The document summarizes research being conducted on incorporating pile setup into pile design using Load and Resistance Factor Design (LRFD). The research aims to identify conditions where pile setup may be used, determine the reliability of pile setup prediction methods, and establish resistance factors. Field data on pile setup is presented from a bridge project in Louisiana. Methods for predicting pile setup are described, including empirical equations and static capacity methods using Cone Penetration Test data. Software tools for pile capacity analysis incorporating pile setup are identified.

1. Neha Verma
Graduate Student
Civil Engineering Program
Louisiana Tech University
1

2.  Objectives of the research
 Problem definition
 Scope of work
 Methodology
 Result Analysis
 Conclusion
 Recommendation
2

3.  Identify the conditions where pile setup may be
incorporated in pile design.
 Determine the reliability associated with pile setup
prediction by LRFD implementation.
 Determine resistance factors based on pile setup.
3

4.  Load and resistance factor design (LRFD) is a
rational approach.
 Load uncertainties, incorporated by load factors.
 Resistance factors account for the uncertainties
associated with material properties.
4

5.  LRFD accounts for variability in both resistance and
load.
 Achieves more uniform levels of safety based on
strength of soil, design methods and foundation
types.
 Provides more consistent design and level of safety
in the superstructure and substructure.
 Implemented to geotechnical structures, but never
to pile-setup.
5

6.  Piles driven in soft soils undergo increase in axial
capacity with time known as “pile setup” or
“freeze”.
 Phenomenon was discovered long back in 1955 by
Reese and Seed.
6

7.  During pile installation soil
around pile displaced
outward and exposed to
large strains.
 Excessive pore pressure is
generated in soil, results in
temporary reduction
strength of soil.
 Excessive pore pressure
begins to dissipate and soil
gains strength, hence pile
capacity increases.
7
Illustration of pile setup phenomenon

8.  Numerous cases from
the history and local
field test data in
showed pile setup.
 The test data on driven
piles at LA-1 relocation
project indicates 30-
100 % growth in pile
capacity .
8
Increase of pile capacity with time based
on field data (Vesic, 1977)

9.  Expenditure on construction of pile foundation
reaches millions of dollars every year (LADOTD).
 The current design based on the 14 day pile
resistance after the initial driving.
 No incorporation of long term pile capacity
increase, due to no recommendations.
9

10.  The accuracy and efficiency of different pile setup
prediction methods can be determined.
 The resistance factors (Ф) based on pile setup can
be useful for future research and guidelines.
 Incorporation of pile setup, cost effective design of
pile foundation:
a) Reducing length of pile.
b) Varying cross-section of pile.
c) Choice in using heavy or light driving equipment.
10

11.  The load data for the production pile is from phase
1-B of LA-1 relocation project in Leeville,
Louisiana.
 Construction of a 4-mile long high-level bridge
with connecting ramps and interchanges.
 The substructure of the project comprises of 16",
24", & 30" prestressed concrete (PSC) piles.
11

12. Summary of restrike records
 A total of 115
restrike records on
95 piles.
 63 records short
term restrike less
50 hrs EOD.
 21 long-term
records of more
than two weeks.
12

13. Load testing summary for nine test piles at
LA-1 relocation project (LADOTD)
 Location of test piles
represent the soil
 The test piles were
monitored during
driving by PDA (Pile
Driving Analyzer)
 Analyzed using
CAPWAP software (Case
Pile Wave Analysis
Program) .
13

14. 14
Pile Pile Type
Restrike
Date
Time
(Hrs)
Penetration
Length (ft)
Soil Type
Rskin
(kips)
Rtip
(kips)
Rult
(kips)
NC44-07 16" SQ. PPC 10/2/2006 24 81.37 Major clay with sand 110 43 153
NC40-04 16" SQ. PPC 10/11/2006 24 81.92 Major clay with sand 208 17 225
NC36-04 16" SQ. PPC 10/21/2006 24 77.08 Major clay with sand 231 29 260
NC33-04 16" SQ. PPC 10/26/2006 24 71.44 Major clay with sand 294 17 310
NC29-02 24" SQ. PPC 1/20/2007 744 114.31 Major clay with sand 353 70 422
NC29-02 24" SQ. PPC 3/2/2007 1728 114.31 Major clay with sand 451 59 510
NC29-03 24" SQ. PPC 12/21/2006 24 114.31 Major clay with sand 213 82 294
NC29-03 24" SQ. PPC 12/27/2006 144 114.31 Major clay with sand 271 69 340
NC29-03 24" SQ. PPC 1/17/2007 672 114.31 Major clay with sand 433 72 505
NC29-03 24" SQ. PPC 3/2/2007 1728 114.31 Major clay with sand 450 70 520
: : : : : : : : :
: : : : : : : : :
: : : : : : : : :
Information for NC site (LADOTD)

15. 0
100
200
300
400
500
600
700
0 200 400 600 800 1000 1200 1400 1600 1800 2000
Rult(kips)
Time EOD (HRS)
NC-Rult vs EOD
Pile NC1B-03
Pile NC02-03
Pile NC06-02
Pile NC10-03
Pile NC18-03
Pile NC22-03
Pile NC24-03
Pile NC14-03
Pille NC25-02
Pile NC26-03
Pile NC28-03
Pile NC29-03
Pile NC33-04
Pile NC36-04
Pile NC40-04
Pile NC44-07
Pile NC48-05
Pile NC52-05
Pile NC56-05
Pile NC59-06
Pile NC60-05
Pile NC64-05
Pile NC66-06
Pile NC68-02
Pile NC72-05
Pile NC75-05
15
Total capacity variation with time from the
restrikes at North Connector site

16. 16
0
200
400
600
800
1000
1200
0 100 200 300 400 500 600 700
Rult
Time EOD (Hrs)
SC-Rult vs EOD
Pile SC02-02
Pile SC05-02
Pile SC10-02
Pile SC13-02
Pile SC17-03
Pile SC21-03
Pile SC25-02
Pile SC29-03
Pile SC33-03
Pile SC37-03
Pile SC41-03
Pile SC45-02
Pile SC49-02
Pile SC52-03
Pile SC54-03
Pile SC56-02
Pile SC59-03
Pile SC61-04
Total capacity variation with time from the
restrikes at South Connector site

17. 17
0
200
400
600
800
1000
1200
0 1000 2000 3000 4000 5000 6000
Rult(Kips)
Time EOD (Hrs)
Mainline S-Rult Vs EOD
Pile 20S-02
Pile 23S-02
Pile 27S-03
Pile 31S-03
Pile 34S-02
Pile 37S-03
Pile 40S-01,04
Pile 41S-03
Pile 53S-02
Pile 58S-03
Pile 61S-03
Pile 64S-01
Pile 65S-03
Pile 69S-03
Pile 73S-02
Pile 78S-03
Total capacity variation with time from the
restrikes at Mainline-S site

18. 0
50
100
150
200
250
300
350
400
450
500
0 500 1000 1500 2000
Rult(Kips)
Time EOD (Hrs)
Ramp N1-Rult Vs EOD
Pile N1-
02
Pile N1-
05
Pile N1-
09
Pile N1-
12
Pile N1-
14
Pile N1-
17
18
Total capacity variation with time from the
restrikes at Ramp-N1 site

19.  The LA-1 relocation project site is dominated by
clay along with silts and sand traces.
 The mudline is about 1-3 feet below the water
table.
 The liquidity indexes 20 and 40, depth of 70 feet.
 Compressive strengths from the unconsolidated
undrained tests 0.1 to 0.5 tsf (tons square feet).
19

20. 20
Typical boring log data (LADOTD)

21. 21
Typical CPT log data (LADOTD)

22.  Empirical Equations.
1. Skov-Denver method.
2. Rate-based method.
 Static pile capacity
methods.
1. LCPC method (CPT
based).
2. Schmertmann method
(CPT based).
3. de-Ruiter and Beringen
method (CPT based).
4. α-method (Total stress
Based).
22

23.  Semi-logarithmic relation between pile capacity and
time proposed by Skov and Denver (1988):
 where A (0.5-0.7 LA clayey soils) is a dimensionless
setup factor and based on soil type.
 Q and Q0 can either be the total or the shaft pile
capacity (setup) at time t and t0.
Q(ult) = Q(t) (predicted skin friction) + T(t0) (measured
tip resistance at the reference time)
23

24.  Developed from the combined restrike data of all the
production piles of the LA-1 Relocation project;
 S(t) = Skin Friction at time t;
 S(t0) = Measured Skin Friction at time (t0=24 hrs);
 t = time elapsed since the end of initial driving; and
 t0 = Reference time i.e. 24 hrs.
Q(t) = S(t) (predicted skin friction) + T(t0) (measured tip
resistance at the reference time)
24

25. 25
Setup prediction for 25 hrs-7 day by Rate-based method
REFERENCE PILE INFORMATION PREDICTION
RATE-BASED
Pile
Time
(Hrs)
Pile Name
Ref Skin
Friction
(S0)
(Tons)
Tip
Resistance
(tons) at (t =
24Hrs)
Ref Time
(t0)(Hrs)
Skin
Friction
(Tons)
Total
Capacity
(Tons)
NC66-06 120 NC68-02 70 7.5 24 100 107
NC64-05 98 NC68-02 70 7.5 24 94 101
NC60-05 43 NC68-02 70 7.5 24 77 84
NC59-06 25 NC68-02 70 7.5 24 70 78
NC52-05 27 NC44-07 55 21.5 24 56 77
NC48-05 42 NC44-07 55 21.5 24 60 81
NC29-03 144 NC29-03 106.3 40.85 24 160 200
NC28-03 148 NC29-03 106.3 40.85 24 161 202
NC26-03 46 NC29-03 106.3 40.85 24 118 159
NC25-02 46 NC29-03 106.3 40.85 24 118 159
: : : : : : : :
: : : : : : : :
: : : : : : : :

26.  Proposed by Bustamante and Gianeselli (1982), based
on pile type, soil type and pile cone tip elevation;
or
where, ks1=30-150 (pile type,soil type and installation
process);
kb1 = 0.15-0.60 (installation procedure and soil type);
qeq = equivalent average cone friction (qc);
As = pile-soil surface area; and
At = pile toe area.
26

27.  Proposed by Schmertmann (1978)
Shaft resistance in soil:
Ultimate pile toe resistance:
Ultimate pile capacity is given as:
Where, K is the ratio of unit pile shaft resistance to unit cone sleeve
friction and function of penetration depth;
b is pile width; D is pile penetration length;
qc1 is minimum of the averages of qc values from 0.7 to 4D below
the pile tip;
qc2 is average of minimum qc values 8D above the pile cone tip
27

28.  Proposed by de-Ruiter and Beringen (1979)
Ultimate pile capacity:
or
where, α is 1 for normally consolidated clay, 0.5 OC;
Su = qca/Nk ;
qca = average qc value over a specified zone method;
Nk = 15 to 20;
Nc = 9;
As = pile-soil surface area; and
At = pile toe area.
28

29.  Also known as Tomlinson method, based on undrained soil
shear strength parameters.
Ultimate capacity:
or
where, α is the empirical adhesion factor based on the
reduction of average undrained shear strength cu;
Nc is a dimensionless bearing factor, the pile diameter and
length of the pile, taken as 9 for deep foundation.
29

30. 1. Louisiana Pile Design and Cone penetration
Test (LPD-CPT).
2. DRIVEN 1.2.
30

31.  LTRC, the load capacity/analyzing driven precast
concrete piles.
 Schmertmann, de-Ruiter and Beringen method and
LCPC method.
 Input data in order of depth, tip resistance and
sleeve resistance.
 Pile data information.
 Output results: plots of CPT data, soil classification
plot and variation of bearing capacity plot.
31

32. Plot of CPT data Plot of soil classification
32

33. 33
Variation of ultimate capacity based on three CPT
methods

34. 34
Mainline-S segment based on de-Ruiter and
Beringen method.
Pile Name Pile Type
Penetration
Length (ft.)
Test Data
Used
End
Bearing
Capacity
(tons)
Pile Shaft
Capacity
(tons)
Ultimate
Bearing
Capacity
(tons)
20S-02 24" SQ. PPC 81.01 CPT-366 28 104 132
23S-02 24" SQ. PPC 83.15 CPT-55 85 155 240
27S-03 24" SQ. PPC 83.15 CPT-55 85 155 240
31S-03 24" SQ. PPC 90.41 CPT 55b 95 195 290
34S-02 24" SQ. PPC 89.84 CPT 380 60 180 240
37S-03 30" SQ. PPC 144.84 CPT-57 640 420 1060
40S-02 30" SQ. PPC 72.74 CPT-387 360 140 500
41S-03 30" SQ. PPC 75.57 CPT-387 290 150 440
45S-03 30" SQ. PPC 161.58 CPT-392 260 430 690
47S-03 30" SQ. PPC 157.58 CPT-392 260 430 690
: : : : : : :
: : : : : : :
: : : : : : :

35.  FHWA and new version of the SPILE program.
 Tomlinson (α-method), Nordlund, Thurman,
Meyerhof, Cheney and Chassie.
 Input parameters: unit weight, undrained shear
strength, SPT N value and pile infomation.
 DRIVEN 1.2 presents the output in graphical and
theoretical form.
35

36. 36
Sample soil classification plot based on DRIVEN 1.2 program

37. 37
Variation of bearing capacity based on DRIVEN 1.2 program

38. 38
Mainline-S segment based on α-
method
α- method Prediction
Pile Name Pile Size
Penetration
Length (ft.)
Boring Log
Used
Rtip (tons) Rskin (tons) Rult (tons)
20S-02 24" SQ. PPC 81.01 B-54 135.54 76.71 212.25
23S-02 24" SQ. PPC 83.15 B-55B 10.8 176.33 187.13
27S-03 24" SQ. PPC 83.15 B-55B 10.8 176.33 187.13
31S-03 24" SQ. PPC 90.41 B-56 7.02 139.215 146.235
34S-02 24" SQ. PPC 89.84 B-56 7.02 136.765 143.785
37S-03 30" SQ. PPC 144.84 B-56 25.03 421.57 446.6
40S-02 30" SQ. PPC 72.74 B-58 9 161.72 170.72
41S-03 30" SQ. PPC 75.57 B-58 9 166.25 175.25
: : : : : : :
: : : : : : :
: : : : : : :

39.  To estimate the reliability associated with the pile
setup predictions by different methods.
 Calibration of LRFD has been performed.
39

40.  Calibration is the process of assigning values to
resistance factors and load factors.
 In the present work, LRFD components are
determined based on reliability theory.
 The level I probabilistic method, Mean-Value-First-
Order-Second-Moment method (MVFOSM).
40

41.  Statistical parameters are determined using the
measured and predicted capacity predictions, which are
prerequisites for determining resistance factors.
λRi is the bias factor:
Rm is the measured resistance; and
Rn is the predicted or nominal resistance.
 Statistical uncertainties mean, standard deviation and
coefficient of the variance.
41

42. λR is average resistance bias factor:
σR is the resistance standard deviation:
COVR is the resistance coefficient of the variance:
where , N is the number of cases.
42

43. 43
Bias factor based on rate-based method for 14 days and
above prediction
Measured Capacity Predicted Capacity Bais Factor (λ)
Pile
Time
(Hrs)
Rskin (Tons)
Rult
(Tons)
Rskin
(Tons)
Rult
(Tons)
Skin
Friction
(λ)
Total
Capacity
(λ)
NC29-02 744 176.4 211.2 271.1 279.3 0.651 0.756
NC29-02 1728 225.7 255.1 196.2 237.1 1.150 1.076
NC29-03 672 216.5 252.5 196.1 236.9 1.104 1.066
NC29-03 1728 225.1 260.1 196.2 237.1 1.147 1.097
NC28-02 218 166.7 200.0 178.0 218.8 0.936 0.914
NC14-03 644 214.5 265.0 329.2 364.7 0.652 0.727
NC10-03 285 195.5 228.0 127.5 165.0 1.533 1.382
NC10-03 323 233.0 268.0 129.6 167.1 1.798 1.604
SC54-03 648 369.4 475.0 318.1 413.1 1.161 1.150
: : : : : : : :
: : : : : : : :
: : : : : : : :

44.  To determine RF, latest AASHTO LRFD Specifications (Paikowsky
et al. 2004) are adopted for load statics and load factor to make the
pile foundation design consistent with the bridge super structure
design.
 In the present work, the reliability analysis is performed for a factor
of safety equal to 2.5.
 As per specified in the AASHTO LRFD Specifications the live load
factor (γL) and dead load factor (γD) are taken as 1.75 and 1.25,
respectively.
 Four reliability indices are selected 2, 2.5, 3, and 2.33; which are
corresponding to four different dead load to live load ratios i.e.,
QD/QL = 1, 2, 3 and 4.
44

45. Where, COVQD coefficients of variation for the dead load = 0.1;
COVQL coefficients of variation for the live load = 0.2;
λQD is the bias factor for the dead load = 1.15;
λQL is the bias factor for the live load = 1.05; and
COVR is the coefficients of variation of resistance (R).
45

46.  The axial design capacity of pile may be presented
as:
 In the future, pile foundation can be designed by
using the resistance factors, ϕ, based on pile setup,
and nominal resistance (Rn) of the pile.
46
𝑃𝐷𝑒𝑠𝑖𝑔𝑛 = 𝜙𝑅 𝑛

47.  The reliability analyses performed based on different elapsed
times.
 Skov-Denver and rate-based method, first interval: 25 hours
to 7 days (1 week) after the end of driving and second
interval: 7 days to 14 days (2 weeks) and more after the end
of driving.
 For static pile capacity methods, first interval: 48 hours after
the end of driving, second interval : after 48 hours to 7days
(1 week), and third interval: at 14 days (2 weeks) and more.
47

48.  Summary of statistical analysis.
 Summary of resistance factors.
48

49. 49
Summary of statistical analysis for ultimate capacity

50. 50
Summary of resistance factors for skin friction

51. 51
Summary of resistance factors for ultimate capacity

52.  The values of the statistical parameters were different
for different methods and different elapsed time.
 Also values of the resistance factors are different for all
the methods; therefore, different resistance factors
must be used for different elapsed times.
 No much difference was found between resistance
factors of skin friction and ultimate capacity. It implies
that tip resistance does not play significant role.
 Resistance factors based on the static pile capacity
method showed prominent variations for different time
intervals.
52

53.  A Large volume of long-term restrike or long-waiting load
testing data is required in order to improve the accuracy and
reliability of the prediction model.
 Implementation of LRFD on accurate setup predictions will
generate more reliable factors. Therefore, attention must be
paid to get the long-term data through dynamic monitoring
and static and statnamic load testing.
53

54. 54