper measurements show satisfactory
correlations with in-situ measured values especially when soil samples are
immediately sealed in testing cans with filter papers at the field site and
sample disturbance is kept to a minimum.
2.4 Prediction of Potential Heave
Heave prediction is the process of determining an estimated free-field
heave (heave under no surcharge). The free-field heave is the displacement
that will occur at the ground surface due to the swelling characteristic of
expansive soils at and below the ground surface. This heave prediction
ultimately allows for the determination of a site as low, moderate, high or very
high for risk of heave beneath a slab or other structure. The prediction of
future or remaining heave is an essential step of the repair protocol
developed by Thompson (1997). The prediction of potential heave is also
important during the design phase of geotechnical investigations to help
determine foundation and floor types which will perform best on expansive
soils.
35


As part of the establishment of the repair protocol, multiple testing was
performed at a test site to compare various heave prediction procedures.
The procedures compared include the consolidation swell test method
(ONeill, 1980), the soil suction method (McKeen, 1992), the Cation
Exchange Capacity method (Lytton,1994) and the swell index method
(Fredlund,1993). Based on availability and the time required to run the
testing, the percent swell method by ONeill and the suction method by
McKeen were selected as alternative procedures to use in heave predictions.
According to Thompson (1997), where soil samples are similar, the data
indicate that the adjusted swell method yields predicted heave values about
the same as the suction method.
2.4.1 Percent Swell Method
The percent swell method is the more traditional of the two methods
used to predict heave in the Denver, Colorado area (Perko, Thompson and
Nelson, 2000). This method is based on the representation of soil suction
changes through equivalent changes in effective stress, and
utilizes one-dimensional swell-consolidation measurements at surcharge
pressures approximating the overburden weight of soils and/or bedrock at the
sample depth.
36


The procedure used to calculate heave by the percent swell method is as
follows:
Total Heave Estimate = Ht (in or mm)
Ht = £Hi
Hi = estimated heave for each soil layer (in or mm)
Hi = tj (%sWj)
t| = thickness of layer i (in or mm)
%sWj = average swell for soils tested in layer i (%)
Design Heave = Hd (in or mm)
Hd = 0.70 Ht
0.70 = correction factor to adjust for field heave
This method involves dividing the soil profile within the active zone
into layers. The active zone established for the protocol was 20 feet below
the ground surface (Thompson, 1997). The heave of each layer is calculated
based on percent swell and them summed over the entire profile. The
correction factor of 0.70 was added by Thompson (1997) to account for the
difference between observed field heave and the heave predictions from
laboratory swell test in the Denver area.
2.4.2 McKeen Soil Suction Method
The other alternative for determining potential heave according to the
repair protocol is the soil suction method developed by McKeen (1992). This
method regards soil suction as the primary stress state variable and relates
soil heave directly to suction changes (Perko, Thompson and Nelson, 2000).


The procedure used to calculate heave by the McKeen soil suction method is
as follows:
Total heave estimate = Ht (ft)
Ht = £AHi
AHi = vertical heave of layer i (ft)
AHi = Ch Ah Atf s
Ch = suction compression index for layer i
Ah = suction change in layer i (pF)
At = thickness of layer i (ft)
f = lateral restraint factor
s = load effect factor
The suction method of heave prediction requires the determination of
the suction change for each layer of soil within the active zone. This change
in suction must be determined from the difference between measured suction
values and an estimated final suction profile. The final suction profile used in
the repair protocol for heave predictions was estimated by Thompson (1997)
based on observed suction values at heaving sites within the Denver area.
This final suction profile is based on a depth of wetting of 20 feet as shown in
Figure 2.8 (presented previously).
This method also requires the determination of the suction
compression index for each layer and the assignment of values for a lateral
restraint factor (f) and a coefficient for load effect (s). The lateral restraint
factor, f = (1+2Ko)/3, for Denver area soils has been found to be: f = 0.5 for
highly fractured clays and f = 0.83 for massive clay with little to no fractures.
The coefficient (s) is determined from the swell pressure,
38


s = 1.0 0.01 (%SP). Where swell pressure is not determined an (s) value of
0.9 is assigned (Thompson, 1997).
2.4.3 Suction Compression Index
The suction compression index (Ch) is the slope of the volume change
versus soil suction curve or Ch = aV/Vm Ah. If the suction compression
index is not measured in the laboratory by performing Coefficient of Linear
Extensibility (COLE) test and soil suction tests, there are two methods of
estimating the suction index. The first method is based on COLE results from
research conducted by McKeen (1977) and allows the suction index to be
calculated using three regression equations when the clay mineralogy and
the clay percentage are known (Wray, 1997).
for smectite clays: Ch = 0.00018 (% of clay) 0.000098
for illite clays: Ch = 0.00047 (% of clay) 0.00351
for kaolinite clays: Ch = 0.00056 (% of clay) 0.00433
The second method for estimating the suction compression index
involves using a chart developed by McKeen (1980), shown on Figure 2.15.
The soil properties required to use the chart are the plasticity index (PI), the
percent of the soil passing the #200 sieve, and the soils Cation Exchange
Capacity (CEC). The CEC is not commonly used in commercial testing
laboratories, but can be estimated at PI117 (Wray, 1997). The chart also
requires the Activity (Ac) and the Cation Exchange Activity (CEAc) of the clay
be determined. These values are respectively the PI divided by the % clay
and the CEC divided by the % clay. Plotting these values on the chart gives
the Ch for soils with 100% clay content. For real soils, the value of Ch
39


obtained from the chart must be multiplied by the % clay in the soil.
P.I. = 51 percent
clay = 52 percent
CEC = 31.7 me/100 g
Ac = 0.981
CEAc = 0*610
yh = [0.53(0.163)] = 0.085
Example:
o
<
IU
o
ISA
(0.096)
X
(0.2E0)
(0,033)
>
t 1.0
TL
(0.163)
ZB
(0.033)
12 A [0.096)
(0.061)
KB
(0.061)
z
o
0.1
j______I___i i i.. i-l.
0.1
0.5 1.0
ACTIVITY, AC
2.0 3.0
Figure 2.15
Chart for Estimation of Suction Compression Index (Ch)
(from McKeen, 1980)


McKeen (1992) has utilized suction data from several field monitoring
sites to compare the suction-water content slope and the suction
compression index. From this comparison a line can be drawn representing
the 85th percentile based on Ch> Figure 2.16. This line allows for Ch to be
estimated based on the slope of the suction versus water content curve,
Ah/AW, as represented by the following equation.
Ch = (-0.02673)(aIi/aw) 0.38704
A
y
X
o
o
c
t-H
£
o
a
o
U
c
o
u
3
w
-20 -15 -10 -5 . 0
Ah/Aw
Figure 2.16
Comparison of Suction Compression index to the Slope of the Suction
vs. Moisture Content Curve and Empirical Relationship of Ch
(from McKeen, 1992)
41


Perko, Thompson and Nelson (2000) conducted a study to verify
previously established empirical relationships used to determine Ch This
study involved the results of CLOD test of clay and claystone bedrock from
the Denver, Colorado area. In general, the results of this study matched
McKeens empirical relationship, thus supporting its validity and versitality
(Perko, Thompson and Nelson, 2000). The relationship established by
McKeen (1992) was shown to be more realistic and accurate for soils in the
mid-range of suction to water content ratios. The average Ch measured in
the Denver area was shown on Figure 2.17 to be a function of bedrock
geology (Perko, Thompson and Nelson, 2000).
42


Average Ch
-0.20
-0,18
-0-.16
-0.14
-0.12
-0.10
-0.08
-0.06
-0,04
-0.02
0.00
Figure 2.17
Influence of Denver Area Bedrock Types
on the Suction Compression Index
(from Perko, Thompson and Nelson, 2000)
43


2.4.4 Comparison of Heave Predictions using
Swell Method vs. Suction Method for
Denver Area Expansive Soil Sites
The percent swell and McKeen soil suction methods of heave
prediction were implemented by Thompson and McKeen (1995) during an
extensive subsurface investigation of residential structures undergoing
movement in a particular subdivision southwest of Denver, located within the
Designated Dipping Bedrock Area (DDBA). Heave predictions were made
using a suction index determined from empirical relations between the
suction water content slope, aIi/aw. The suction change was determined
from the difference in the pre- and post-construction suction profiles from
three relatively deep (50 feet) test holes. The percent swell method
calculated a potential heave of 26.5 inches. Using the suction method
presented by McKeen, a total heave estimate of 26.2 inches was calculated.
The actual measured field heave at the same boring location was 24.8
inches. This data indicates that for this site the predicted total heave
estimates by the swell and suction methods were very similar to each other,
as well as to actual measured field heave.
44


3. Updated Denver Area Soil Suction Data
Approximately six years have past since the original protocol for
estimating future heave at distressed residential sites was established by
Thompson (1997), and about five years have past since suction data were
reviewed by McOmber and Thompson (2000) to evaluate the original
protocol. During these past five years CTL/Thompson, Inc. has continued to
investigate subsurface soil conditions at residential sites where damage to
the basement floor slab or structure has occurred. These warranty-type
investigations have included the use of both swell/consolidation and soil
suction measurements for prediction of remaining heave. Since 1996,
CTL/Thompson, Inc. has also continued to obtain both swell/consolidation
and soil suction measurements during design-level soils investigations
(McOmber and Thompson, 2000). These investigations have resulted in the
accumulation of new soil suction data which can be utilized to update and
improve the understanding of soil suctions trends and the depth of wetting in
the Denver area. Ultimately, analysis of this new suction data will provide
better understanding of potential heave which, when combined with
engineering judgement, will improve design recommendations.
3.1 Pre-Construction Suction Data
Pre-construction laboratory test data were obtained from 14 design-
level soils investigations in the Denver area where expansive soils were
encountered. These investigations typically consisted of multiple exploratory
borings drilled to depths between 15 feet to 35 feet below existing ground
45


surface with a truck-mounted drill rig and 4-inch continuous flight auger. Soil
and bedrock samples were obtained at 5 feet intervals by driving a modified
California sampler (2.5-inch O.D.) into the soils and bedrock using a 140-
pound hammer falling 30 inches. After the samples were logged in the field,
they were returned to the laboratory where they were identified by a
geotechnical engineer and set-up for testing. Laboratory testing included
swell/consolidation, soil suction, moisture content, dry density, gradation,
soluble sulfates and Atterberg Limits.
Laboratory test results, soil or bedrock type and boring logs were
collected from CTL/Thompson, Inc. records for each of the 14 post-
construction design-level investigations. This information was then combined
with the data from the 51 pre-construction sites used in McOmber and
Thompsons (2000) evaluation of the repair protocol. The data were
compiled in Microsoft Excel spreadsheets. A total of 396 pre-construction
soil suction measurements were used in the analysis. Appendix A contains
the Excel spreadsheets of the raw pre-construction data.
3.2 Post-Construction Suction Data
During the period from February 1998 to September 2002, 133
warranty-type investigations were performed at residential sites in the Denver
area where heave distress had occurred. These investigations typically
consisted of drilling one or two exploratory borings with a truck-mounted or
limited access drill rig and 4- or 2.5-inch continuous flight auger to depths of
20 feet to 40 feet below existing grade. In most cases, soil samples were
obtained at 5 feet intervals by driving a modified California sampler (2.5-inch
46


O.D.) into the soils and bedrock using a 140-pound hammer falling 30
inches. Occasionally samples were also obtained by pushing thin-walled
Shelby tube samplers into the soils. After the samples were logged in the
field, they were returned to the laboratory where they were identified by a
geotechnical engineer and set-up for testing. Laboratory testing included
swell/consolidation, soil suction, moisture content, dry density and Atterberg
Limits.
Laboratory test results, date of drilling, soil or bedrock type, boring
logs, moisture profiles and suction profiles were collected from
CTL/Thompson, Inc. records for each of the 133 post-construction warranty-
type investigations. This information was then combined with the data from
the 41 post-construction residential sites used in McOmber and Thompsons
(2000) evaluation of the repair protocol. The data were compiled in Microsoft
Excel spreadsheets. A total of 945 post-construction soil suction
measurements were used in the analysis. Appendix B contains the Excel
spreadsheets of the raw post-construction data.
47


4. Influence of Soil Type on Denver Area
Suction T rends
Subsoils from both pre- and post-construction investigation were
classified according to the Unified Soil Classification System, which is based
on the percent of fines (passing the #200 sieve), Liquid Limit (LL) and
Plasticity Index (PI). Results of field penetration test (blow counts) and soil
structure were used to classify subsoils as overburden or bedrock. For the
purpose of determining soil suction profiles for different soil types, all clay
soils (greater than 50 percent fines) were grouped together. Therefore the
suction profile for clay soil represents suction measurements from both lean
and fat clays (CL and CH), as well as weathered claystone samples. The
suction profile for claystone represents suction measurements from both
claystone and interbedded claystone/sandstone samples. The profile for
sands and sandstone represents suction measurements for clayey to silty
sands (SC and SM) and sandstone bedrock. Sand and sandstone samples
are not well represented in the data set since most of the data was obtained
from expansive soil sites where clays and claystone bedrock are most
prevalent.
4.1 Pre-Construction Suction Profiles by Soil Type
The figures presented in this section represent soil suction
measurements from design level soils investigations in the Denver area.
Figure 4.1 shows all 396 suction vs. depth data points and the average fit
line at 5 foot intervals. Soil suction values for pre-construction conditions are
48


between 3.9 to 4.4 pF, with a few values outside of this range. The average
pre-construction soil suction represented by the line is about 4.18 pF.
Figures 4.2, 4.3 and 4.4 show the Pre-construction soil suction
measurements and regression log fit profiles for clay, claystone and sand
and sandstone soil types. The line equations used to represent the suction
profiles of each soil type were developed using Microsoft Excel to perform
regression analysis of soil suction as a function of depth with a natural
logarithm relationship. This method was used to fit a line to the suction
data in an effort to account for all suction data of a particular soil type and
has been established by McOmber and Thompson (2000). Due to the
relatively small data set of sand and sandstone suction measurements, these
two soil types were combined to be represented by one profile.


Suction (pF)
2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5
Figure 4.1
Pre-Construction Suction Measurements of
All Soil and Bedrock with Average Fit Line
50


Suction (pF)
2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5
10
^ 15
a.
a>
a
20
25
30
35
a M AAA A i AA AA
>< II o o )79Ln(x) h R2=( 4.1968 .0003 A 3 4 4 4 4 k Mi
MAI kAA AMAA m
A M A
Mi , A i A

A Pre- "Log Fit
A

Figure 4.2
Pre-Construction Suction Measurements for Clay Soil
with Logarithmic Fit Line
51


2.0 2.5 3.0
Suction (pF)
3.5 4.0 4.5
5.0 5.5
Figure 4.3
Pre-Construction Suction Measurements for Claystone Bedrock
with Logarithmic Fit Line
52


Suction (pF)
2.0 2.5 3.0 3.5 4.0 4.5 5.0
Figure 4.4
Pre-Construction Suction Measurements for Sand
and Sandstone with Logarithmic Fit Line
53


The soil suction profiles for the clay soil and the claystone bedrock,
represented on Figures 4.2 and 4.3, are almost vertical lines indicating that
soil suction is relatively constant with depth for pre-construction conditions.
The log fit for pre-construction suction measurements ranges from 4.21 to
4.22 pF for the clay soils, and 4.22 pF near the surface (5 feet deep) to 4.16
pF at a depth of 35 feet for the claystone bedrock. The suction values
represented by the updated pre-construction profiles for the clay soil and the
claystone bedrock are within or very close to the suction limits for Denver
area soils suggested by Thompson (1997). The log fit for the combined sand
and sandstone bedrock suction measurements is shifted to the left at 3.8 to
3.96 pF. This shift in suction is typical of sandy soils as compared to clay
soils and claystone, and shows the necessity of sorting suction
measurements by soil type for trend analysis.
In order to better represent a pre-construction suction profile for
expansive soils, the suction data for the clay soils and claystone bedrock
were combined as shown on Figure 4.5. Pre-construction soil suction
measurements generally varied between 3.8 pF to 4.5 pF, with a few suctions
outside of this range. The log fit line for the combined clay and claystone
profile ranged from 4.22 pF near the surface to 4.17 pF at depth. Therefore it
appears that the suction profile is controlled more by the clays near the
surface and by the claystone at depth. This is what would be expected given
soil profiles in the Denver area. The pre-construction suction profile for
expansive soils represents a suction equilibrium based on Denver area
geologic and climatic conditions (McKeen, 1992).
54


The dashed line on either side of the log fit regression line for the pre-
construction clay and claystone suction data shown on Figure 4.6 represent
plus and minus one standard deviation of the log fit line. One standard
deviation lines are used to help account for the variability of the suction data.
There appears to be slightly more variability in the soils above 15 feet than
for those below 15 feet.
55


Suction (pF)
2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5
Figure 4.5
Pre-Construction Suction Measurements for Clay
and Claystone with Logarithmic Fit Line
56


Suction (pF)
Figure 4.6
Pre-Construction Suction Measurements for Clay and Claystone
with Plus and Minus One Standard Deviation
57


4.2 Post-Construction Suction Profiles
by Soil Type
The figures presented in this section represent soil suction
measurements from post-construction warranty-type soils investigations in
the Denver area. Figure 4.6 shows all 945 suction vs. depth measurements
and the average fit line. The majority of the post-construction suction
measurements are between 3.0 to 4.5 pF, with a few data points outside of
this range. The average post-construction soil suction represented by the
line varies from about 3.4 pF near the surface to 4.2 at a depth of 40 feet.
Figures 4.7, 4.8 and 4.9 show the post-construction soil suction
measurements and regression log fit profiles for clay, claystone and sand
and sandstone soil types. The line equations used to represent the suction
profiles of each soil type were developed using Microsoft Excel to perform
regression analysis of soil suction as a function of depth with a natural
logarithm relationship. This method was used to fit a line to the suction
data in an effort to account for all suction data of a particular soil type and
has been established by McOmber and Thompson (2000). Due to the
relatively small data set of sand and sandstone suction measurements, these
two soil types were combined into one profile. Post-construction sand and
sandstone suction measurements are not well represented in this study
because the majority of the warranty-type investigation sites were underlain
by expansive clays soils and claystone or interbedded claystone/sandstone
bedrock.
58


Suction (pF)
2.0 2.5 3.0 3.5 4.0 4.5 5.0
Figure 4.7
Post-Construction Suction Measurements of All Soil
and Bedrock with Average Fit Line
59


2.0
2.5
Suction (pF)
3.0 3.5 4.0
4.5
5.0
Figure 4.8
Post-Construction Suction Measurements for Clay Soil
with Logarithmic Fit Line
60


Suction (pF)
2.0 2.5 3.0 3.5 4.0 4.5 5.0
Figure 4.9
Post-Construction Suction Measurements for Claystone
Bedrock with Logarithmic Fit Line
61


Suction (pF)
2.0 2.5 3.0 3.5 4.0 4.5 5.0
Figure 4.10
Post-Construction Suction Measurements for Sand and
Sandstone with Logarithmic Fit Line
62


The soil suction profile for the clay soil represented on Figure 4.8 is
limited to soils above 24 feet below the ground surface. Clay soils were not
present below this depth at the post-construction sites. The log fit for post-
construction suction measurements ranges from 3.49 pF near the surface (5
feet deep) to 3.82 pF at a depth of 25 feet for clay soil, and 3.96 pF at
shallow depth (8 feet deep) to 4.16 pF at a depth of 40 feet for claystone
bedrock. The suction values for post-construction conditions, represented in
the profiles for the clay soil and the claystone bedrock, are within or very
close to the suction limits for Denver area soils suggested by Thompson
(1997). The log fit for the combined sand and sandstone bedrock suction
measurements is a nearly vertical line at 3.73 pF to 3.65 pF.
In order to better represent a post-construction suction profile for
expansive soils, the suction data for the clay soils and claystone bedrock
were combined as shown on Figure 4.11. Post-construction soil suction
measurements for the combined clay and claystone generally varied between
3.0 pF to 4.5 pF, with a few suctions outside of this range. The shape of the
post-construction combined clay and claystone profile shows soil moisture
changes have cause a shift toward lower suction measurements above 20 to
25 feet. The log fit line for the combined clay and claystone profile ranged
from 3.47 pF near the surface to 4.21 pF at depth. This trend is indicative of
moisture and suction variations within the active zone as described by
McKeen (1992).
The dashed line on either side of the log fit regression line for the
post-construction clay and claystone suction data shown on Figure 4.12
63


represent plus and minus one standard deviation of the log fit line. One
standard deviation lines are use to help account for the variability of the
suction data.
64


Suction (pF)
2.0 2.5 3.0 3.5 4.0 4.5 5.0
Figure 4.11
Post-Construction Suction Measurements for Clay and
Claystone with Logarithmic Fit Line
65


Suction (pF)
Figure 4.12
Post-Construction Suction Measurements for Clay and Ciaystone
with Plus and Minus One Standard Deviation
66


4.3 Comparison of Pre- and Post-Construction
Suction Profiles
McOmber and Thompson (2000) have shown that a comparison of the
pre- and post-construction suction profiles is an effective method to
determine the depth to constant suction for Denver area soils. The depth to
constant suction can be interpreted as the depth at which the pre- and post-
suction profiles intersect. Plots of the pre- and post- construction suction
profiles for the clay soil and the claystone bedrock, represented by the
updated suction data, are shown on Figures 4.13 and 4.14
The pre- and post-construction suction profiles for clay soil (Figure
4.13) do not intersect between depths of 4 to 25 feet below the ground
surface. This indicates that analysis of clay soil alone does not provide
sufficient data to determine depth to constant suction. The pre- and post-
construction profiles for claystone bedrock intersect around 35 feet below the
ground surface.
The typical soil profile for expansive soil sites in the Denver area
consists of clay soil over claystone or interbedded claystone/sandstone
bedrock. Therefore, in order to determine the depth to constant suction for
expansive soil sites, the pre- and post- construction suction profiles for
combined clay and claystone were plotted together as shown on Figure 4.15.
67


Suction (pF)
2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5
Figure 4.13
Comparison of Pre- and Post-Construction
Suction Profiles for Clay Soil
68


Suction (pF)
2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5
0
10
15
B 20
Q.
Q
25
30
35
40
AAA A A A
A AAAJ aa
o oo oso 000 o & osmm lo A , O mbao
o O OO 0 o o o <> O0BMH o o o o o ft A
o o < * 0 >1 Soooo o O EGA A
o o o
a ooe&tt
- o Post- Log Fit Post- a Pre- Log Fit Pre- o
oo ooo 30 O
I O 00
Figure 4.14
Comparison of Pre- and Post-Construction
Suction Profiles for Claystone Bedrock
69


Suction (pF)
Figure 4.15
Comparison of Pre- and Post-Construction
Suction Profiles for Clay and Claystone
70


The comparison of pre- and post-construction suction profiles for clay
and claystone show an intersection around 34 feet below the ground surface.
The inherent variability in the range of suction values when dealing with a
large data set and the fact that the majority of the suction data was obtained
from soil samples at 5 feet intervals from depths of 4 to 39 feet makes it
difficult to determine a depth to constant suction to the nearest foot.
4.3.1 Variation in Depth to Constant Suction
To help account for variability in the suction data, a comparison of the
pre- and post-construction suction profiles for clay and claystone with plus
and minus one standard deviation lines were plotted on Figure 4.16. This
comparison shows the plus one standard deviation lines intersecting around
a depth of 28 feet and the minus one standard deviation lines intersecting
around 34 feet. This comparison of the standard deviation profiles for the
pre- and post-construction conditions suggests a depth to constant suction of
28 to 34 feet. The depth to constant suction indicated by the updated soil
suction data shows an increase in the depth of wetting as compared to the
depth of wetting of 16 to 22 feet, indicated by Thompson (1997); and of 21
feet shown by McOmber and Thompson (2000) in Figure 2.8.
71


Suction (pF)
2.0 2.5 3.0 3.5 4.0 4.5 5.0
Figure 4.16
Comparison of Plus and Minus One Standard Deviation of
Pre- and Post-Construction Profiles for Clay and Claystone
72


4.3.2 Suggested Adjustment of Final Suction
Profile
The original final suction profile estimated by Thompson (1997) was
based upon a depth of wetting of 20 feet. This study has shown that the
depth to constant suction indicated by the updated suction data is around 28
to 34 feet. Figure 4.17 shows the comparison of pre- and post-construction
suction profiles for clay and claystone together with the original final suction
profile as estimated by Thompson (1997). Analysis of the post-construction
suction measurements as compared to the original final suction profile
indicates that the final suction profile falls below, i.e. to the left of, about 55
percent of the suction data which indicates that the final suction profile
should be adjusted to represent an increase in the depth to constant suction.
A suggested adjustment to the final suction profile is shown on Figure
4.18. The adjustment to the final profile was based upon an estimated depth
to constant suction of 30 feet. The final suction values at 4 to 10 feet remain
at 3.5 pF and suction values at 30 to 40 feet remain at 4.2 pF as estimated
by Thompson (1997). However, the final suction values between 10 to 30
feet were shifted to the left in order to better represent the increase in the
depth of wetting. The adjusted final suction profile falls below, i.e. to the left
of, about 72 percent of the post-construction suction data and below about 93
percent of the pre-construction data.
73


Suction (pF)
Figure 4.17
Comparison of Pre- and Post-Construction Suction Profiles
and Original Final Suction Profile
74


Suction (pF)
Figure 4.18
Comparison of Pre- and Post-Construction Suction Profiles
and Adjusted Final Suction Profile
75


5. Influence of Age of Structure on Soil
Suction Trends
Another task of this study was to determine what influence the age of
the residential structure has on the suction trends and the depth to constant
suction. Theoretically older structures which have been subject to longer
periods of irrigation or wetting should show a trend in which the suction
profile is shifted to the left indicating lower suction measurements, i.e.
increased moisture in the subsurface soils. In order to determine if this trend
is occurring, the age of the residence for each of the 174 expansive soil sites,
represented by the McOmber and Thompson (2000) study and the updated
data collected for this study, was obtained by comparing the date of
investigation to the date of construction. Soil suction data were then sorted
according to the age of structure at the time of investigation. Figure 5.1
shows the age distribution, in years, for the Denver area expansive soil sites.
The graph shows a peak in the age distribution around 3 years old, therefore
the suction data are some what skewed towards residences that were 5
years old or younger.
5.1 Age of Structure Suction Profiles
The suction data were grouped in approximate three year intervals (1
to 3 years, 4 to 6 years and 7 to 10 years) to better distribute the suction data
and generate log fit profiles. Figure 5.2 shows the suction measurements
and profiles for each of these age groups. A comparison of the suction
profiles by age indicates that there is a slight shift to the left in the 4 to 6 year


age group as compared to the 1 to 3 year age group. However the profile for
the older group (7 to 10 years) shows a shift to the right above 15 feet and to
the left below 15 feet. The theory that older residential sites should show
lower suction values with time, i.e. a shift to the left,
is supported by the suction data from residences 6 years old and younger,
but not by the data from residences 7 years or older.
77


Age Distribution of Data
Age of House (years)
Figure 5.1
Age Distribution of Suction Data
78


2.0 2.5
Suction (pF)
3.0 3.5 4.0 4.5 5.0
0
10
15
£ 20
cx
0)
25
30
35
40
(2)(lj> (3)
1.0

a
y = 0.2935Ln(x) +3.151
R2 = 0.3206
V = 0.2977Ln(x) + 3.1085
R2= 0.1926
y = 0.2126Ln(x) + 3.3354
R-=0.T5T4"
rP
-G....!--
1
Ba
in ft i Bin j ^ n
!<3E3 OO
O O
BCOOD BDStSBGX'
j-B.EB
o o cooBiUpecsss' o
IOO
o oai
o a o o
t
OOQ
O O
1-3 years
o 4-6 years
7-10 years
(1) 1-3 yrs (Log)
(2) 4-6 yrs (Log)
(3}_ 7-10 yrs (Log)
In o JO
O I ] a OEffiM
a bo a ana ^SSQKflBSR} a o
Figure 5.2
Suction Profiles for Clay and Claystone
According to Age of Structure
79


5.2 Variation in Depth to Constant Suction
According to Age of Structure
In order to determine the influence of age on the depth to constant
suction, the three suction profiles for each age group were compared to the
pre-construction suction profile for clay and claystone as shown on Figure
5.3. This comparison shows an approximate depth to constant suction of 32
feet for sites where residences were 1 to 3 years old; of 35 feet for sites
where residences were 4 to 6 years old; and of around 40 feet or more where
residences were 7 to 10 years old. This indicates that there is an increase in
the depth to constant suction over time, thus showing deeper moisture
migration at expansive soil sites with older residences.
80


2.0 2.5
Suction (pF)
3.0 3.5 4.0
4.5 5.0
10
15
B 20
Q.
a
25
30
35
40
o 1-3 years
o 4-6 years
7-10 years
(1) 1-3 yrs (Log)
(2) " 46 yrs (Log)
(^)7-10 yrs (Log)
Pre-Construction
(Log)
Figure 5.3
Suction Profiles for Clay and Claystone According to Age
of Structure Compared to Pre-Construction Suction Profile
81


6. Summary and Conclusions
6.1 Summary
A study was undertaken to propose a modified protocol for soil suction
heave prediction. The protocol was based on the original protocol
developed by Thompson (1997) and updated pre- and post-construction soil
suction data, from Denver area expansive soil sites.
The five main research tasks of this study include:
1. Summarize the advantages and methods of using soil suction to
estimate potential heave;
2. Collect and compile updated pre- and post-construction suction
measurements obtained from warranty-type and design level
investigations during the past four years;
3. Collect the date of construction and date of investigation for
each of the warranty-type investigations;
4. Determine the influence of soil type and age of the structure on
the soil suction trends and depth to constant suction; and
5. Adjust the final suction profile to better represent any changes
in the updated suction profiles and depth to constant suction.
The first task was accomplished by review of previous studies
involving soil suction and heave prediction. The various soil suction test
methods were evaluated for their applicability to consulting engineering
practice. The soil suction method was compared to the consolidation swell
method for use in heave prediction at Denver area expansive soil sites.
82


The second task was conducted through obtaining pre-and post-
construction soil suction measurements together with depth, soil type and
other soil parameters from 14 updated Denver area design level
investigations and 133 updated warranty-type investigations conducted by
CTL/Thompson, Inc.
The third task was accomplished through determination of age of the
structure by obtaining the date investigation and date of construction for each
residential site.
The fourth task was achieved through sorting the pre- and post-
construction suction data by soil type and age of structure. Suction profiles
were generated for each soil type and age group. The difference in suction
measurements for clay/claystone and sand/sandstone soils made it
necessary to utilize only clay and claystone suction values to determine a
depth to constant suction for Denver area expansive soil sites. A depth to
constant suction was estimated from the comparison of the pre- and post-
construction profiles. The influence of time on the suction trends and depth
to constant suction was determined through analysis of the suction data by
age of the residence.
The fifth task was accomplished through adjusting the final suction
profile to represent an increase in depth of wetting shown by the updated
suction data. The adjustment to the final suction profile will help improve
heave prediction for design and repair purposes in the Denver area.
83


6.2 Conclusions
A modified soil suction heave prediction protocol has been developed
based upon adjustment to the final suction profile. The suggested
adjustment was made using a depth to constant suction estimated by
analysis of soil suction data by soil type only. This study has produced the
following conclusions:
1. Soil suction is indeed a valuable tool for prediction of soil moisture, soil
behavior and potential heave:
The filter paper test has been shown to be a simple and practical
method to perform suction test for consulting engineering practice. The
estimation of a final suction profile is an essential component in the
calculation of potential heave. The McKeen soil suction method of heave
prediction yields similar heave estimates as the swell consolidation method.
2. Suction measurements from sand and sandstone samples should not be
used in analyzing suction trends at expansive soil sites:
The sand and sandstone suction profile exhibits lower suction values
around 3.6 to 3.7 pF which tends to shift the profile of all soil and bedrock to
the left and outside of the reasonable range of suction values at expansive
soil sites. Clay and claystone suction measurements should be combined to
determine a suction profile for expansive soil sites.
3. The depth to constant suction (depth of wetting) has increased by 7 to 13
feet compared to previous studies of Denver area expansive soil sites:
The updated suction data for combined clay and claystone shows a
depth to constant suction of 28 to 34 feet below the ground surface.
84


4. The final suction profile used for heave prediction should be adjusted to
account for the increase in depth of wetting:
The final suction profile was shifted to the left and down between the
depths of 10 and 30 feet to represent an estimated depth of constant suction
of 30 feet, otherwise the shallow and deep final suction values remained the
same as the original final suction profile.
5. The depth to constant suction increases with time after construction:
Suction profiles for residential sites grouped in 3 years intervals show
an increase in the depth to constant suction from 32 to 35 to 40 feet below
the ground surface, when compared with the pre-construction profile for clay
and claystone.
The results of this study were based upon analysis of soil suction data
by soil type only. There are several other factors, as discussed in the
Literature Review of this thesis, which influence soil suction trends and the
depth to constant suction. This suction data should also be analyzed by
some of these other factors prior to implementation of the adjusted final
suction profile to heave prediction for the purpose of design and repair of
structures built on expansive soils.
6.3 Suggestions for Future Study
Soil suction remains a relatively new tool in consulting engineering
practice. As more pre- and post-construction investigations utilize suction
measurements, more data will become available for future studies to better
understand soil/moisture interactions within the subsurface of Denver area
expansive soil sites. This data should be compiled and analyzed periodically
to interpret suction trends and validate depth to constant suction for
estimation of the final suction profile.
85


Perhaps the suction data can be analyzed by other factors to better
represent the swell potential, i.e. risk of heave. For example, sorting the soil
according to the McKeen Classification System, which uses suction and
moisture to group soils into five different categories of swell risk, might be a
good method.
This study was limited to data from residential sites which were 10
years in age or younger. Future studies could utilize suction data from
building sites older than 10 years to determine the influence of many years of
wetting on the suction trends and depth of wetting. Eventually there may be
a new soil/moisture equilibrium established after wetting has occurred for
numerous years.
86


Appendix A
Raw Pre-Construction Suction Data
87


TOTAL PRE-CONSTRUCTION DATA
Drill Date DeDth ffi) Suction ton Soil Tvpe
May-97 1 4.265 CLAYSTONE
Sep-96 4 3.380 SAND, SILTY (SM)
Sep-96 4 4.050 CLAY, SANDY (CL)
Sep-96 4 4.360 CLAY, SANDY (CL)
Sep-96 4 3.770 SAND, SILTY (SM)
Dec-96 4 4.230 FILL, CLAY, SANDY
Dec-96 4 4.370 CLAY, SANDY (CL)
Dec-96 4 4.170 FILL, CLAY, SANDY
Dec-96 4 4.250 CLAYSTONE
Dec-96 4 4.470 CLAY, SANDY (CL)
Dec-96 4 4.500 CLAY, SANDY (CL)
Nov-96 4 4.240 CLAY, SANDY (CH)
Nov-96 4 4.030 FILL, CLAY, SANDY
Nov-96 4 4.450 CLAY, SANDY (CH)
Nov-96 4 4.430 CLAY, SANDY (CH)
May-97 4 4.245 FILL, CLAY, SANDY
May-97 4 3.952 CLAYSTONE
May-97 4 4.076 CLAYSTONE
May-97 4 4.260 CLAYSTONE
May-97 4 4.290 CLAYSTONE
May-97 4 4.371 CLAYSTONE
Aug-97 4 4.237 WEATHERED CLAYSTONE
Aug-97 4 3.725 CLAY, SANDY (CL)
Aug-97 4 3.473 CLAY, SANDY (CL)
Aug-97 4 4.030 CLAY, SANDY (CL)
Aug-97 4 3.969 CLAY, SANDY (CL)
Aug-97 4 3.537 CLAY, SANDY (CL)
Apr-97 4 4.098 FILL, CLAY, SANDY
Apr-97 4 3.771 CLAY, SANDY (CL)
Apr-97 4 4.013 INTERBEDDED CLAYSTONE/SANDSTONE
Apr-97 4 4.126 WEATHERED CLAYSTONE/SANDSTONE
Jul-97 4 4.112 CLAY, SANDY (CL)
Jul-97 4 4.393 WEATHERED CLAYSTONE
Jul-97 4 4.044 WEATHERED CLAYSTONE
Jul-97 4 4.092 CLAY, SANDY (CL)
Jul-97 4 4.460 CLAY, SANDY (CL)
Oct-97 4 4.121 CLAYSTONE
Oct-97 4 4.248 CLAYSTONE
Oct-97 4 4.233 CLAYSTONE
Oct-97 4 4.017 FILL, CLAY, SANDY
Oct-98 4 4.300 CLAYSTONE
Oct-98 4 4.010 CLAY, SANDY (CL)
88