See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/346940180

Calculation of Heave of Deep Pier Foundations

Conference Paper · October 2011

CITATIONS READS

0 262

4 authors, including:

John D. Nelson Kuo Chieh Chao

Colorado State University Asian Institute of Technology
98 PUBLICATIONS   827 CITATIONS    81 PUBLICATIONS   409 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Foundation failure on Collapsible Soil View project

Expansive Soils View project

All content following this page was uploaded by Kuo Chieh Chao on 11 December 2020.

The user has requested enhancement of the downloaded file.

 Unsaturated Soils: Theory and Practice 2011
Jotisankasa, Sawangsuriya, Soralump and Mairaing (Editors)
Kasetsart University, Thailand, ISBN 978-616-7522-77-7

Calculation of heave of deep pier foundations

Calculation of heave of deep pier foundations
J.D. Nelson
Engineering Analytics, Inc., Fort Collins, Colorado, USA, [email protected]
J.D. Nelson
K.C. Chao
Engineering Analytics,
Engineering Analytics,Inc.,
Inc.,Fort
FortCollins,
Collins,Colorado,
Colorado,USA, [email protected]
USA, [email protected]
K.C. Chao
D.D. Overton
Engineering Analytics, Inc., Fort Collins, Colorado, USA, [email protected]
Engineering Analytics, Inc., Fort Collins, Colorado, USA, [email protected]
D.D. Overton
R.W. Schaut
Engineering Analytics, Inc., Fort Collins, Colorado, USA, [email protected]
Engineering
R.W. Schaut Analytics, Inc., Fort Collins, Colorado, USA, [email protected]
Engineering Analytics, Inc., Fort Collins, Colorado, USA, [email protected]

ABSTRACT Pier and grade beam foundations are a commonly used foundation type in highly expansive
soils. The design of such foundations is one of the most important and challenging aspects of geotechnical
engineering. Few methods have been developed to enable the calculation of pier heave. Existing methods
consider only uniform soil profiles, and piers with limited length to diameter ratios. These methods are re-
stricted with regard to evaluation of more complex aspects of pier heave. A finite element method of analy-
sis was developed to compute pier movement in expansive soils having variable soil profiles, complex wet-
ting profiles, large length-to-diameter ratios, and complex pier configurations and materials. The model has
been named APEX (for Analysis of Piers in EXpansive soils). This paper describes the method of analysis
and demonstrates its validity using several case histories. The results of pier design computations using the
APEX model are compared with those of both conventional rigid pier analyses and elastic pier analyses. A
series of simplified design charts developed using APEX are presented to facilitate its use. The results show
the versatility of the model with regard to variable soil profiles and wetting zones. Piers of different materi-
als such as sleeved or cased piers are considered and the effects of adhesion factors are demonstrated.

KEYWORDS: Expansive soils, deep foundation, pier design, heave prediction, finite element method

1 INTRODUCTION on finite element analyses by Poulos and

Davis (1980). This method, however, was devel-
Foundations on expansive soils pose a unique chal- oped for uniform soil profiles and piers with limited
lenge to the geotechnical engineer. They almost al- length to diameter ratios. This method is limited in
ways cost more than foundations on ordinary soils, its application to more complex soil profiles and
and the site investigation and foundation design pier geometry. More recently a finite element
costs more as well. method of analysis was developed to compute pier
The Expansion Potential, EP, of an expansive movement in expansive soils having variable soil
soil is defined on the basis of percent swell exhib- profiles, complex wetting profiles, large length-to-
ited in a consolidation-swell test and the swelling diameter ratios, and complex pier configurations
pressure of the soil (Nelson et al., 2007). For sites and materials. The model has been named APEX
having a low to moderate EP, a variety of founda- (for Analysis of Piers in EXpansive soils). This pa-
tion systems have been proposed and used. How- per describes that model and demonstrates its valid-
ever, for sites with high to very high expansion po- ity using case histories. The results of pier design
tential the most reliable method for foundation computations using the APEX model are compared
design is the use of pier and grade beam founda- with those of both conventional rigid pier analyses
tions. The grade beam forms a structural member of and elastic pier analyses. A series of simplified de-
the foundation and must be designed to mitigate the sign charts developed using APEX are presented to
effects of differential pier movement on the super- facilitate its use. The results show the versatility of
structure. The grade beam may be in the form of a the model with regard to variable soil profiles and
reinforced basement wall or a stiff beam supported wetting zones.
by the piers. Thus, an important part of the design of
the foundation is the calculation of the heave of the
individual piers. 2 FREE-FIELD HEAVE PREDICTION
Few methods have been developed to enable the
calculation of pier heave. A readily available one is Free-field heave is the fundamental parameter on
presented in Nelson and Miller (1992). It is based which pier heave is calculated. Therefore, a review

59
of free-field heave calculation will be presented of the curve shown in Figure 1 is the compression
first. index, Cc, and that of the rebound portion of
the

Figure 2. Vertical stress states in soil profile.

Figure 1. Terminology and notation for oedometer tests.

Free-field heave is the heave that will occur at the

surface of a soil profile if no surcharge or stress is
applied. Methods for calculating free-field heave
have been developed that use either oedometer tests
or soil suction tests. The authors believe that the
method based on oedometer test results has particu-
lar advantages and that is the method used in this
paper.

2.1 Prediction of free-field heave by oedometer

method
A method for prediction of free-field heave using Figure 3. Hypothetical oedometer test results for stress states
oedometer test data was outlined in Nelson and Mil- shown in Figure 2.
ler (1992). A refinement of that method is pre-
sented in Nelson et al. (2006) and was presented in curve is the rebound index, Cs. The volumetric
a panel discussion at the UNSAT 2010 conference strain experienced during inundation is the percent
(Nelson, 2010). The general equation for heave of a swell, %S.
soil layer of thickness, zi, is: Figure 2 shows the vertical overburden stress at
three different depths in a soil profile with similar
 σ  soil throughout. At all points all samples are in a
ρ  C H z i log f 
 (1) condition of zero lateral strain with a vertical over-
 σ cv i burden stress equal to σ′vo. If a consolidation-swell
where:  = free-field heave, test is conducted on a sample identical to that at
CH = heave index, depth, ZA, at an inundation stress, (σ′i)A = (σ′vo)A, the
f = in-situ effective stress state, sample will swell by an amount %SA as shown in
cv = swelling pressure from the constant- Figure 3. Similarly, for a sample at depth ZB, the
volume oedometer test; and sample would be subjected to an inundation stress,
zi = layer thickness. (σ′i)B = (σ′vo)B, and the sample would swell by an
amount %SB. Obviously, if a sample is tested at an
inundation pressure equal to the constant-volume
2.2 Determination of heave index, CH swelling pressure, σ′cv, the sample will not swell and
The heave index, CH, can be determined from con- the test data would define point C in Figure 3.
solidation-swell test data along with data from con- To arrive at the form of Equation (1), it is con-
stant-volume tests. Because, the constant-volume venient to start with the general equation for heave
test data can be approximated from consolidation- in a soil stratum of thickness, zi. That is,
swell test data, all of the soil property data that is ρ  ε v  Δz i  %S  Δz i (2)
needed can be obtained from that test alone.
An example of consolidation-swell test data is
shown in Figure 1. The slope of the loading portion

60
 For a layer of thickness, zi, the overburden pres- tained from the consolidation-swell test with point
sure and applied stress will influence the amount of C obtained from the constant volume test. How-
swell that will occur when that layer becomes wet- ever, to do so is generally not practical, mainly be-
ted. Because that stress varies from layer to layer, it cause it is almost impossible to obtain two identical
is necessary to define a relationship between the samples from the field. A relationship between σ′cv
amount of swell that occurs and the imposed stress and σ′cs has been proposed so that the value of the
when the soil is wetted, i.e., the inundation pres- heave index can be determined from a single con-
sure. That relationship is defined by the line ABC solidation-swell test (Nelson et al., 2006).
in Figure 3. For practical purposes, the line ABC
can be defined by a straight line connecting point A
(the point defined by the percent swell in a consoli- 2.3 Relationship between ′cv and ′cs
dation-swell test) and point C (the point corre- A relationship between σ′cv and σ′cs has been pro-
sponding to the constant volume swell pressure, posed that is of the form.
σ′cv. The slope of that line is denoted by the heave
index, CH, where: σ' cv  σ' i  λ(σ'cs σ' i ) (6)
%SA %SA
(3) The rationale behind this equation is that the val-
CH  
log σ'cv  logσ'i A  σ' cv  ue of σ′i must be less than σ′cv, otherwise heave
log  would not have occurred. Also it is reasonable to
 ( σ'i ) A  expect that σ′cv will be less than σ′cs. Equation (6)
If values of CH and σ′cv are known, the vertical proposes that the value of σ′cv falls between σ′i and
strain, or percent swell, that will occur during inun- σ′cs by the proportionality defined by the value of λ.
dation at any depth z in a soil profile can be deter- The authors’ experience has indicated that a reason-
mined from Equation (3). For the case of free field able value for λ is 0.6 for the clay soil in the Front
heave when the soil at depth z is inundated, the Range area of Colorado, USA. Since the value of λ
stress on the soil is the overburden stress, (σ′vo)z. is dependent upon mineralogy of the clay soil, the λ
This value is, therefore, the inundation stress, σ’i, in value should be determined for soil on a regional
the field and Equation (3) can be rewritten as Equa- basis.
tion (4).
 σ'  2.4 Depth of wetting/ degree of wetting
(ε v ) z  %S z  C H log  cv  (4)
 σ' vo z 
The depth of soil that is contributing to heave at a
particular point of time depends on two factors.
Therefore, for a layer of soil of thickness, zi, These are the depth to which water contents in the
that exists at a depth z to its midpoint, the maximum soil have increased since the time of construction,
heave that will occur due to expansion of that stra- and the expansion potential of the various soil
tum during complete inundation would be obtained strata. As water migrates through a soil profile dif-
by substituting Equation (4) into Equation (2). ferent strata become wetted, some of which may
Thus, have more swell potential than others. Conse-
quently, the zone of soil that is contributing to
 σ'  heave varies with time.
ρ oi  C H  Δzi log cv  (5)
 σ' vo  z 
The amount of heave that will occur at a particu-
lar time depends on the manner in which the
In actual application of Equation (5), a soil pro- groundwater migrates in the soil and the expansion
file will be divided into layers of thickness, z, the potential of the soil at depth. Movement of the soil
value of heave for each layer will be computed, and surface will begin almost immediately after con-
the incremental values will be added to determine struction, whereas some time will be required for
the total heave. It should be noted that the value of the soil at deeper depths to become wetted. Thus,
(σ′vo)z to be used in Equation (5) is the stress at the the surface of the soil will begin to heave almost
midpoint of the layer at depth z. In a soil with no immediately, but movement of piers will be de-
applied load, that value would be the overburden layed, sometimes by several years.
pressure. If a load is applied to the soil, for exam- The term “active zone” has been in common us-
ple, by a footing or stiffened slab, the value of (σ′vo)z age in the field of expansive soils. However, the
to be used in Equation (5) is the overburden stress usage of that term has taken different meaning at
plus the applied stress. different times and in different places. Therefore,
The line ABC in Figure 3 and the heave index, for purposes of clarity and consistency, the follow-
CH, can be determined by conducting consolida- ing five definitions have been put forth.
tion-swell and constant volume oedometer tests on Active Zone, ZA, is that zone of soil that is con-
samples of the same soil and connecting point A ob- tri-buting to heave due to soil expansion at a par-
ticular point in time. The depth of the active zone

61
will vary as heave progresses, and therefore, it var- wetting will occur to below the depth of potential
ies with time. heave.
Zone of Seasonal Moisture Fluctuation, Zs, is
that zone of soil in which water contents change 2.4.2 Vadose zone modeling
seaso-nally due to climate changes. For sites at which the depth of potential heave is
Zone of Wetting, Zw, is that zone in which water large, the degree of wetting typically will be less
contents have increased over the pre-construction than fully saturated (Chao et al., 2006; Overton et
equilibrium conditions. Factors contributing to this al., 2006). Design of foundations for these condi-
could include capillary rise after the elimination of tions must consider the design life of the structure,
evapo-transpiration from the surface, infiltration due the depth of wetting that can occur during the de-
to irrigation or precipitation, or introduction of wa- sign life, and the degree of saturation, and thus the
ter from off-site. Underground sources may include portion of potential heave that can develop during
broken water lines, development of perched water the design life.
tables, or flow through permeable strata that are re- The depth of wetting and corresponding degree
charged at distant locations. of saturation can be calculated using readily avail-
Depth of Potential Heave, Zp, is the depth to able software, such as Vadose/W, SVFlux, or Hy-
which the overburden vertical stress equals or ex- drus 2-D. Using the results of these analyses, the
ceeds the swelling pressure of the soil. This repre- amount of heave that is expected to occur in the par-
sents the maximum depth of the active zone that tially wetted zone can be calculated. This amount
could occur. of heave will be less than that calculated assuming
Design Active Zone, ZAD, is the active zone for full wetting for the entire depth of potential heave.
which the foundation is designed. This is the zone Example results of the analysis of wetting due to
of soil that is expected to have become wetted dur- irrigation and precipitation are shown in Figure 4.
ing the design life. It may be equal to the depth of Analyses were conducted using the computer pro-
potential heave, or some shallower depth if water gram Vadose/W (GEO-SLOPE, 2007). Vadose/W
migration analyses indicate that the entire depth of can model both saturated and unsaturated flow in
potential heave will not become wetted. response to climatic and environmental conditions,
making it possible to analyze water migration as a
2.4.1 Saturated water content profile function of time. The program considers the effect
Construction of buildings and pavements in arid re- of precipitation and infiltration, surface seepage, ru-
gions typically results in a reduction of evaporation. noff and ponding, plant transpiration, potential and
Additionally, the introduction of irrigation typically actual evaporation, snow accumulation and melt,
exceeds the evapo-transpiration of the vegetation. ground freezing and thawing, and groundwater re-
These factors as well as others result in the devel- charge.
opment of a wetting front that progresses downward A one-dimensional profile that consisted of ho-
in the soil. Below the wetting front the water con- mogeneous claystone was used for this example
tent is the same as that which existed prior to intro- Actual site data were used in the analysis. These in-
duction of the water source. However, above the cluded initial water contents prior to development,
wetting front the water contents are higher, and the post-development irrigation values, and climate
soil may be saturated or unsaturated. The difference data. Climate data were generated using the Clim-
in soil suction between the wetter and drier zones Gen program (Washington State University, 2002)
will result in downward flow of water, and the wet- based on 20 years of historical data from a nearby
ting front will continue to move downward until an NOAA certified weather station. The soils proper-
impermeable boundary or a water table is reached ties used in the analyses are presented in Table 1.
(McWhorter and Nelson, 1979). Once an imperme- The Vadose/W model was used to model a pe-
able boundary is reached, the water table will propa- riod of 100 years. The results of the analyses are
gate upward to the surface, thusly, forming a shown on Figure 4. The volumetric water content
perched water table. Full wetting of the soil profile of the claystone is predicted to increase up to about
would be expected to occur if the soil above the 32 percent near the ground surface. Wetting is pre-
wetting front is saturated and the wetting front ad- dicted to occur only to a depth of approximately
vances to below the depth of potential heave. 11 meters in 100 years.
Where a rising groundwater table is anticipated, the
full wetting conditions should be used to make cal- Table 1. Soil properties for claystone
Saturated hydraulic conductivity, Kv
culations (Houston et al., 2001). (cm/sec)
2.0 × 10-8
If full wetting is not expected to occur, analyses
Porosity, n 0.37
must be conducted to determine the water content
profile at the end of the design life. If such analyses Displacement pressure, hd (m) 30
are not conducted it should be assumed that full Pore size distribution index, λ 0.218

62
2.4.3 Simplified hand calculations 
 k ( f )  Z p   2  3
McWhorter and Nelson (1979) developed a method  f  n   (11)
 
of analysis to model the movement of the wetting  k s  Z p  hd 
front. Below the wetting front the pressure head in
the soil is dictated by the soil suction, which in turn, If k(θf) ≤ ks, Equation (11) can be further simpli-
is a function of the water content. They derived the fied as Equation (12).
following generalized relationship between 10 15
Volumetric Water Content (%)
20 25 30 35 40
the 0
hydraulic conductivity and volumetric water con-
tent.
5
( 2 3  )
  r  

Depth Below Ground Surface (m)

ke  k s  f  (7) 1/2000

 n  r
10
 8/2010

where: ke = effective hydraulic conductivity, 15 8/2020

ks = saturated hydraulic conductivity,
θf = final volumetric water content, 8/2040

θr = residual volumetric water constant, 20

8/2060
n = porosity, and
λ = pore size distribution index. 8/2080
25
8/2099
According to Darcy’s law, for the case where wa-
ter is continuously ponded to a small depth at the 30
ground surface, the infiltration at the surface, qi, in a Figure 4. Water content profiles for claystone calculated by Va-
homogeneous and isotropic medium is represented dose/W.
by Equation (8):


 h  Zp  23
qi  ke 1   (8)  f  n  (12)
 z Z h 
 p d 
in which h = pressure head and z = elevation head, Calculations were performed for the homogene-
and other symbols are as previously defined. ous claystone profile that was discussed in Sec-
McWhorter and Nelson (1979) showed that a tion 2.4.2. Using the soil properties presented in
good approximation for the pressure head, h, at the Table 1 and a value of Zp of 19.1 meters presented
wetting front is to assume that the suction is equal in Table 2, the results of these analyses indicated
to the displacement pressure, hd, of the soil. Assum- that a final constant volumetric water content of
ing the pressure head due to the ponding at the sur- 34% can be used for design. This result is com-
face is negligible, and that the wetting front has pared with the Vadose/W results in Figure 5. It is
progressed to the depth of potential heave. Equa- seen that this is a conservative estimate for the wet-
tion (8) can be expressed as follows: ting profile. However, it is less conservative than
 assuming fully saturated conditions.
h 
q i  k e 1  d  (9)
 Z 
 p 
2.5 Rate of heave
The infiltration, qi, does not completely contrib- The free-field heave of the soil for the saturated fi-
ute to downward migration of the wetting front. nal water content, as discussed in Section 2.4.1, was
Some of it is retained in the pore spaces of the soil calculated using Equation (5). The calculations re-
and increases the water content. This increase in sult in 384 mm of predicted free-field heave. The
water content is, of course, the factor that causes soil parameters used for the analyses are presented
heave. in Table 2.
Combining Equations (7) and (9) yields, The final water content profile calculated by the
 
 Vadose/W program and that calculated using Equa-
 q  Zp   2  3  tion (12) as presented in Figure 5 were also used to
 f   r  n   r  i    (10) calculate the free-field heave.
Z h 
  k s  p d    For soils that are not fully wetted, the percent swell
  and the swelling pressure will be less then that
Equation (10) can be simplified for unsaturated measured after saturation in the consolidation–swell
clay soils as follows: test. Therefore, in calculating heave those values

63
must be corrected for the actual degree of satura- soil, however, there would be a greater difference
tion. Normalized degree of saturation versus the for the hand calculation.
normalized percent swell for initial degree of satura-
tion of 42, 52, 66, and 84 percent, are shown in Fig-
ure 6 (Chao, 2007). These curves can be used to 3 PIER HEAVE PREDICTION
calculate the percent swell for a non-fully wetted
soil. Figure 7 shows a pier and grade beam foundation
system in which the basement wall serves as the
Table 2. Claystone Parameters Used for Heave Prediction grade beam. A void space must be maintained be-
Depth of tween the grade beam so that as the soil heaves
Total Percent Swelling Potential
Heave it
Unit Swell, Pressure,
Index, CH Heave, Zp
Weight S% σ′cs Volumetric Water Content (%)
10 15 20 25 30 35 40
Mg/m 3
% kPa - m 0

1.86 4.2 608 0.046 19.1 A B

Table 3. Normalized Final Degree of Saturation from 5

Vadose/W Model

Depth Below Ground Surface (m)

C D
Depth Normalized Final
(meters) Degree of Saturation
0 -2 0.94
2-4 0.94 10

4-6 0.92
6-8 0.90
8 - 10 0.85
10 - 12 0.24 15

Table 4. Calculated values of free field heave

zp = 19.1 m
Free-Field Heave
Water Content Profile
(mm) 20

Saturated Profile 384 A: Initial Water Content Profile

B: Saturated Water Content Prof ile
Vadose/W 287 C: Water Content Profile f rom VADOSE/W Model
Simple Hand Calculation 362 D: Water Content Profile f rom Hand Calculation

Figure 5. Comparison of design water content profiles.

The swell pressure can be corrected in a similar
fashion. Reichler (1997) showed that an e-log p 1.0
curve from a partially saturated consolidation-swell
Initial DOS = 42%
0.9 Initial DOS = 51%
test has the same shape as that from a fully wetted 0.8 Initial DOS = 66%

consolidation-swell test.
Normalized Percent Swell, S%N

Initial DOS = 84%

0.7
Using the revised percent swell as calculated in 0.6
the above discussion, the consolidation-swell line 0.5
can be shifted downward parallel to the fully wetted 0.4
line and the corrected value of swell pressure for the 0.3
partially wetted soil can be calculated. The values 0.2
of normalized degree of saturation for the Va- 0.1
dose/W water content profile are presented in Ta- 0.0
ble 3. The calculated free-field heave using this 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0

methodology was 287 mm for the water content Degree of Saturation, S

profile in 100 years obtained from the Vadose/W Figure 6. Normalized percent swell vs. degree of saturation
program. (modified from Chao, 2007).
Using the final design volumetric water content
of 33 percent that was determined using Equa- will not exert uplift forces on the grade beam and
tion (11), and the method presented above for calcu- piers.
lation of heave for partially saturated soils, the cal- Piers can be of different types. They may be
culated free-field heave was 362 mm. straight shaft drilled piers, they may be drilled piers
The calculated values of free-field heave for the with belled ends, they may be steel piles that are
three scenarios of wetting are compared in Table 4. driven or pushed, they may be helical piers, or they
The Vadose/W value is 25 percent less than that for may be micropiles. The principle on which the
the fully wetted condition whereas the hand calcula- piers are designed is to found them in a sound stra-
tion is only 6 percent less. For a more permeable tum at sufficient depth so as to provide sufficient

64
anchorage to minimize movement under the uplift d = diameter of the pier.
forces exerted by the expansive soil.
If a stable non-expansive stratum exists sufficiently The maximum tensile force generated in the pier,
near the surface, the pier may be designed as a rigid Pmax, will occur at the bottom of the depth of poten-
member anchored in that stratum so as to prevent tial heave and will be equal to
movement. With rigid pier design, it is assumed
that the pier does not move. Pmax  Pdl  f u Z AD πd (15)
If the depth of potential heave is large, the design where, fu = the uplift skin friction = α1×σ′cv and all
length of a pier designed as a rigid pier will be too other terms were defined in Equation (14).
long to be practical. Typically some movement of A more complete treatment of rigid pier design is
the pier would be tolerable. Economic design of the presented in Nelson and Miller (1992).
pier would be to calculate the required length of Rigid pier design works well if the stratum of ex-
pier pansive soil is not thick and is underlain by a stable

Figure 7. Typical pier and grade beam foundation system.

such that the tolerable movement is not exceeded.

The pier should then be designed as an elastic
member in an elastic medium. The methods for de-
sign of a rigid pier and for calculating heave of an
elastic pier are presented below. Figure 8. Forces acting on a rigid pier.
3.1 Rigid pier analysis non-expansive stratum. However, in a deep deposit
The forces acting on a rigid pier in expansive soil of expansive soil, the required pier length ap-
are shown in Figure 8. The principle of the design proaches a value equal to twice the depth of poten-
is that the negative skin friction below the depth of tial heave. In such cases the design rigid pier length
potential heave plus the dead load, P, must resist the is generally not practical for a light structure, and it
uplift pressures produced by the swelling pressures would be appropriate to use a shorter pier designed
exerted on the pier above that point. according to elastic pier design.
The equation for required length of a rigid
straight shaft pier as presented in Nelson and Mil-
3.2 Elastic pier analysis
ler (1992) is,
Elastic pier design is presented in Nelson and Mil-
1  Pdl  ler (1992). The heave prediction method for calcu-
L  Z AD 
fs α1σcv Z AD  πd  (14)
lating pier heave, as presented therein, is based on a
 
finite element solution for pier heave developed by
where: ZAD = the depth of the design active zone, Poulos and Davis (1980). The material presented
fs = the negative skin friction below the by Poulos and Davis was modified by Nelson and
depth of potential heave, Miller (1992) to make it more easily usable by the
α1 = a coefficient of adhesion between the design engineer. The material below is modified
pier and the soil, further from Nelson and Miller (1992).
σ′cv = the swelling pressure from the constant The uplift skin friction along the side of the pier
volume pressure;, may be considered to be uniform along the length of
Pdl = dead load on the pier, and the pier or to increase with depth as shown in Fig-

65
ure 9. A uniform distribution would be assumed for For the case where the skin friction is uniform
the situation where the soil within the depth of the with depth, the force, PFS, is equal to,
design active zone has generally the same swelling
PFS  f um Lπd (17)
pressure throughout. A linearly increasing distribu-
tion would be appropriate for several strata of soils
where the deeper soils exhibit a higher expansion
potential. 3.3 APEX method
Figure 9 shows normalized pier heave for a The elastic pier method is somewhat limited in that
straight shaft pier plotted as a function of the pier it considers only one soil throughout the depth of
length normalized against the depth of the design the pier, and it was developed for L/d ratios of about
active zone for a straight shaft pier. Similar curves 20 or less. In many cases the various strata of ex-
are presented in Figure 10 for belled piers having a pansive soil that are penetrated by the pier exhibit
bell diameter twice that of the shaft. widely varying properties. Also, with increasing use
of micropiles in expansive soil applications, L/d
ratios well

Figure 11. Normalized force in straight shaft piers vs. L/ZAD.

Figure 9. Normalized straight shaft pier heave vs. L/ZAD.

Figure 12. Normalized force in belled piers vs L/ZAD.

in excess of 20 are common. In order to provide for

a more versatile method of analysis a finite element
based model was developed The code for that
analysis is termed APEX for Analysis of Piers in
Figure 10. Normalized belled pier heave vs. L/ZAD. EXpansive soils. The development of that code is
presented below. A detailed description of the code
Figures 11 and 12 show the normalized maxi- is presented in Nelson et al. (2010). Design charts
mum tensile force in straight shaft and belled piers developed using that code are demonstrated and an
plotted as a function of the pier length normalized example is presented to illustrate the use of this
against the depth of the design active zone The method to analyze and design piers.
maximum force is normalized to a force, PFS, com-
puted by applying the uplift friction over the entire 3.3.1 The field equations with soil swelling
length of the pier. Material properties can be specified point-wise
For the case where the skin friction varies line- throughout the analysis domain. The swell is as-
arly with depth, the force, PFS, is equal to, sumed to be isotropic, and is accounted for by writ-
PFS  0.5f um Lπd (16) ing the constitutive equations as:

66
 1 shear to its smallest allowable value. That value
ε rr  σ rr  vσ θθ  σ zz   ε iso (18) would be the smaller of either that which will cause
E
slip, or that which will cause soil failure.
Limiting values for shear as defined by Coulomb
1
ε θθ  σ θθ  vσ zz  σ rr   ε iso (19) friction (slip) and Mohr-Coulomb failure (soil fail-
E ure) are in terms of normal and tangential compo-
nents of stress. For computational efficiency, these
1 relationships are converted to equivalent nodal
ε zz  σ zz  vσ rr  σ θθ   ε iso (20) forces with no loss in the generality of the theory.
E
For the Mohr-Coulomb failure theory, it is also nec-
where: iso = isotropic swelling strain, essary to determine soil strain, which poses a
rr, , zz = components of stress and strain unique problem which will be discussed as a sepa-
in cylindrical coordinates, and rate issue.
E = modulus of elasticity of the soil.

Figure 13. Boundary conditions: (a) soil boundary

conditions, (b) pier-soil boundary condition.

3.3.2 The boundary conditions Figure 14. Schematic of pier and soil interface: (a) initial-no
force on pier, (b) soil heave-upward force on pier, (c) pier
The pier-soil interface must be modeled such that heave-resultant force on pier is zero.
either slip between the soil and the pier (Coulomb
friction) or failure within the soil adjacent to the pier 3.3.3 Adjustment in pier heave
(Mohr-Coulomb failure) can take place. Axial strain
in the pier is assumed to be negligible relative to The pier must be in equilibrium, and therefore, the
strain in the soil. tangential forces exerted on the pier by the soil must
The boundary conditions at the pier-soil interface equal the total external load on the pier. Figure 14
specify a relationship between the nodal displace- illustrates how Hp is adjusted in order to bring about
ment and nodal force, rather than the actual value of equilibrium. Figure 14a illustrates the boundary
one or the other. In the APEX code, this specifica- conditions before the soil swells. In this state there
tion is used for the vertical component of displace- is no uplift force on the pier. Figure 14b illustrates
ment at nodes where the soil would be in contact the state of the mixed boundary conditions after
with the pier. It is shown schematically in Fig- swelling takes place but before any pier heave. The
ure 13, and has the following form: shear forces create an upward force on the pier, and
the pier is no longer in equilibrium (for simplicity,
Ft = k(Hp − Ut) (21) the pier is shown as having no external load). To
where: Ft = the nodal force tangent to pier; Hp = the bring the pier into equilibrium, it must be allowed
pier heave; Ut = the nodal displacement tangent to to move up, creating both upward and downward
pier; and k = the parameter used to adjust shear forces acting on the pier. This is illustrated in Fig-
stress, considered similar to a spring constant of the ure 14c. For given k values, the relationship be-
connection between Ut and Hp. tween heave and total force is linear, thus this ad-
The pier heave (Hp) is assumed to be the same justment requires an insignificant amount of
for all nodes because of the pier’s rigidity. Large computer time.
values of k require Ut to be approximately equal to
Hp. In these cases, the soil has the same displace- 3.3.4 Soil failure and shear strain
ment as the pier. This no-slip condition is used as If the stress in the soil adjacent to the pier exceeds
the initial condition for all such nodes. However, that which is necessary to cause soil failure by the
when the force at any node exceeds that necessary Mohr-Coulomb theory, and there has been no slip
to cause either slip between soil and pier, or soil of the soil based on Coulomb friction, the soil will
failure adjacent to the pier, k is reduced to bring the fail in shear and the shear stress on the side of the

67
pier will equal the shear strength of the soil. Loss in Typical APEX output is shown in Figure 16 for a
soil strength due to strain is incorporated into APEX soil profile with uniform expansion potential with
by a linear decrease in the apparent cohesion and depth. The cumulative heave profile shown in Fig-
angle of internal friction from their peak values to ure 16a is the input for the APEX program. Fig-
residual values at a specified shear strain. ure 16b shows the distribution of slip along the pier.
Coulomb friction is defined as a linear relation- In this case slip occurred over the entire length.
ship between normal force and skin friction. The This is reflective of the fact that the shear strength
friction is defined by an adhesion factor, . Ben- of the soil was sufficiently high that failure occurred
venga (2005) showed that the value of  can vary as slipping between the pier and the soil at the pier-
widely with most values between about 0.4 to 0.8. soil interface. For a soil with a lower shear strength,
Shear strength is defined by the conventional shear failure would have occurred in the soil at some dis-
strength equation. Figure 15 shows the two rela- tance into the soil away from the interface.
tionships. Figure 16c shows the distribution of shear stress
along the pier. In the upper portion of the pier the
3.3.5 Slip or soil failure shear stresses are positive. This defines the uplift
zone and the lower portion defines the anchorage
It is not always predictable whether the soil will slip
zone. Figure 16d shows the axial force in the pier.
along the pier or failure will occur in the soil. Fig-
The maximum value occurs at the change from the
ure 15 illustrates the allowable shear stress as a
uplift to the anchorage zone. This is the tensile
function of the normal stress for both mechanisms.
force in the pier for which the reinforcing steel must
Whether one mechanism or the other governs, de-
be designed.
pends on the normal stress. Thus, it is necessary to
monitor both mechanisms throughout the process of
iteration.

3.3.6 The iterative process

An iterative solution procedure is used. Simultane-
ous adjustments are made of the pier heave and the
k parameters during each iteration.

3.3.7 Typical APEX results Figure 15. Strength envelopes for slip and soil failure modes.

0 0

2
2

4
4
Depth (m)

Depth (m)

6
8

8
10

12 10
0 25 50 75 100 125 150 -50 -25 0 25 50 75 100

Cumulative Heave (mm) Slip (mm)

(a) (b)
0 0

Uplift
2 2
Zone

4 4
Depth (m)

Depth (m)

6 6
Anchorage
Zone
8 8

10 10
-75 -50 -25 0 25 50 75 100 0 50 100 150 200 250
Shear Stress (kPa) Axial Tensil Force (kN)
(c) (d)

Figure 16. Typical output from APEX Program: (a) cumulative heave used as input, (b) variation of slip along pier, (c) shear stress
distribution along pier, (d) axial force distribution.

68
 heave of a 750 mm diameter pier with a dead load
of

(a) (b)

Figure 17. Soil heave profiles for validation cases: (a) Manufac-
turing Building (b) CSU Test Site.

Figure 18. Elevation survey data in hyperbolic form compared

4 VALIDATION OF THE FINITE ELEMENT with heave computed by APEX for Manufacturing Building.
BASED ANALYSIS
435 kN ranges from 345 to 380 mm for values of
The APEX code was validated by comparing com- E equal to 5,000 kPa and 10,000 kPa, respectively.
puted results with field-measured data from project At the time of our investigation the total ultimate
sites where measurements of pier heave or tensile pier heave had not yet been realized. The survey
force had been recorded and where reliable soil pro- data was extended into the future to predict the ul-
file characteristics were available. Two of the vali- timate pier heave by means of a hyperbolic function.
dation cases are discussed below. These include a t
manufacturing building in Colorado, USA, and the p  (24)
a  bt
Colorado State University Expansive Soil Test Site.
The total free-field heave and the distribution of the This technique has been used previously for analy-
soil heave with depth was determined using the me- sis of pier movement in expansive soil (Chao, 2007;
thods described previously. Details of these case Overton et al., 2007). Figure 18 shows the eleva-
histories are presented in Nelson et al. (2010). Fig- tion survey data and the hyperbolic function for two
ure 17 shows the calculated soil heave distribution groups of piers that experienced large amounts of
for each case. heave.
The range of values of pier heave calculated using
the APEX code is also shown on Figure 18. It is
4.1 Manufacturing Building seen that very good agreement exists between the
The Manufacturing Building is a heavily-loaded computed value of heave and that determined using
concrete and masonry building located on an ex- measured values.
pansive soil site in Colorado, USA. Structural dis-
tress was experienced several years after construc- 4.2 CSU Expansive Soil Test Site
tion (Attwooll et al., 2006; Overton et al., 2007). The Colorado State University Expansive Soils Test
The foundation consists of drilled concrete piers Site is located in an area of the Pierre Shale forma-
having a diameter of 750 mm installed to a depth of tion. The soil consists of approximately 0.5 m of
7.0 m. The soil profile at the site consists of ap- organic silt and 0.9-1.2 m of clay underlain by clay-
proximately 3.8 to 4.3 m of silty/sandy clay under- stone.
lain by claystone of the Pierre Shale formation. The Four 350 mm diameter piers were installed at the
piers were constructed below the basement level site to a depth of 7.6 m in August 1995 (Cha-
with the top of the piers being about 6 m below the pel, 1998). Strain gauges were mounted in the cen-
top of original grade. Thus, the piers are located en- ter of each pier at depths of 1.8, 3.1, 4.3, and 5.5 m
tirely within the Pierre Shale. below the ground surface. Irrigation water was
The free-field heave profile was calculated for a placed on the site during summer and fall months
uniform soil as shown on Figure 17a. Using the during the period from September 1995 through Oc-
APEX code and that heave profile, the ultimate tober 1997. The free-field heave adjacent to this
example pier was measured to be 64 mm, and the

69
pier heave was 9 mm during the time period from ry structures, piers with diameters of 250 to 300 mm
August 1995 to August 1997. are used. Micropile diameters are about 100 to
110 mm. Pier lengths of 6 m for the larger diameter
piers were common several years ago but they have
frequently experienced intolerable heave. Pier
lengths in the range of 8 to 12 m are currently com-
mon. Micropiles are being used in some instances.
The authors are aware of some micropiles as long as
30 m with a diameter of 100 mm being installed.
More common micropile lengths are around 15 to
20 m. Thus, common values of L/d range from
about 10 to 150. The design charts that are pre-
sented below consider the range from 20 to 80
which represents the middle of the range of actual
pier sizes.
The design curves presented below were devel-
oped for a zero dead load condition.
The curves presented below were developed us-
ing a value for the adhesion factor, α, of 0.4. The
sensitivity of the results to the α parameter will be
discussed at a later point.
Figure 19. Measured versus predicted axial force in the concrete Figure 20 shows pier heave normalized with re-
pier for the CSU Test Site.
spect to free-field heave as a function of pier length
In this case, the APEX model was validated by normalized with respect to the depth of the Design
comparing the measured force in the pier to the Active Zone, ZAD. Figure 21 is an enlargement of
force calculated using APEX. Readings taken of the that portion of Figure 20 for values if L/ZAD greater
strain gauges in October 1997 were used to calcu- than 1.0. This figure facilitates use of the design
late the tensile forces in the pier. The total force in charts at a larger scale. These figures were prepared
the pier was calculated using the values of strain for the case where the cumulative free-field heave
along with the cross-sectional area and modulus of distribution is linear with depth. This would be the
elasticity of the concrete and the steel in the pier. case where the heave for each incremental soil
The measured and computed forces in the pier layer, zi, is the same. The stiffness of the soil is ex-
are shown in Figure 19. There is generally good pressed in terms of atmospheres. Thus, EA=Es/1
agreement between the measured and calculated atm where Es is the modulus of elasticity of the soil,
values for the upper three strain gauges. and atmospheric pressure is expressed in the same
The high strain gauge reading at a depth of 5.5 m units as Es.
indicates that the concrete at this location was at Figure 22 shows similar curves for the case
least partially cracked. For a partially cracked where the expansion potential of the soil is constant
and/or yielded section, the force at that depth would with depth. In that case the cumulative heave pro-
be somewhere between the two values indicated by file is non-linear as shown in Figure 22.
the solid and open points. In Figure 20 it is seen that for L/ZAD less than
Good agreement is seen between the measured about 1.5, pier heave is more for the stiffer soil;
and computed values of force at the upper three whereas the reverse is true for larger values of
strain gauges. The curve falls appropriately be- L/ZAD. Figure 23 shows the effect of soil stiffness
tween the upper and lower bounds of the measured on pier heave. The effect of stiffness is greatest at
force at a depth of 5.5 m. L/ZAD of 1.3 to 1.5 where the maximum difference
is a factor of 1.5 for L/d = 20 and 2 for L/ZAD = 80.
At larger values of L/ZAD the stiffer soil exhibits less
5 PIER DESIGN CURVES pier heave than the less stiff soil. This effect of
stiffness can be explained by the fact that for the
Curves of the nature shown above in Figures 9 shorter pier the uplift is producing more heave and
through 12 are being developed using the APEX the stiffer soil in the uplift zone has more influence.
program for use in facilitating pier design. Concrete For the larger piers, the stiffer soil in the anchorage
piers used in expansive soil applications generally zone has more influence which then resists the up-
fall within a fairly narrow range of sizes and condi- lift forces.
tions. The largest diameter drilled concrete piers in The effect of stiffness for the nonlinear heave
expansive soils that the authors have seen are those distribution is also plotted on Figure 23 and it is
for the manufacturing building discussed above seen that the effect of soil stiffness is very pro-
(750 mm). Most commonly for light one or two sto- nounced around values of L/ZAD equal to 1.0. De-

70
pending on slenderness ratio, the heave varies by a ted and the stiffness is lower. Thus, use of the oe-
factor of 2 to 2.7. It is of interest that in Figure 23 dometer test results may be reasonable.
the curves for larger values of L/ZAD converge and Figure 24 demonstrates the sensitivity of the pier
are less sensi- heave to the value of the adhesion factor, α. In Fig-
1.0 ure 24a the soil was assumed to have a shear
0.9 0.30

0.8
0.25
0.7

0.6
0.20
ρp/ρo

0.5 ZAD
ZAD

ρp/ρo
0.4 0.15

0.3
0.10
L/d = 20
0.2
20
L/d = 20 20
0.1
0.05
α = 0.4 80
0.0 80 80
α = 0.4 80
0.00 0.50 1.00 1.50 2.00 2.50 3.00 3.50
L/ZAD 0.00
1.00 1.50 2.00 2.50 3.00 3.50
Figure 20. Pier heave as a function of pier length - linear free- L/ZAD
field heave distribution. Figure 21. Pier heave for long piers - linear free-field heave dis-
tribution.
1.0
3.0
0.9

0.8 L/d = 80
2.5
Nonlinear Heave
0.7 Distribution

0.6 2.0
ρp/ρo

0.5 Linear Heave

Distribution
ρp(200)/ρp(50)

1.5
0.4

0.3
200 1.0
100
EA = 50
0.2 L/d = 20

0.1 L/d = 20 0.5

L/d = 80
0.0
0.00 0.25 0.50 0.75 1.00 1.25 1.50 1.75 2.00 2.25
0.0
L/ZAD 0.0 0.5 1.0 1.5 2.0 2.5
Figure 22. Pier heave as a function of pier Length - nonlinear L/Z AD
free-field heave distribution. Figure 23. Effect of Soil Stiffness on Pier Heave.

tive to the slenderness ratio. strength such that shear failure would occur in the
Thus, the stiffness of the soil can have a signifi- soil for higher values of α. In Figure 24b the soil
cant effect on pier heave. This shows the impor- shear strength has been assigned a high value such
tance of measuring the modulus of elasticity for the that soil shear failure does not occur for any values
soil. The modulus of elasticity can be determined of α. In that case the value of α has little or no effect
from the slope of the oedometer test results if the for piers shorter than 1.25xZA. For longer piers the
soil is assumed to be linearly elastic. However, that higher values of α produce greater heave. This im-
would represent the stiffness of a saturated soil. If plies that the uplift skin friction is producing the
the foundation soil is not saturated it will be stiffer. greater effect for the longer piers. This is also the
For wetting from the surface, ZA is smaller initially case for the soil with the lower shear strength as
and increases with time. Thus, L/ZA is large imme- shown in Figure 24a. In Figure 24a, however, at
diately after construction and decreases with time. larger values of α, failure occurs in the soil near the
By the time that the depth of wetting reaches the pier-soil interface. The uplift shear stress on the
value of ZAD the soil above the wetting front is wet- pier is, therefore, lower and the pier heave is lower.

71
6 EXAMPLE FOUNDATION DESIGN straight shaft concrete piers with a diameter of
300 mm, and a tolerable heave of 25 mm. For de-
To demonstrate the use of the design charts that sign purposes, it will be assumed that at the end of
have been presented above, an example calcula- the design life full wetting will occur throughout the
tion was entire depth of potential heave. A value of 0.4 for α
1.0
Table 5. Soil properties used in heave calculations for example
0.9 foundation design
0.8
Consolidation-Swell
Water Total Test(1)
Soil Type Es
0.7 Content Density Percent Swelling
EA = 100
Swell Pressure
L/d = 80
0.6 (%) (Mg/m3) (kPa) (%) (kPa)
Z AD
Weathered
ρ p/ρo

0.5
α = 0.1 - 0.6 Claystone 12 1.9 9,400 2.0 350
a = 0.7 (0–5 m)
0.4
a = 0.8
Soil Slip up to α = 0.6 Claystone
9 1.8 11,200 3.5 550
0.3 (5–10 m)
Soil Failure at α >0.6 Sandy
0.2
α = 0.6 - 0.8 Claystone 8 1.8 120,000 1.86 305
0.1
0.4 (>10m)
0.25 0.1
(1) Inundation Pressure = 48 kPa
0.0
0.00 0.25 0.50 0.75 1.00 1.25 1.50 1.75 2.00 2.25
L/ZAD Cumulative Heave (mm)
(a) 0 50 100 150 200 250
1.0 0

0.9
1
0.8
2
0.7
EA = 100
3
L/d = 80
0.6
ZAD
ρp/ρo

4
0.5
α = 0.1 - 0.8
5
Depth (m)

0.4

0.3 6
No soil shear
failure
0.2
α = 0.8 7
0.6 0.4
0.1 0.25 0.1
8
0.0
0.00 0.25 0.50 0.75 1.00 1.25 1.50 1.75 2.00 2.25 9
L/ZAD
(b)
10
Figure 24. Sensitivity of pier heave to : (a) lower soil shear
strength allows shear failure, (b) high soils shear strength forces
slip. 11
Figure 25. Cumulative heave profile for example calculation.
done for a pier and grade beam foundation on a
typical soil profile. The soil profile and soil proper- was assumed. The required length of pier will be
ties are shown in Table 5. This profile is similar to computed using rigid pier, elastic pier and APEX
an actual site near Denver, Colorado. The profile at analyses.
the site consists of 5 m of weathered claystone over
unweathered claystone. At a depth of 10 m the
6.1 Rigid pier analysis
claystone became sandy and much less expansive.
The depth of potential heave was computed to be The required length of a rigid pier is calculated by
10.1 m. The free-field heave was computed using equating the uplift forces as shown in Figure 8 to
Equation (5) and was calculated to be 192 mm. The the anchorage skin friction forces. The uplift skin
cumulative heave profile is shown in Figure 25. friction is equal to
For the example calculation, the top of the pier fu = α1 σ′cv (25)
was assumed to be at the ground surface and dead
load was neglected. The piers were specified to be

72
 Where α1 is the adhesion factor in the uplift zone. timated and revised if necessary after the required
As noted, a value of 0.4 was assumed. length is computed. The depth of potential heave
The negative (anchorage) skin friction can be Zp was computed to be 10.1 m and this was taken to
calculated by the equation be the design active zone ZAD. If we assume that
the pier will be approximately 50% larger than that,
fs = αsσ′h (26)
L would be 15 m. Thus,
Where αs is the adhesion factor in the anchorage 0.7

zone, and σ′h is the lateral stress acting on the pier.

The lateral pressure will be taken as being equal to 0.6
the swelling pressure of the sandy claystone.
The value of αs should be similar to that of α1. 0.5
Therefore, a value of 0.4 was used. The required Sleeved Pier
pier length for a rigid pier was calculated to be 0.4
18.7 m. Unsleeved Pier

ρp/ρo
0.3
6.2 Elastic Pier Analysis
Figure 9 was used to compute the required length of 0.2
(ρp/ρo)allowable = 0.13
an elastic straight shaft pier. Because the claystone
has a high swelling pressure the pier-soil interaction 0.1

was considered to be constant with depth. This cor- LReq'd = 11.4 m L Req'd = 15.3 m
responds to case A in Figure 9. The normalized al- 0.0
lowable pier heave is 4 6 8 10 12 14 16 18 20
Pier Length (m)
ρp / ρ = 25.4 / 192 = 0.13 (27) Figure 26. Example pier heave computed from APEX program.
Using case A in Figure 9, at p/o = 0.13
L /Zp = 1.8 (28) L assumed 15 m
  50 (31)
d 0.3 m
The depth of potential heave, Zp, was previously
calculated to be 10.1 meters. Thus, the required As shown on Figure 21, and interpolating between
pier length, Lreqd, is, the curves for L/d = 20 and 80,
Lreq’d = 1.8 × 10.1 = 18.2 meters (29) L req'd
 1.52 (32)
Z AD
6.3 APEX Analysis Thus,
The pier design charts shown in Figures 21 and 22 Lreq’d = 1.52 x 10.1 = 15.3 m (33)
will be used to determine the required pier length.
That result will be compared with the value com- The initial assumption of L/d=50 is reasonable and
puted using the APEX finite element program di- does not need to be refined.
rectly. Next it will be assumed that the curve for cumu-
The shape of the cumulative free-field heave lative heave is non-linear such as that in Figure 22.
curve shown in Figure 25 will dictate whether to At a value of p/o = 0.13 the curves for L/d = 20
use Figure 21 or Figure 22 to represent the actual and L/d = 50 coincide. Thus, from Figure 22,
condition. This example will compare the results
from both charts with the results determined directly L req'd
 1.10 (34)
from the APEX program. The stiffness of the soil Z AD
averages 10,860 kPa. Since one atmosphere is ap-
proximately equal to 100 kPa, and
10,860 kPa Lreq’d = 1.10 x 10.1 = 11.1m (35)
EA   108.6 (30)
100 kPa These results are compared with the output from
the APEX program in Figure 26. As seen from that
Therefore, the curve for EA = 100 in the design figure, the results from Figure 21 are the most accu-
charts will be used. From Equation (27), the allow- rate. It is evident from this example that the shape
able normalized pier heave is 0.13. At this value, of the cumulative heave curve can have a significant
the curves for L/d = 20 and L/d = 80 in Figure 21 effect on the results. The curve shown in Figure 25
diverge and thus, it is necessary to assume a value is nearly linear and the small amount of curvature
of L/d in order to use the charts. A value will be es- did not have a major influence.

73
 In some instances, the upper portion of the pier is REFERENCES
cased in a material such as PVC in order to reduce
the skin friction in the uplift zone. The results from Attwooll, B., Reins, J., and Lykosh, P. (2006). Heave at manu-
the APEX program from the concrete pier used in facturing facility: observations and response. Proc. of the
4th Int. Conf. on Unsaturated Soils, April 2-6, 2006, Care-
this example, but having an upper cased section 6 m free, Arizona, ASCE Geotechnical Special Publication
long with a value for α = 0.10, are shown also in 147(1), 323-346.
Figure 26. For that case the required pier length is Benvenga, M.M. (2005). Pier-soil adhesion factor for drilled
reduced to 11.4 m. Whether it is more economical shaft piers in expansive soil. M.S. Degree, Colorado State
to install the sleeved pier or just to drill the pier an- University, Fort Collins, Colorado (J.D. Nelson, advisor).
other 4 m deep is a function of the construction Chao, K.C. (2007). Design principles for foundations on expan-
sive soils. Ph.D. Degree, Colorado State University, Fort
costs. Also the value of  for soil against PVC is Collins, Colorado.
closer to 0.3. Chao, K.C., Overton, D.D., and Nelson, J.D. (2006). The ef-
fects of site conditions on the predicted time rate of heave.
American Society of Civil Engineers, Unsaturated Soils
7 CONCLUSION Conf. 2006. Special Publication 147, 2086-2097.
Chapel, T.A. (1998). Field investigation of helical and concrete
Piers in Expansive Soils. Thesis submitted in partial re-
An essential component of the ability of a foun- quirement for the M.S. Degree, Colorado State University,
dation to provide adequate support of a structure is Fort Collins, Colorado.
the expected movement of that foundation. Just as Chen, F.H. (1988). Foundations on Expansive Soils. Elsevier.
the settlement of a foundation on soft soil must be New York, NY.
calculated and analyzed, so must the heave of a GEO-SLOPE International, Ltd. (2007). GEO-STUDIO VA-
DOSE/W software package for seepage analysis, Version
foundation on expansive soils. This paper has pre- 7.15. Calgary, Alberta, Canada.
sented the existing state of the art with respect to Houston, S.L., Houston, W., Zapata, C., and Lawrence, C.
analyzing heave of deep pier foundations in expan- (2001). Geotechnical engineering practice for collapsible
sive soils. It has presented the design tools that are Soils. Geotechnical and Geological Engineering, Kluwer
readily available to the practicing engineers for pre- Academic Publications, 19:333-355.
dicting pier heave, and has presented a rigorous and McWhorter, D.B. and Nelson, J.D. (1979). Unsaturated flow
beneath tailings impoundments. J. Geotech. Eng. Div.,
useful finite element method of analysis. That ASCE, November, 105(GT11):1317-1334.
method has been termed the APEX method. Nelson, J. D. (2010). Heave Prediction in Expansive Soils.
Prior to development of the APEX method the Panel Discussion, UNSAT 2010, Barcelona, Spain.
two most commonly used methods that have been Nelson, J. D., Chao, K. C., and Overton, D. D. (2007). Defini-
published are the rigid pier method (Chen, 1988) tion of expansion potential for expansive soils. Proc. of the
and the elastic pier method (Nelson and Miller, 3rd Asian Conf. on Unsaturated Soils. Nanjing, China. 21-
23 April. 587-592.
1992). The rigid pier method assumes equilibrium Nelson, J.D. and Miller, D.J. (1992). Expansive Soils: Prob-
of the pier, and hence, no pier movement. It pro- lems and Practice in Foundation and Pavement Design,
vides for an overly conservative design and long John Wiley and Sons, New York, NY. USA.
pier lengths. The elastic pier method allows for Nelson, J.D., Reichler, D.K., and Cumbers, J.M. (2006). Pa-
some tolerable amount of pier heave. However, in rameters for heave prediction by oedometer tests. Proc. of
its currently used form it is limited to use in simpli- the 4th Int. Conf. on Unsaturated Soils. Carefree, Arizona.
USA. April. ASCE, Geotechnical Special Publication, No.
fied soil profiles and uniform piers. It has also limi- 147: 951-961.
tations with respect to the range of slenderness ra- Nelson, J.D., Thompson, E.G., Schaut, R.W., Chao, K.C.,
tios for which it applies. Overton, D.D., and Dunham-Friel, J.S. (2010) Design con-
The APEX program presented in this paper is a siderations for piers in expansive soils. ASCE J. Geotech.
more versatile and robust method of analysis, and and Geoenvironmental Engr. (Submitted for publication
Sept, 2010).
represents an improvement over the other two meth- Overton, D.D., Chao, K.C., and Nelson, J..D.. (2006). Time
ods. It addresses the limitations of those methods rate of heave prediction for expansive soils. American Soci-
and allows for pier analysis within complex soil ety of Civil Engineers, Proc. of GeoCongress 2006 Conf.
profiles where soil properties and/or water contents Atlanta. February.
vary with depth. It is also useful for piers with com- Overton, D. D., Chao, K.C., and Nelson, J.D. (2007). Heave
plex construction details such as segmented mi- distress of a manufacturing building. Proc. of the Geo-
Denver 2007 Conf., Denver. February. In Contemporary Is-
cropiles. sues in Deep Foundations (GSP 158).
When compared with the existing rigid and elas- Poulos, H.G., and Davis, E.H. (1980). Pile Foundation Analy-
tic pier analysis methods, APEX generally predicts sis and Design. John Wiley, New York. USA.
lower pier heave values when considering a stan- Reichler, D.K. (1997). Investigation of variation in swelling
dard straight shaft concrete pier. The magnitude of pressure values for an expansive soil. Thesis submitted in
the difference depends upon the shape of the soil partial requirement for the Master Degree, Colorado State
University, Fort Collins, Colorado.
heave profile. It is believed that the APEX method Washington State University. (2002). Biological Systems Engi-
is more realistic and gives more accurate results neering Department, Climgen – Climate Data Generator Us-
than the other design methods. ers Manual, Pullman, Washington.

74
View publication stats