##### Document Text Contents

Page 1

TE

C

H

N

IC

A

L

N

O

TE

The Application of PDA/CAPWAP to

Ensure Quality and Capacity in Driving

Long Steel H-piles

The methodology of application of Pile Driving Analyser (PDA) test and Case Pile Wave Analysis

Programme (CAPWAP) analysis for quality control and capacity assessment in driving long

steel H-piles is fully illustrated in the project ‘Sun Yat Sen Memorial Park and Swimming Pool

Complex’. The deficiencies of using the Hiley Formula as a field control to assess pile capacity

have been well known. These deficiencies are more prominent when the pile is long. In the use

of the Hiley Formula for long driven piles, out of range of final set table has been a common

phenomenon and this has rendered the formula not applicable. The paper discusses the use of

a hybrid method making use of advanced method (PDA/CAPWAP) in conjunction with simple

method (Hiley Formula) for field control and pile capacity assessment. It has been shown that this

method will not only improve the reliability of the pile foundation but also contribute greatly to

the overall cost-effectiveness of installation of the piles by reducing the chance of over-driving

and damage to the pile. This method also enables the successful use of hydraulic hammers in

taking final sets.

Keywords: Pile Driving Analyser (PDA), Case Pile Wave Analysis Programme (CAPWAP),

Hiley Formula, Steel H-piles, Final Set Table, Hydraulic Hammer

W W LI

Architectural Services Department,

the HKSAR Government

M K WONG

Architectural Services Department,

the HKSAR Government

Y K CHAN

Architectural Services Department,

the HKSAR Government

The Hong Kong Institution of Engineers Transactions, Vol 18, No 2, pp34-49

Paper T0819-201002; Received 8 March 2010; Accepted 13 July 2010

Introduction

For pile driving in Hong Kong, the Hiley Formula has been widely used as

a field control to assess pile capacity despite its deficiencies. All dynamic

formulae suffer from inherent problems including poor representation of

the hammer, driving system, pile and soil. These deficiencies are more

prominent when the pile is long. In the use of the Hiley Formula for long

driven piles, ie pile length that is greater than about 55 m to 60 m, out

of range of final set table has been a common phenomenon and this

has rendered the formula not applicable. A pile-driving criterion that

is applicable to long driven piles as well as user-friendly is called for.

Accurate determination of the driving criterion will not only improve the

reliability of the pile foundation but also contribute greatly to the overall

cost-effectiveness of installation of the piles by reducing the chance of

over-driving and damage to the pile.

Quality Control on Pile Driving

Fung et al. (2004b) investigated the reliability of CAPWAP analysis in

pile capacity prediction as compared with results from static load tests

in local soil conditions. It has been shown that CAPWAP analysis is a

fairly accurate method for driven pile capacity prediction.

Fung et al. (2004b) also compared pile capacity predicted by CAPWAP

analysis and by the Hiley Formula (where the coefficient of restitution (e)

and drop efficiency (E

h

) were obtained by back analysis from CAPWAP

results of trial piles) of 327 piles from 18 sites. The results revealed

that 94% of the piles have capacities predicted by the Hiley

Formula deviating from their corresponding CAPWAP capacities by less

than ± 10%.

So and Ng (2009) pointed out that among the pile capacity prediction

methods of PDA, the Hiley Formula and the HKCA Formula, PDA gave

the best estimation. It could be used in lieu of a static loading test for

verifying pile capacity when the piles are long.

In the last decade, the Architectural Services Department (ArchSD)

used Pile Driving Analyser (PDA) test and CAPWAP analysis for quality

control and capacity assessment of driven piles installation. PDA has

been used to detect the integrity of piles, hammer efficiency and give

a rough indication of pile capacity. The data of force and velocity of

piles obtained from the PDA tests can be further analysed by CAPWAP

to give a more accurate assessment of pile capacity.

Carrying out PDA test and CAPWAP analysis for each pile for acceptance

may affect the site progress of projects with restricted driving time to a

certain extent. Therefore, a simple tool to assess pile capacity during the

final set of the driving process is still in need. Despite various shortcomings

of the Hiley Formula as mentioned above, it has been shown that the

Hiley Formula can take up the role after calibration by the CAPWAP

analysis to determine the efficiency of hammer (E

h

) and the coefficient

of restitution of hammer cushion (e) of the hydraulic hammer (Fung et

al., 2004a). Strictly speaking, the ArchSD’s method is a hybrid method

making use of advanced method (CAPWAP) in conjunction with simple

method (Hiley Formula). The resulting method is therefore handy to use

but avoids the drawbacks of the Hiley Formula. The details of the pile

acceptance process are as follows:

Simple Ground Conditions – The efficiency of the hammer (E

h

) and

the coefficient of restitution of the hammer cushion (e) of the hydraulic

hammer shall be determined from/verified by CAPWAP analysis of trial

piles. The combination of E

h

and e shall be so chosen such that when

these values are substituted into the Hiley Formula, the average of the

predicted bearing capacity of the trial piles is not higher than 85% of

the average CAPWAP capacity. Once these parameters are determined,

a final set table can be established for the contract. The factor of 85%

accounts for possible deviations of CAPWAP predictions from static load

test results and for the gradual decrease in driving system efficiency due

to deterioration in the condition of the hammer cushion. Normally, the

e value of cushion decreases with sustained use and the efficiency of

the driving system becomes lower when compared with that observed

during trial pile installation (Fung et al., 2004b).

Difficult Ground Conditions – In cases where measured final sets

are out of range of the set table with E

h

and e so chosen, all the piles

falling into this category shall be subject to CAPWAP analysis. To further

ensure quality, at least the pile with the lowest CAPWAP capacity shall

be load tested for acceptance. This provision can eliminate the risk of

pile over-driving which is common for very long pile or difficult ground

conditions, and as a result causing less pile damage.

THE HKIE TRANSACTIONS • Volume 18 Number 234

Page 2

The above methods have been used by the ArchSD as a field control

to assess pile capacity since 2003. From experience obtained to date,

they provide an accurate and workable tool to fulfil the requirement

of quality control and assessment of pile capacity. It also enables the

successful use of hydraulic hammers in taking final sets.

Difficult Ground Conditions – Sun Yat Sen Memorial

Park and Swimming Pool Complex

Ground Conditions

The project site is a challenging and difficult site in difficult ground

condition with bedrock at 56 m to 94 m below ground.

The borehole information indicates that the stratification of site comprises

mainly, Fill, Marine Deposit, Alluvium and Completely Decomposed

Granite.

Fill ranges in thickness from 19 m to 27 m with many boulders, cobbles

and gravels.

Marine Deposit with thickness ranging from 2 m to 5 m lies beneath

the Fill.

Alluvium, which is encountered beneath the Marine Deposit, has a

thickness ranging from 2 m to 6 m.

Completely Decomposed Granite with thickness ranging from 21 m to

62 m lies beneath the Alluvium. Rockhead level lies between 56 m and

94 m below ground.

Difficulties of the Site

It was expected that the steel H-piles would be founded at the residual

soil with SPT N-values > 200, which was in general approximately at

-52 mPD. However, the geological profile of the site shows that the layer

of soils with SPT N-values > 200 is deeper in the central portion of the

site where longer piles were expected. The installation of driven steel

H-piles with some of them longer than 70 m in a site with many boulders

and cobbles would be the main challenges to the Contractor. Since the

piles would be long, out of range of final set table was expected if the

Hiley Formula was used. The presence of many boulders and cobbles

might also affect the progress of works significantly.

Trial Pile Installation

In this project, the Contractor adopted 305 x 305 x 223 kg/m steel

H-pile as foundation. 328 steel H-piles at an estimated average depth

of 56 m were adopted with theoretical safe pile capacity of 3,600 kN.

Eight trial piles were selected for PDA tests with CAPWAP analysis to

determine the efficiency (E

h

) and the coefficient of restitution of the

hammer cushion (e) of the 20-tonne hydraulic hammer for final set.

These piles were spread evenly across the site including the area where

the longest pile depth is expected. Before the start of pile driving, visual

inspection of the type and quality of the hammer cushion was carried

out. The ram and helmet of the hammer were weighed. The results of

correlation are shown in Appendix A.

It was observed that all trial piles were out of range of the final set

table calculated using the correlated Hiley Formula based on the E

h

and e determined by CAPWAP analysis. In view of the trial pile results,

the Contractor then chose to use the acceptance method for difficult

ground conditions, ie carrying out PDA test and CAPWAP analysis for

every pile for assessing pile capacity and load testing the pile with the

lowest CAPWAP capacity for acceptance.

It was also observed that for long piles where the bearing strata are

deep, the CAPWAP capacities are lower (some of them are even lower

than 7,200 kN) because the Contractor was reluctant to drive the piles

to a deeper level to achieve a higher capacity for fear of damaging the

piles. This, however, may be acceptable due to potential set-up effect.

PDA Tests

In order to start the pile cap works early by phases, the site area was

originally divided into four zones and further increased to six zones later

due to difficulties encountered in driving piles to achieve the design

capacity in some areas of the site. PDA tests were carried out for each

pile by a testing firm employed by the Contractor. Based on the results of

the PDA tests, the piles with lower capacity (determined by Case method)

were selected to be tested by PDA again by an independent testing firm

employed by the ArchSD. CAPWAP analyses were also carried out for

these piles and the pile with the lowest CAPWAP capacity of each zone

was load tested for acceptance. Table 1 below summarises the number

of piles selected for PDA test for each zone.

Table 1 – PDA Tests Carried out for Different Zones

Zone No No of Piles No of PDA Tests No of PDA Tests

Carried out by Carried out by

the Contractor (%) the ArchSD (%)

1 64 64 (100%) 9 (14%)

2 97 97 (100%) 16 (16%)

2A 6 6 (100%) 6 (100%)

3 77 77 (100%) 18 (23%)

4 132 132 (100%) 14 (11%)

4A 9 9 (100%) 1 (11%)

First Loading Test

After the completion of pile installation for the first zone, loading test

was carried out for the pile with the lowest CAPWAP capacity (C9A-1,

6,715 kN). The loading test was carried out on 16 December 2008,

seven days after the PDA test. The maximum test load applied was

7,200 kN. The load test results failed to satisfy the contract specification.

According to the contract specification, if a pile fails the static load

test, additional load tests need to be carried out for other two piles for

acceptance. Therefore, piles with the second and third lowest CAPWAP

capacity of the zone, ie C5A-3 (6,728 kN) and C4A-3 (7,147 kN), were

chosen for static load test. Both piles passed the static load tests. The

borehole logs and test results of the three piles are shown in Fig 1 and

(SPT of 200 blows in 250 mm penetration is denoted as 200/250)

Figure 1 – Borehole Logs

THE HKIE TRANSACTIONS • Volume 18 Number 2 35

Page 3

Set-up Effect

The time lag between the PDA test and the static load test for piles no

C5A-3 and C4A-3 are 15 days and 24 days respectively. Increase in the

pile capacity may occur after PDA test due to soil set-up.

The set-up effect of this site has been studied. PDA tests and CAPWAP

analyses were also carried out for 51 piles at different time interval after

the end of driving in order to investigate the set-up effect. Static load

tests had not been carried out on these piles. The results show that pile

capacity increases ranging from 1 – 52% between restrike tests carried

out after time lapse in the range of 3 to 187 days. It is also noted that

there was a decrease in the set per blow of the second restrike tests.

If more energy had been used to drive the piles in the second restrike

tests to produce a larger set per blow, the CAPWAP prediction would

have been even higher (ie higher set-up effect). The plot of pile capacity

variations with log time is shown in Fig 5.

Fung et al. (2006) pointed out that the timing of both the static load

test and the dynamic restrike test relative to the end of driving (set-up

Table 2 respectively. C9A-1 was re-driven to a depth of 72.7 m (12.7

m deeper than the original depth). PDA test was carried out again and

no sign of damage was observed. The CAPWAP capacity was increased

from 6,715 kN to 7,741 kN.

In Table 2, it can be seen that pile no C9A-1 with CAPWAP capacity

of 6,715 kN, which is less than the maximum test load of 7,200 kN,

marginally passed the total settlement requirement but failed to satisfy

the residual settlement requirement. However, for piles C5A-3 and C4A-3,

although their CAPWAP capacities (6,728 kN and 7,147 kN respectively)

are also lower than 7,200 kN, they passed the loading test. Same as other

testing methods, CAPWAP can only compare with load testing method

within a certain tolerance. From Table 2, it is however clearly indicated

that the lower the CAPWAP value, the lower the load test result, and

CAPWAP can predict the lower bound pile capacity confidently. The two

piles C5A-3 and C4A-3 that passed the loading test may also be due to

the following considerations:

Pile Capacity Not Fully Mobilised during the Dynamic Load Testing

It should be noted that dynamic load testing (eg PDA test) only indicates

the activated or mobilised pile capacity at the time of testing. At very low

set per blow, dynamic test methods tend to produce lower bound capacity

estimates because the resistance is not fully activated particularly at and

near the toe as the movement is less than the quake value (quake is the

minimum relative movement between the pile and the soil for activation

of ultimate resistance) (Goble and Likins, 1996). Using dynamic testing

to determine the capacity of a pile with low set per blow is analogous to

a static load test that could not be run to failure due to limited loading

capacity or limited reaction system capability (Hussein et al., 2002). In

both cases, the loads determined are the proof loads of the piles but

not the ultimate loads.

The set per blow for piles no C5A-3 and C4A-3 are very small, ie 0.2 mm

and 0.6 mm respectively. Therefore, the pile capacity may not be fully

mobilised during the dynamic load testing. This argument is supported by

the CAPWAP analysis results which show that the shaft friction near the

toe for piles C5A-3 and C4A-3 has not been fully mobilised (Figs 2 to 4).

Table 2 – Loading Test for Zone 1

C9A-1 60.0 1.8 9/12/08 6,715 16/12/08 7 7,200 80.903 80.268 20.067 24.309

C5A-3 60.3 0.2 9/12/08 6,728 24/12/08 15 7,200 81.336 65.167 16.292 3.864

C4A-3 60.4 0.6 9/12/08 7,147 2/1/09 24 7,200 81.435 60.839 15.21 0.589

Actual

Residual

Settlement

(mm)

Pile No

for

Loading

Test

Embedded

Length of

Pile (m)

Set/Blow

(mm)

Date of

PDA Test

CAPWAP

Capacity

(kN)

Date of

Loading

Test

No of

Days after

PDA Test

Test Load

(kN)

Allowable

Total

Settlement

(mm)

Actual

Total

Settlement

(mm)

Allowable

Residual

Settlement

(mm)

Figure 4 – Shaft Resistance Distribution for C9A-1

Figure 2 – Shaft Resistance Distribution for C5A-3

Figure 3 – Shaft Resistance Distribution for C4A-3

THE HKIE TRANSACTIONS • Volume 18 Number 236

Page 4

was increased to 385. The theoretical safe loading capacities of piles

for each zone are shown in Table 3.

The loading test results of piles for Zone 2, 2A, 3, 4 and 4A as summarised

in Table 3 all satisfied the contract specification.

effect), and the capacity mobilisation (set per blow) have a significant

effect on pile capacity prediction. The CAPWAP analysis can predict the

static load test results very accurately for piles where the influence of pile

capacity mobilisation and set-up effect is minimal. When the influence

of the two factors is large, CAPWAP analysis underestimates the static

load test results fairly significantly.

Triantafyllidis (2001) pointed out that in assessing pile capacity using the

Hiley Formula, if the pile is relatively long, it is worth considering only

that portion of the pile that is affected for the duration of the impact (L

c

).

During impact between hammer and pile, waves are generated travelling

towards the pile toe. In the case of friction acting at the perimeter of

the pile or toe resistance, these waves are partly reflected, generating

waves travelling upwards to the pile top and causing separation between

hammer and pile.

L

c

is related to the duration of the impact where pile and hammer are

still in contact (t

c

) (ie compression stresses are applied at the pile top)

and speed of wave propagation in the pile material (c) as follows:

2L

c

= ct

c

From Triantafyllidis (2001), it is noted that L

c

depends on weight of

hammer, maximum compressive stress/impact velocity, pile impedance

and skin friction on the pile shaft while driving.

The PDA results of all piles in this project site indicate that the time,

t

c

, where pile head is in compression, is greater than 2L/c, where L

is the total pile length. This implies that the whole pile length was

affected for the duration of impact. This may be due to the fact that a

20-tonne hydraulic hammer was used for driving in this project instead

of a 6.2-tonne diesel hammer as quoted in the illustrative examples of

Triantafyllidis (2001).

The Use of Lower Capacity Piles

Some of the trial piles were driven to 76 m deep but still could not

achieve the pile capacity of 3,600 kN. The Contractor proposed to use

a mixture of 305 x 305 x 180 kg/m and 305 x 305 x 223 kg/m steel

H-piles instead of all 305 x 305 x 223 kg/m steel H-piles. 305 x 305 x

180 kg/m steel H-piles can be driven to final set at shallower depths. In some

areas where the bearing strata are deep, the Contractor proposed to use

lower capacity steel H-piles, ie 2,800 kN instead of 2,950 kN for

305 x 305 x 180 kg/m steel H-piles. The use of lower capacity steel

H-piles can result in shorter piles and facilitate the final set process

as well as reducing the risk of damaging long piles due to prolonged

driving. Since the cost and time incurred as a result of damaged piles,

removal of damaged piles and re-design of pile caps outweighed the

additional cost due to increase in number of piles, therefore, there is

an overall benefit in cost and time though the number of steel H-piles

Figure 5 – Set-up Effect of Piles at Different Time Intervals

Table 3 – Six Zones of Piling Construction

Zone No Theoretical Safe No of Pile

Loading Capacity (kN)

1 3,600 64

2 3,600 97

2A 3,600 6

3 3,450 77

4 2,950 132

4A 2,800 9

Total 385

Underground Obstructions

Obstructions were observed in some of the ground investigation

boreholes. While driving the steel H-piles, a large amount of obstruction

was encountered which caused damage of piles. The total number of

damaged piles was 76. There were no obvious signs before the piles

were damaged, even though the Contractor monitored the driving using

PDA. During the installation of trial piles in difficult ground conditions,

it may be worth using inclinometer reservation channels welded to the

steel H-pile profiles in order to monitor deviations. Some damaged piles

are shown in Fig 6. Figs 7 and 8 show extraction of damaged piles using

vibration clamp and hydraulic jack methods respectively. Pre-boring using

Odex method with temporary casing of 610 mm diameter was adopted

to penetrate through obstructions. The pre-bored hole was backfilled

with 10 mm aggregate and sand before driving steel H-pile. However,

the Contractor reported that the piles would easily hit the boulder again

at the pre-bored level. They revised their method by installing the steel

H-piles at the pre-bored level before backfilling the hole with aggregate

and sand. After backfilling, the steel H-piles were driven to final set and

PDA tests were carried out. Five pre-boring rigs were used to pre-bore

the obstructions and the total number of pre-bored piles was 193 at an

average depth of 40 m.

Founding Stratum

The average embedded pile length for this site is 60.6 m with the

maximum length of 79.0 m. Most of the steel H-piles were founded on

Figure 6 – Photo of Some Damaged Piles

THE HKIE TRANSACTIONS • Volume 18 Number 2 37

Page 5

References

1. Fung, W.K., Wong, C.T., Wong, M.K., A study on capacity predictions for driven

piles. The HKIE Transactions. Volume 11, No 3, pp10-16. Hong Kong (2004a).

2. Fung, W.K., Wong, C.T., Wong, M.K., Assessment of Load Carrying Capacity of

Driven Piles – A Practical Approach. The Structural Engineer. The Institution

of Structural Engineers. Volume 82, No 20. (2004b).

3. Fung, W.K., Wong, C.T., Wong, M.K., Response to Discussion by Victor Li and

Joley Lam on Observations on Using the Energy Obtained from Stress-wave

Measurements in the Hiley Formula. The HKIE Transactions. Volume 13, No

2, pp65-69. Hong Kong (2006).

4. Goble, G.G., Likins, G.E., On the Application of PDA Dynamic Pile Testing.

STRESSWAVE Conference 1996. Orlando, FL (1996).

5. Hussein, M.H., Sharp, M.R., Knight, W.F., The Use of Superposition for Evaluating

Pile Capacity. Deep Foundations 2002: An International Perspective on Theory,

Design, Construction and Performance (Geotechnical Special Publication No 116).

pp6-21. (Edited by O’Neill, M.W., Townsend, F.C.). American Society of Civil

Engineers. Orlando, FL (2002).

6. So, A.K.O., Ng, C.W.W., Performance of Long-driven H-piles in Granitic Saprolite.

ASCE Journal of Geotechnical and Geoenvironments engineers. Volume 135,

No 2, pp246-258. (2009).

7. Triantafyllidis, T., On the Application of the Hiley formula in driving long-piles.

Geotechnique. Volume 51, No 10, pp891-895. (2001).

8. Zhang, L.M., Dasaka, S.M., Uncertainties in Geologic Profiles vs Variability in

Pile Founding Depth. Journal of Geotechnical and Geoenvironmental Engineering

(doi: 10.1061/(ASCE)GT.1943-5606.0000364). ASCE. Retrieved 10 April, 2010.

Appendix A

Correlation of Parameters for the Hiley Formula

bearing strata with N values greater than 200. The depth of penetration

into N > 200 zone of some piles is as much as 15 m. This agrees with

the findings of Zhang and Dasaka (2010). Therefore, the rule of thumb

used by many engineers that steel H-piles could be founded on strata

with three consecutive N values greater than 200 may not be applicable

especially for 223 kg/m steel H-piles founded on soils with weak top

layers or if pre-boring is required.

Observations

(i) PDA test and CAPWAP analysis are valuable tools used for field

control and pile capacity assessment especially when the piles are

long, resulting in out of range of final set table.

(ii) In the situation where the bearing strata are deep and the risk

of damaging piles while driving is high, using piles with less pile

capacity can result in shorter piles which will reduce the risk of

damaging long piles, facilitate the final set process and shorten

installation time.

(iii) Whenever using CAPWAP (same for the other dynamic formulae)

to predict pile capacity, the set per blow and the timing of both

dynamic restrike test and static loading test relative to the end of

driving should be taken into consideration as pile capacity may not

be fully mobilised if set per blow is small. Also, an increase in the

pile capacity may occur after PDA test due to soil set-up. Therefore,

when piles having similar CAPWAP capacities are considered for

loading test, the pile with the largest set per blow should be chosen

for test.

(iv) The set-up effect of piles for this site is significant. It can also be

taken into consideration in pile design.

(v) Steel H-piles may not be founded on strata with three consecutive

N values greater than 200 especially for 223 kg/m steel H-piles

founded on soils with weak top layers or if pre-boring is required.

Figure 7 – Extraction of H-pile by Vibration Clamp

Figure 8 – Extraction of H-pile by Hydraulic Jack

Pile No C

p

+ C

q

Set/ Pile CAPWAP Hiley Capacity

(mm) 10 Blows Length Capacity for E

h

= 0.93,

(mm) (m) (kN) e = 0.65 (kN)

DC9-2 54 16 52.0 8,131 6,681

C3F-1 56 3 65.0 7,529 6,517

C6H-2 59 3 66.9 7,234 6,185

C8A-1 54 24 67.0 7,205 6,262

C9G-1 56 27 82.3 7,177 5,810

CAP4-5 55 25 67.0 7,159 6,146

C10D-3 55 34 60.3 7,153 6,083

C11G-1 56 36 72.2 6,851 5,786

Average = 7,305 6,184

7,305 x 0.85 =

6,209 > 6,184

C

c

= 5 mm for Plastic Cushion

20t Hammer Weight = 197.2 kN

Helmet Weight = 31.22 kN

Drop Height = 1.5 m

THE HKIE TRANSACTIONS • Volume 18 Number 238

TE

C

H

N

IC

A

L

N

O

TE

The Application of PDA/CAPWAP to

Ensure Quality and Capacity in Driving

Long Steel H-piles

The methodology of application of Pile Driving Analyser (PDA) test and Case Pile Wave Analysis

Programme (CAPWAP) analysis for quality control and capacity assessment in driving long

steel H-piles is fully illustrated in the project ‘Sun Yat Sen Memorial Park and Swimming Pool

Complex’. The deficiencies of using the Hiley Formula as a field control to assess pile capacity

have been well known. These deficiencies are more prominent when the pile is long. In the use

of the Hiley Formula for long driven piles, out of range of final set table has been a common

phenomenon and this has rendered the formula not applicable. The paper discusses the use of

a hybrid method making use of advanced method (PDA/CAPWAP) in conjunction with simple

method (Hiley Formula) for field control and pile capacity assessment. It has been shown that this

method will not only improve the reliability of the pile foundation but also contribute greatly to

the overall cost-effectiveness of installation of the piles by reducing the chance of over-driving

and damage to the pile. This method also enables the successful use of hydraulic hammers in

taking final sets.

Keywords: Pile Driving Analyser (PDA), Case Pile Wave Analysis Programme (CAPWAP),

Hiley Formula, Steel H-piles, Final Set Table, Hydraulic Hammer

W W LI

Architectural Services Department,

the HKSAR Government

M K WONG

Architectural Services Department,

the HKSAR Government

Y K CHAN

Architectural Services Department,

the HKSAR Government

The Hong Kong Institution of Engineers Transactions, Vol 18, No 2, pp34-49

Paper T0819-201002; Received 8 March 2010; Accepted 13 July 2010

Introduction

For pile driving in Hong Kong, the Hiley Formula has been widely used as

a field control to assess pile capacity despite its deficiencies. All dynamic

formulae suffer from inherent problems including poor representation of

the hammer, driving system, pile and soil. These deficiencies are more

prominent when the pile is long. In the use of the Hiley Formula for long

driven piles, ie pile length that is greater than about 55 m to 60 m, out

of range of final set table has been a common phenomenon and this

has rendered the formula not applicable. A pile-driving criterion that

is applicable to long driven piles as well as user-friendly is called for.

Accurate determination of the driving criterion will not only improve the

reliability of the pile foundation but also contribute greatly to the overall

cost-effectiveness of installation of the piles by reducing the chance of

over-driving and damage to the pile.

Quality Control on Pile Driving

Fung et al. (2004b) investigated the reliability of CAPWAP analysis in

pile capacity prediction as compared with results from static load tests

in local soil conditions. It has been shown that CAPWAP analysis is a

fairly accurate method for driven pile capacity prediction.

Fung et al. (2004b) also compared pile capacity predicted by CAPWAP

analysis and by the Hiley Formula (where the coefficient of restitution (e)

and drop efficiency (E

h

) were obtained by back analysis from CAPWAP

results of trial piles) of 327 piles from 18 sites. The results revealed

that 94% of the piles have capacities predicted by the Hiley

Formula deviating from their corresponding CAPWAP capacities by less

than ± 10%.

So and Ng (2009) pointed out that among the pile capacity prediction

methods of PDA, the Hiley Formula and the HKCA Formula, PDA gave

the best estimation. It could be used in lieu of a static loading test for

verifying pile capacity when the piles are long.

In the last decade, the Architectural Services Department (ArchSD)

used Pile Driving Analyser (PDA) test and CAPWAP analysis for quality

control and capacity assessment of driven piles installation. PDA has

been used to detect the integrity of piles, hammer efficiency and give

a rough indication of pile capacity. The data of force and velocity of

piles obtained from the PDA tests can be further analysed by CAPWAP

to give a more accurate assessment of pile capacity.

Carrying out PDA test and CAPWAP analysis for each pile for acceptance

may affect the site progress of projects with restricted driving time to a

certain extent. Therefore, a simple tool to assess pile capacity during the

final set of the driving process is still in need. Despite various shortcomings

of the Hiley Formula as mentioned above, it has been shown that the

Hiley Formula can take up the role after calibration by the CAPWAP

analysis to determine the efficiency of hammer (E

h

) and the coefficient

of restitution of hammer cushion (e) of the hydraulic hammer (Fung et

al., 2004a). Strictly speaking, the ArchSD’s method is a hybrid method

making use of advanced method (CAPWAP) in conjunction with simple

method (Hiley Formula). The resulting method is therefore handy to use

but avoids the drawbacks of the Hiley Formula. The details of the pile

acceptance process are as follows:

Simple Ground Conditions – The efficiency of the hammer (E

h

) and

the coefficient of restitution of the hammer cushion (e) of the hydraulic

hammer shall be determined from/verified by CAPWAP analysis of trial

piles. The combination of E

h

and e shall be so chosen such that when

these values are substituted into the Hiley Formula, the average of the

predicted bearing capacity of the trial piles is not higher than 85% of

the average CAPWAP capacity. Once these parameters are determined,

a final set table can be established for the contract. The factor of 85%

accounts for possible deviations of CAPWAP predictions from static load

test results and for the gradual decrease in driving system efficiency due

to deterioration in the condition of the hammer cushion. Normally, the

e value of cushion decreases with sustained use and the efficiency of

the driving system becomes lower when compared with that observed

during trial pile installation (Fung et al., 2004b).

Difficult Ground Conditions – In cases where measured final sets

are out of range of the set table with E

h

and e so chosen, all the piles

falling into this category shall be subject to CAPWAP analysis. To further

ensure quality, at least the pile with the lowest CAPWAP capacity shall

be load tested for acceptance. This provision can eliminate the risk of

pile over-driving which is common for very long pile or difficult ground

conditions, and as a result causing less pile damage.

THE HKIE TRANSACTIONS • Volume 18 Number 234

Page 2

The above methods have been used by the ArchSD as a field control

to assess pile capacity since 2003. From experience obtained to date,

they provide an accurate and workable tool to fulfil the requirement

of quality control and assessment of pile capacity. It also enables the

successful use of hydraulic hammers in taking final sets.

Difficult Ground Conditions – Sun Yat Sen Memorial

Park and Swimming Pool Complex

Ground Conditions

The project site is a challenging and difficult site in difficult ground

condition with bedrock at 56 m to 94 m below ground.

The borehole information indicates that the stratification of site comprises

mainly, Fill, Marine Deposit, Alluvium and Completely Decomposed

Granite.

Fill ranges in thickness from 19 m to 27 m with many boulders, cobbles

and gravels.

Marine Deposit with thickness ranging from 2 m to 5 m lies beneath

the Fill.

Alluvium, which is encountered beneath the Marine Deposit, has a

thickness ranging from 2 m to 6 m.

Completely Decomposed Granite with thickness ranging from 21 m to

62 m lies beneath the Alluvium. Rockhead level lies between 56 m and

94 m below ground.

Difficulties of the Site

It was expected that the steel H-piles would be founded at the residual

soil with SPT N-values > 200, which was in general approximately at

-52 mPD. However, the geological profile of the site shows that the layer

of soils with SPT N-values > 200 is deeper in the central portion of the

site where longer piles were expected. The installation of driven steel

H-piles with some of them longer than 70 m in a site with many boulders

and cobbles would be the main challenges to the Contractor. Since the

piles would be long, out of range of final set table was expected if the

Hiley Formula was used. The presence of many boulders and cobbles

might also affect the progress of works significantly.

Trial Pile Installation

In this project, the Contractor adopted 305 x 305 x 223 kg/m steel

H-pile as foundation. 328 steel H-piles at an estimated average depth

of 56 m were adopted with theoretical safe pile capacity of 3,600 kN.

Eight trial piles were selected for PDA tests with CAPWAP analysis to

determine the efficiency (E

h

) and the coefficient of restitution of the

hammer cushion (e) of the 20-tonne hydraulic hammer for final set.

These piles were spread evenly across the site including the area where

the longest pile depth is expected. Before the start of pile driving, visual

inspection of the type and quality of the hammer cushion was carried

out. The ram and helmet of the hammer were weighed. The results of

correlation are shown in Appendix A.

It was observed that all trial piles were out of range of the final set

table calculated using the correlated Hiley Formula based on the E

h

and e determined by CAPWAP analysis. In view of the trial pile results,

the Contractor then chose to use the acceptance method for difficult

ground conditions, ie carrying out PDA test and CAPWAP analysis for

every pile for assessing pile capacity and load testing the pile with the

lowest CAPWAP capacity for acceptance.

It was also observed that for long piles where the bearing strata are

deep, the CAPWAP capacities are lower (some of them are even lower

than 7,200 kN) because the Contractor was reluctant to drive the piles

to a deeper level to achieve a higher capacity for fear of damaging the

piles. This, however, may be acceptable due to potential set-up effect.

PDA Tests

In order to start the pile cap works early by phases, the site area was

originally divided into four zones and further increased to six zones later

due to difficulties encountered in driving piles to achieve the design

capacity in some areas of the site. PDA tests were carried out for each

pile by a testing firm employed by the Contractor. Based on the results of

the PDA tests, the piles with lower capacity (determined by Case method)

were selected to be tested by PDA again by an independent testing firm

employed by the ArchSD. CAPWAP analyses were also carried out for

these piles and the pile with the lowest CAPWAP capacity of each zone

was load tested for acceptance. Table 1 below summarises the number

of piles selected for PDA test for each zone.

Table 1 – PDA Tests Carried out for Different Zones

Zone No No of Piles No of PDA Tests No of PDA Tests

Carried out by Carried out by

the Contractor (%) the ArchSD (%)

1 64 64 (100%) 9 (14%)

2 97 97 (100%) 16 (16%)

2A 6 6 (100%) 6 (100%)

3 77 77 (100%) 18 (23%)

4 132 132 (100%) 14 (11%)

4A 9 9 (100%) 1 (11%)

First Loading Test

After the completion of pile installation for the first zone, loading test

was carried out for the pile with the lowest CAPWAP capacity (C9A-1,

6,715 kN). The loading test was carried out on 16 December 2008,

seven days after the PDA test. The maximum test load applied was

7,200 kN. The load test results failed to satisfy the contract specification.

According to the contract specification, if a pile fails the static load

test, additional load tests need to be carried out for other two piles for

acceptance. Therefore, piles with the second and third lowest CAPWAP

capacity of the zone, ie C5A-3 (6,728 kN) and C4A-3 (7,147 kN), were

chosen for static load test. Both piles passed the static load tests. The

borehole logs and test results of the three piles are shown in Fig 1 and

(SPT of 200 blows in 250 mm penetration is denoted as 200/250)

Figure 1 – Borehole Logs

THE HKIE TRANSACTIONS • Volume 18 Number 2 35

Page 3

Set-up Effect

The time lag between the PDA test and the static load test for piles no

C5A-3 and C4A-3 are 15 days and 24 days respectively. Increase in the

pile capacity may occur after PDA test due to soil set-up.

The set-up effect of this site has been studied. PDA tests and CAPWAP

analyses were also carried out for 51 piles at different time interval after

the end of driving in order to investigate the set-up effect. Static load

tests had not been carried out on these piles. The results show that pile

capacity increases ranging from 1 – 52% between restrike tests carried

out after time lapse in the range of 3 to 187 days. It is also noted that

there was a decrease in the set per blow of the second restrike tests.

If more energy had been used to drive the piles in the second restrike

tests to produce a larger set per blow, the CAPWAP prediction would

have been even higher (ie higher set-up effect). The plot of pile capacity

variations with log time is shown in Fig 5.

Fung et al. (2006) pointed out that the timing of both the static load

test and the dynamic restrike test relative to the end of driving (set-up

Table 2 respectively. C9A-1 was re-driven to a depth of 72.7 m (12.7

m deeper than the original depth). PDA test was carried out again and

no sign of damage was observed. The CAPWAP capacity was increased

from 6,715 kN to 7,741 kN.

In Table 2, it can be seen that pile no C9A-1 with CAPWAP capacity

of 6,715 kN, which is less than the maximum test load of 7,200 kN,

marginally passed the total settlement requirement but failed to satisfy

the residual settlement requirement. However, for piles C5A-3 and C4A-3,

although their CAPWAP capacities (6,728 kN and 7,147 kN respectively)

are also lower than 7,200 kN, they passed the loading test. Same as other

testing methods, CAPWAP can only compare with load testing method

within a certain tolerance. From Table 2, it is however clearly indicated

that the lower the CAPWAP value, the lower the load test result, and

CAPWAP can predict the lower bound pile capacity confidently. The two

piles C5A-3 and C4A-3 that passed the loading test may also be due to

the following considerations:

Pile Capacity Not Fully Mobilised during the Dynamic Load Testing

It should be noted that dynamic load testing (eg PDA test) only indicates

the activated or mobilised pile capacity at the time of testing. At very low

set per blow, dynamic test methods tend to produce lower bound capacity

estimates because the resistance is not fully activated particularly at and

near the toe as the movement is less than the quake value (quake is the

minimum relative movement between the pile and the soil for activation

of ultimate resistance) (Goble and Likins, 1996). Using dynamic testing

to determine the capacity of a pile with low set per blow is analogous to

a static load test that could not be run to failure due to limited loading

capacity or limited reaction system capability (Hussein et al., 2002). In

both cases, the loads determined are the proof loads of the piles but

not the ultimate loads.

The set per blow for piles no C5A-3 and C4A-3 are very small, ie 0.2 mm

and 0.6 mm respectively. Therefore, the pile capacity may not be fully

mobilised during the dynamic load testing. This argument is supported by

the CAPWAP analysis results which show that the shaft friction near the

toe for piles C5A-3 and C4A-3 has not been fully mobilised (Figs 2 to 4).

Table 2 – Loading Test for Zone 1

C9A-1 60.0 1.8 9/12/08 6,715 16/12/08 7 7,200 80.903 80.268 20.067 24.309

C5A-3 60.3 0.2 9/12/08 6,728 24/12/08 15 7,200 81.336 65.167 16.292 3.864

C4A-3 60.4 0.6 9/12/08 7,147 2/1/09 24 7,200 81.435 60.839 15.21 0.589

Actual

Residual

Settlement

(mm)

Pile No

for

Loading

Test

Embedded

Length of

Pile (m)

Set/Blow

(mm)

Date of

PDA Test

CAPWAP

Capacity

(kN)

Date of

Loading

Test

No of

Days after

PDA Test

Test Load

(kN)

Allowable

Total

Settlement

(mm)

Actual

Total

Settlement

(mm)

Allowable

Residual

Settlement

(mm)

Figure 4 – Shaft Resistance Distribution for C9A-1

Figure 2 – Shaft Resistance Distribution for C5A-3

Figure 3 – Shaft Resistance Distribution for C4A-3

THE HKIE TRANSACTIONS • Volume 18 Number 236

Page 4

was increased to 385. The theoretical safe loading capacities of piles

for each zone are shown in Table 3.

The loading test results of piles for Zone 2, 2A, 3, 4 and 4A as summarised

in Table 3 all satisfied the contract specification.

effect), and the capacity mobilisation (set per blow) have a significant

effect on pile capacity prediction. The CAPWAP analysis can predict the

static load test results very accurately for piles where the influence of pile

capacity mobilisation and set-up effect is minimal. When the influence

of the two factors is large, CAPWAP analysis underestimates the static

load test results fairly significantly.

Triantafyllidis (2001) pointed out that in assessing pile capacity using the

Hiley Formula, if the pile is relatively long, it is worth considering only

that portion of the pile that is affected for the duration of the impact (L

c

).

During impact between hammer and pile, waves are generated travelling

towards the pile toe. In the case of friction acting at the perimeter of

the pile or toe resistance, these waves are partly reflected, generating

waves travelling upwards to the pile top and causing separation between

hammer and pile.

L

c

is related to the duration of the impact where pile and hammer are

still in contact (t

c

) (ie compression stresses are applied at the pile top)

and speed of wave propagation in the pile material (c) as follows:

2L

c

= ct

c

From Triantafyllidis (2001), it is noted that L

c

depends on weight of

hammer, maximum compressive stress/impact velocity, pile impedance

and skin friction on the pile shaft while driving.

The PDA results of all piles in this project site indicate that the time,

t

c

, where pile head is in compression, is greater than 2L/c, where L

is the total pile length. This implies that the whole pile length was

affected for the duration of impact. This may be due to the fact that a

20-tonne hydraulic hammer was used for driving in this project instead

of a 6.2-tonne diesel hammer as quoted in the illustrative examples of

Triantafyllidis (2001).

The Use of Lower Capacity Piles

Some of the trial piles were driven to 76 m deep but still could not

achieve the pile capacity of 3,600 kN. The Contractor proposed to use

a mixture of 305 x 305 x 180 kg/m and 305 x 305 x 223 kg/m steel

H-piles instead of all 305 x 305 x 223 kg/m steel H-piles. 305 x 305 x

180 kg/m steel H-piles can be driven to final set at shallower depths. In some

areas where the bearing strata are deep, the Contractor proposed to use

lower capacity steel H-piles, ie 2,800 kN instead of 2,950 kN for

305 x 305 x 180 kg/m steel H-piles. The use of lower capacity steel

H-piles can result in shorter piles and facilitate the final set process

as well as reducing the risk of damaging long piles due to prolonged

driving. Since the cost and time incurred as a result of damaged piles,

removal of damaged piles and re-design of pile caps outweighed the

additional cost due to increase in number of piles, therefore, there is

an overall benefit in cost and time though the number of steel H-piles

Figure 5 – Set-up Effect of Piles at Different Time Intervals

Table 3 – Six Zones of Piling Construction

Zone No Theoretical Safe No of Pile

Loading Capacity (kN)

1 3,600 64

2 3,600 97

2A 3,600 6

3 3,450 77

4 2,950 132

4A 2,800 9

Total 385

Underground Obstructions

Obstructions were observed in some of the ground investigation

boreholes. While driving the steel H-piles, a large amount of obstruction

was encountered which caused damage of piles. The total number of

damaged piles was 76. There were no obvious signs before the piles

were damaged, even though the Contractor monitored the driving using

PDA. During the installation of trial piles in difficult ground conditions,

it may be worth using inclinometer reservation channels welded to the

steel H-pile profiles in order to monitor deviations. Some damaged piles

are shown in Fig 6. Figs 7 and 8 show extraction of damaged piles using

vibration clamp and hydraulic jack methods respectively. Pre-boring using

Odex method with temporary casing of 610 mm diameter was adopted

to penetrate through obstructions. The pre-bored hole was backfilled

with 10 mm aggregate and sand before driving steel H-pile. However,

the Contractor reported that the piles would easily hit the boulder again

at the pre-bored level. They revised their method by installing the steel

H-piles at the pre-bored level before backfilling the hole with aggregate

and sand. After backfilling, the steel H-piles were driven to final set and

PDA tests were carried out. Five pre-boring rigs were used to pre-bore

the obstructions and the total number of pre-bored piles was 193 at an

average depth of 40 m.

Founding Stratum

The average embedded pile length for this site is 60.6 m with the

maximum length of 79.0 m. Most of the steel H-piles were founded on

Figure 6 – Photo of Some Damaged Piles

THE HKIE TRANSACTIONS • Volume 18 Number 2 37

Page 5

References

1. Fung, W.K., Wong, C.T., Wong, M.K., A study on capacity predictions for driven

piles. The HKIE Transactions. Volume 11, No 3, pp10-16. Hong Kong (2004a).

2. Fung, W.K., Wong, C.T., Wong, M.K., Assessment of Load Carrying Capacity of

Driven Piles – A Practical Approach. The Structural Engineer. The Institution

of Structural Engineers. Volume 82, No 20. (2004b).

3. Fung, W.K., Wong, C.T., Wong, M.K., Response to Discussion by Victor Li and

Joley Lam on Observations on Using the Energy Obtained from Stress-wave

Measurements in the Hiley Formula. The HKIE Transactions. Volume 13, No

2, pp65-69. Hong Kong (2006).

4. Goble, G.G., Likins, G.E., On the Application of PDA Dynamic Pile Testing.

STRESSWAVE Conference 1996. Orlando, FL (1996).

5. Hussein, M.H., Sharp, M.R., Knight, W.F., The Use of Superposition for Evaluating

Pile Capacity. Deep Foundations 2002: An International Perspective on Theory,

Design, Construction and Performance (Geotechnical Special Publication No 116).

pp6-21. (Edited by O’Neill, M.W., Townsend, F.C.). American Society of Civil

Engineers. Orlando, FL (2002).

6. So, A.K.O., Ng, C.W.W., Performance of Long-driven H-piles in Granitic Saprolite.

ASCE Journal of Geotechnical and Geoenvironments engineers. Volume 135,

No 2, pp246-258. (2009).

7. Triantafyllidis, T., On the Application of the Hiley formula in driving long-piles.

Geotechnique. Volume 51, No 10, pp891-895. (2001).

8. Zhang, L.M., Dasaka, S.M., Uncertainties in Geologic Profiles vs Variability in

Pile Founding Depth. Journal of Geotechnical and Geoenvironmental Engineering

(doi: 10.1061/(ASCE)GT.1943-5606.0000364). ASCE. Retrieved 10 April, 2010.

Appendix A

Correlation of Parameters for the Hiley Formula

bearing strata with N values greater than 200. The depth of penetration

into N > 200 zone of some piles is as much as 15 m. This agrees with

the findings of Zhang and Dasaka (2010). Therefore, the rule of thumb

used by many engineers that steel H-piles could be founded on strata

with three consecutive N values greater than 200 may not be applicable

especially for 223 kg/m steel H-piles founded on soils with weak top

layers or if pre-boring is required.

Observations

(i) PDA test and CAPWAP analysis are valuable tools used for field

control and pile capacity assessment especially when the piles are

long, resulting in out of range of final set table.

(ii) In the situation where the bearing strata are deep and the risk

of damaging piles while driving is high, using piles with less pile

capacity can result in shorter piles which will reduce the risk of

damaging long piles, facilitate the final set process and shorten

installation time.

(iii) Whenever using CAPWAP (same for the other dynamic formulae)

to predict pile capacity, the set per blow and the timing of both

dynamic restrike test and static loading test relative to the end of

driving should be taken into consideration as pile capacity may not

be fully mobilised if set per blow is small. Also, an increase in the

pile capacity may occur after PDA test due to soil set-up. Therefore,

when piles having similar CAPWAP capacities are considered for

loading test, the pile with the largest set per blow should be chosen

for test.

(iv) The set-up effect of piles for this site is significant. It can also be

taken into consideration in pile design.

(v) Steel H-piles may not be founded on strata with three consecutive

N values greater than 200 especially for 223 kg/m steel H-piles

founded on soils with weak top layers or if pre-boring is required.

Figure 7 – Extraction of H-pile by Vibration Clamp

Figure 8 – Extraction of H-pile by Hydraulic Jack

Pile No C

p

+ C

q

Set/ Pile CAPWAP Hiley Capacity

(mm) 10 Blows Length Capacity for E

h

= 0.93,

(mm) (m) (kN) e = 0.65 (kN)

DC9-2 54 16 52.0 8,131 6,681

C3F-1 56 3 65.0 7,529 6,517

C6H-2 59 3 66.9 7,234 6,185

C8A-1 54 24 67.0 7,205 6,262

C9G-1 56 27 82.3 7,177 5,810

CAP4-5 55 25 67.0 7,159 6,146

C10D-3 55 34 60.3 7,153 6,083

C11G-1 56 36 72.2 6,851 5,786

Average = 7,305 6,184

7,305 x 0.85 =

6,209 > 6,184

C

c

= 5 mm for Plastic Cushion

20t Hammer Weight = 197.2 kN

Helmet Weight = 31.22 kN

Drop Height = 1.5 m

THE HKIE TRANSACTIONS • Volume 18 Number 238