HKIE Transactions

ISSN: 1023-697X (Print) 2326-3733 (Online) Journal homepage: http://www.tandfonline.com/loi/thie20

Assessment of Pile Integrity by Low-Strain Stress

Wave Method

Jian-hua Yin BSc MSc PhD PEng MHKIE MCGS , Jian-xin Yuan & Ming-gui Liu

To cite this article: Jian-hua Yin BSc MSc PhD PEng MHKIE MCGS , Jian-xin Yuan & Ming-gui
Liu (1999) Assessment of Pile Integrity by Low-Strain Stress Wave Method, HKIE Transactions,
6:1, 42-49

To link to this article: https://doi.org/10.1080/1023697X.1999.10667791

Published online: 09 Apr 2013.

Submit your article to this journal

Article views: 40

View related articles

Full Terms & Conditions of access and use can be found at

http://www.tandfonline.com/action/journalInformation?journalCode=thie20
 Assessment of Pile Integrity by Low-Strain
Stress Wave Method
Jlan~hua Yin 8Se M5c PhD PEng MHKIE MCGS Mlng-gul L1u
The Department of CIvil & Structural Engineering, The Hong Kong The Institute of Rock and 50/1 Mechanics, The Chinese Academy of Sciences
Polytechnic University

Jlan-xln Yuan
The Institute of Rock and Soli Mechanics, The Chinese Academy of Sciences

This paper discusses the application and limitations Of the stress wave method for the assessment of pile Integrity. Equations
for the calculation Of wave reflection and transmission are derived as related to the Interpretation Of test results. An
idealized pile with a necking sec/ion is analyzed using the theoretical equal/ons to show typical wave refledions from the
necking sedion and from the pile toe. The Instrument used for pile Integrity testing is briefly described. Original data from a
field verif/cal/on exercise Of pile Integrity tesl/ng in Beijing, China are presented and interpreted. Original data from a test
performed on a precast prestressed concrete pile canneded In four segments in Hong Kong are also presented and
interpreted. The data measured from the field show that the measured velocity wave Is, in general, more complicated than
that described by the idealized stress wave equation. Factors affedlng test results and limitations Of the testing method are
discussed and summarized in this paper.
Keywords: Pile, Dynamic Pile Testing, Stress Wave, Wave Equal/on, Integrity, Case Study, Geotechnical

Introduction where u(x,t) Is displacement; 'II and tjl are two arbitrary fUnctions of
(x-et) and (x+ct) respectively. Eqn.(1) is known as d'Alemberfs solution.
In pife foundation engIneering, the assessment Of pile integrity Is an The item 'II (x~ct) represents a downward traveling wave and 4l (x+ct)
important aspect of the quality control of pile design and construction. represents an upward traveling wave, both at the same wave speed of
Issues concerning pHe integrity include (al defects in a pile, such as necking. c. The wave speed c is equal to ~(E/p), where E Is the Young's modulus
voids, etc., (b) possible weak soil or voids beneath pile toe, (cl broken of pile and p is the density of pile. Expressions for velocity waves and
pile and (d) checking the actual pile length against the designed pile force waves and their relationships have been derived and presented
length (Yuan 1988a and 1988b, GEO 1996). For example, a cast-In- systematically in Appendix 1. For example, the downward force, pt, Is
place pile may be contaminated with soils due to either the collapse of related 10 downward velocity, vt, by
soils, poor casing, or improper lifting of casing. For driven piles, fractures
may be induced during installation by the dynamiC hammer impact due (2)
to either large tensile stress, mis-alignment of pile or deflection by boulders
underneath. where Z AE/c= A..J(Ep), known as pile impedance. The arrow ·.1.-"" stands
There are different methods for pile integrity testing - (a) coring tests in for downward direction of wave propagation, and "j " for upward direction.
concrete piles, (b) sonic profiling (or logging) tests by measuring Any change In impedance Z at a pile section will cause wave reflection
transmission between two probes In two vertical tubes cast into a plte, and transmission at this pile section. Impedance changes may be due
(c) vibration (impedance) tests, and (d) the method based on the stress to variations In pile section area, A, Young's modulus, E, and/or denSity,
wave theory (TIJou 1984, Ellway 1987). This paper discusses only the p, of pile (or the wave velocity, c). Expressions for reflected waves and
last method based on the stress wave theory In the time domain analysis transmitted waves due to known downward or upward waves have been
with a low strain induced in the pile. The testing based on this method derived and presented systematically in Appendix 1 for practical
is referred to as pile Integrity testing in the following presentation and applications. The reflected velocity waves will arrive at the pile head
discussions. This kind of pile Integrity testing is based on the principle and the velocity of a reflected wave will be doubled. Th is doubled velocity
of one-dimensional longitudinal wave reflections (Kolsky 1953, Smith may be measured with an electronic transducer. The reflected waves
1955, Weltman 1977, Ellway 1987, Yuan 1988a and 1988b, Paquet carry the information of the changes In impedance Inside the pile and
1992, Tham et al. 1996). The location of pile defects may be calculated the supporting condition at the pile toe up to the head of the pile. In
using the traveling time of reflected velocity waves recorded by transducers this way, the defects or abnorma11tles inside the pile may be detected
usually placed on pile head. This paper discusses (a) the stress wave at pile head using electroniC instruments and the stress wave analysis.
method, its applications and limitations, (b) the instrument used for testing, This Integrity testing method is a non-destructive testing method, known
(c) case studies, and (d) factors affecting test results. as either -low-strain" stress wave method, wave ·refledlon", sonic "echo"
(or wave ·echo") or seismic method. In order to discuss the application
Stress Wave Method and limitations of this testing method with a special reference to pUe
Integrity testing in Hong Kong, an Idealized pile with a necking section
The derivation of an one-dimensional stress wave equation for low-strain and free-toe-support Is discussed first as followings.
pile integrity testing can be found In Kolsky (1953), Smith (1955) and
Fig.Ha} illustrates a pile with a decreased impedance at pile section Pw
Tlmoshenko and Goodier (1970). A general solution to the stress wave
The Impedance I Ain pile section PAis equal to I ( In pile section P(' I.e.
equation without considering soli resistance can be written as follows
IA = Zc But the Impedance, I B, at pile sectlon PB is a quarter of ZA' I.e.
(see Appendix 1) ZB = Z/ 4. The pile toe Is assumed to be a free-end.

u(x, I) = 'tI (x - ct) + $ (x + ct) (1 ) tt Is assumed that a short pulse of compressive force wave in a rectangUlar

The Hong Kong Institution Of Engineers

 of velocity. In high-strain pile testing. two force transducers lstrain gauges)
(0) (b) , and two accelerometers are placed at Section k"k to measure the time-
history of both force wave and velocity wave. The data obtained can
be used to assess the integrity and static bearing capacity of the pile.
This high-strain pile testing method will not be discussed in this paper.
~h- "- .. _" - .,-" -" _. k
The time tl can be used to estimate the location of the Section A-B,
that is LA= t 1(.../2, assuming that the wave speed, CAl is known. The velocity
arriving at the pile head at time, t 21 is from the reflection at Section B-
C. As shown in Fig.1 ('0, the velocity Is -0.768v, negative. The time difference
(t 2 -t 11 can be used to estimate the length of the pile section Pgo The
calculation is La = (t 2-t,)cl2, again assuming the wave speed, cB' of Section
Pu is known.
Before the initial wave is reflected from the pile toe, the secondalY waves
are reflected from the Section B-C. However, the absolute value of the
, :, ,
velocity reflected is relatively small and gets smaller with time. For example,
! i
the second reflected wave from Section B-C has a velocity of only -O.276v
: ,
on arriving at the pile head at time t3' Another wave, NO.(12), reflected
,'','
"
.'
:i
: i
, from the pile head is reflected from Section 8-C as No.(13) . This wave
.,
:: arrives at the pile head at time t4 before the initial wave is reflected
from the pile toe. The wave No.(13) is reflected again at the pile head
p (e) , and the velocity value Is doubled to be -O.460v. This value is smaller
S«tion k-k:
than the velocity of O.S20v, reflected from the pile toe and Is not to be
confused with the latter wave.
The wave reflected from the pile toe which is assumed freely supported
is O.820v, arriving at time Is' If there are voids under the pile toe or
the stiffness of Ihe soli is much smaller than the stiffness of the pUe,
the pile toe may be assumed approximately freely supported. But when
the pile is supported on bedrock and there are no voids, the pile toe
shall be assumed fixed. If the pile toe is fixed, the velocity reflected
from the pile toe will be negative. This means that the sign of the velocity
reflected from pile toe may be used to assess the supporting condition
at the pile toe. The total pile length may be estimated as (see Fig.l)
(0 , L = (3)
Pile head, x"'o: 0 1---tU~-;;:P.!------- •

Figure 1 - Wave propagation in an IdeaUsed free-supported pile with a where c( is the wave speed for pile section PC' If cA = Cs = c(1 then L =
necking sedion, (a) an idealised pile, (b) wave propagation in x-t co- t5 c/ 2 .
ordinates, (c) force wave passing Section k-" (d) velocity wave passing With Increase of time, more secondary waves will be reflected from Section
Section k-k, Ie) force at pile head, and If) velocity at pUe head. A-B, Section 8-C and the pile loe. The velocity magnitudes will get smaller
and smaller. If more than one impedance changes exist in the pile, the
shape, P, is induced at the pile head due to, say, the impact of a hammer.
calculatIon method for wave reflections is similar to that for piles with
The force wave is a downward wave. The co rresponding velocity wave
one impedance change, but the calculatlon will be more complicated
Is also compressive and downward. Using Eqn.(2), the magnitude of the
and time-consuming.
velocity is v = P/Z. The reflection and transmission of force and velocity
waves have been calculated in Appendix 2. The traveling paths of the In addition to impedance changes, the decrease in velocity magnitudes
waves are shown In Fig.l (b) . Consider a cross-section (Section k-k) at a can be caused by damping of the pile material and Ihe damping and
distance of Lk from the pile head. It is assumed that the time perIod , .6.t, reSistance of surrounding soils. The signals measured are normally
of the force pulse satisfies hot < tk = L/cA, where cA is the wave speed amplified using an exponential factor In order to identify weak wave
of pile section P ~. Thus there Is no wave superposition at Section k·k. reflections from defects in the deep segment and the pile toe .
The magnitudes of the force and velocity waves passing Section k-k are
shown in F(g.1 (c) and Fig.l(d). Instrument for Pile Integrity Testing
As shown in Fig.l (cl, at Section k-k, the force wave reflected from Section
As discussed in the preceding section, in order to identify defects in a
A-B with ZR = O.25Z A, is -D .GP, negative. The reflected velocity wave Is
pile, the time-history of velocity has to be measured. With recent
O.6v, positive. It is noted that if Zs > ZA' the reflected force wave at Sed ion
development in electronic engineering and computer technology, advanced
k-k would be positive, and the velocity would be negative. The signs of
and computerized instruments have been developed specially for dynamic
the reflected force wave and velocity wave are always opposite.
pile testing (Weltman 1977, PIT 1993, TNO 1993). The Instetute of Rock
When the reflected waves arrive at the pile head, the force will become and Soil Mechanics (lRSM) of the Chinese Academy of Sciences started
zero, but the velocity Is doubled. Therefore, a force transducer at the the development of instruments for dynamic pile testing in the early
pile head cannot measure any force. But velocity transducer (or 1980's (yuan 1988b). A series of RSM instruments have been developed.
accelerometer) may measure a value of 1.2v, which is positive and twice The instrument developed for pile integrity testing is known as the RSM-
of the upward velocity of O.6v at Section k-k. In low-strain testing, an PRT system. The RSM-PRT SYSTEM consists of four major components:
accelerometer is normally placed on the pile head to measure analog (a) the hardware (RSM-24FD - instrument), (b) one accelerometer and a
signals of the time-history of acceleration. The analog signals of the hammer, (c) a notebook computer and (d) a computer program (PRT -
acceleration is converted to digital data instantly through an analog to software). The RSM-24FD instrument includes four functional parts: (a)
digital card and then integrated numerically to obtain the time-history charge amplification, (b) signal filtration, (c) signal amplification, and (d)

TRANSACTIONS' Volume 6 Number 1

analog to digital conversion. A brief descri"ption of the four functional
parts is given below. D ~O.6m D~.8m D~O.6m D~O.6m

(a) Charge Amplification

The specially designed charge amplifier is an electronic device with two No.! No.2 No.3
special actions: (i) to amplify the charge signals measured using an L~!6m L ~ 16m L~ 16m
accelerometer and to convert them Into voltage signals and (iI) to provide
a low Input resistance and a high output resistance for the suitable
resistance adaptation between the output terminal of accelerometer and D~O.6m D~O . 6m D ~O.6m D ~ O.6m
the input terminal of the instrument. The vOltage value of the charge-
amptlfier's output Is directly proportional to the acceleratlon value
measured with an accelerometer attached at the pile head.
No.5 No.6 No.7 No.8
(b) Signal Filtration L~ 16m L~ 12m L~16m L~16m

Signals obtained from a transducer during field testing contain mechanical

and electronic noises. Therefore, the signals have to be filtered with a
D~O.6m D~.8m D~O . 6m D~O.6m
low-pass filter and a high-pass filter before being amplified and converted
into digital data. A low-pass filter can delete noises of high frequencies.
A high-pass filter can eliminate low frequency noises such as vibrations
of the field, pile/soil combination, and the zero-shift of accelerometer's
No.9 No.JO No.11 No.12
output. The filter designed with an electronic circuit is called an analog L~20m L ~ 16m L~16m L~12m
fitter. The PRT computer program includes a special sUb-program as a
digital filter.
Figure 2 - The diameters and lengths Of 12 piles at a test site
(c) Signal Amplification
partiCipants, but the location and type of defects within the 12 piles
Because the energy of each hammer blow transferred to the pile head remained confidential. The velocity of stress wave had been measured
varies over a very large range, it is difficult to select the most su itable on a pile of the same grade of concrete material as the 12 model piles
gain for an amplifier manually. If the gain is fixed during the signal and was found to be 3590m/sec before verification tests started. Data
sampling process after each blow, the output signal of the amplifier may from tests on two typical piles, piles NO.9 and No. 11 are presented
sometimes be too weak to be observed or alternatively it may overflow. below.
The floating-amplifIer adopted in the RSM-PRT system can always
Fig. 3(a) shows the relationship of velocity vs . time measured using RSM
automatically select the most suitable gain at each sampling time interval
on pile No.1 1. Fig. 3 (b) and Fig. 3 (c) show the relationships of velocity
during the whole testing process. Events of signals being too weak to
VS. pile length (based on known wave velocity) using PIT (1993) and
record and overflowing output will never occur.
TNO (1993) instruments on the same pile NO.11 . The time scale in Fig.
3 (a) is 22)1s = 22 x 10-6 sec. The time at the initial peak point, A, is
(d) Analog to Digital Conversion
The analog to digital converter used in the RSM-PRT system is an AID
integrated circuit with 16 bits. The floating amplifier mentioned in (c) (al RSM

produces a gain code varying from 0 to 8. Liu (1995) showed that the
final result of AID conversion using the floating amplifier in the RSM-
r--
PRT system had a dynamic range of 24 binary bits. The fastest sampling
A
W
rate of the AID conversion is 10)15. Paquet (1992) recommends a minimum
of 50)1s of sampling rate for detecting a O.4m separation of two reflectors
(detects). The sampling rate of the RSM-PRT system is lOlls and is good
for detecting a O.08m apart of two reflectors. The transducer used for
. ,. • 128 18G 2413 JOO 360 1Z0
B
(XZ2 lUI)

measuring acceleration at pile head is a type of piezo-electric device

, ,
(bl PIT
• , ,
called piezo-accelerometer. " " " " " " " ( m)

Field Verification - A Case Study

j
The stress wave theory developed for pile integrity testing and the
instrument described In the above section need to be verified for
engineering practice. Field verification using fuJi-scale model piles is always
an effective method for the examination of (a) the reliabi lity and accuracy
of the testing instruments and (b) the validation of the stress wave theory (e) TNO , ,
for pile integrity assessment. Tile Soil Dynamics Division (SOD) of the
VIbration Society of China and the Acoustic Measurement Division (AMD)
of the Acoustics SocIety of China have recently jointly organized a n(ltional
conference on "Testing Instruments and Technology for Pile and Foundation
Dynam ic Testing· from 25 May to 4 June 1995 in Beijing, China. During
the conference, both SOD and AMD organized field verification of testing
instruments and methods for pile integrity testing using instruments from
\ I
""
different countries, including PIT from U.S.A., TNO from Holland, and
RSM from IRSM.
a 2 4 6 B 16 12 14 16 18 26(m)

A total of 12 full-scale concrete model piles were installed at a site in

Beijing. Fig.2 shows the arrangement, diameters and pile lengths of the Figure 3 - Pile No. 11· (a) RSM results, (b) PIT resulls, and (c) rNO
12 plies. The pi le diameters and pile lengths were known to the results.

The Hong Kong Institution of Engineers

104 units = 104 x 22 x 10-6 sec = 2.288 x 10-3 sec. The time at the Testing of a Prestressed Concrete Pile in Hong Kong
second peak paint, B, is 456 units = 456 x 22 x 10-6 sec = 10.032 x
10·' sec. The length of pile is L ~ (t,-t.l x cJ2 ~ (10.032 x 10'- 2.288 Precast and prestressed hollow concrete piles are commonly used in Hong
X 10.3) x 3590/2 = 13.9m. Fig. 3(a) shows a clear reflected velocity wave Kong. Typical pile dimensions are (a) 500mm in outside diameter (OD)
of the same sign (downward) as the Initial velocity. At the end of the and 12Smm in thickness and (b) 8m or 12m in length for each segment.
verification exercise, it was discovered that the pile No.1 1 was only 14m A steel cap (female or male) was pre-installed at each end of the pile
in length, not 16m as claimed. This demonstrates that the pile Integrity segment. Two pile segments can be connected to each other by welding
testing can check the actual pile length against the designed pile length. the two steel caps at the ends of the two piles. The pll e Is normally
The velocity values between Point A and Point B were not zero as the driven into soils by a diesel hammer. The connection (alignment and
ideal case in Fig. 1. These non-zero values were an indication of the welding) can be carried out during the installation of the pile. The integrity
effects of the soil friction resistance on the wave propagation. of the pile such as the quality of the connection and fractures induced
during pile driving are the concern of design engineers (the static bearing
Fig. 4(a) shows the relationship of velocity vs. time measured using RSM
capacity and tension stress are other concerns which can be addressed
on pile NO.9. Fig. 4(b) and Fig. 4(c) show the relationships of velocity
in high-strain pile testing). The fractures inside a pile and breaking of
vs. pile length measured on the same pile using PIT and TNO instruments.
the welding may be caused by mis-alignment of the steel caps, boulders
The designed pile length was 20m. The time scale in Fig. 4(a) is 3111S
underground or poor welding. The low-strain pile testing based on stress-
= 31 x 10·6 sec. The time at the initial peak point, A, is 88 units = 88 x
wave theory may be used to assess the integrity of the driven piles and
31 x 1O-6sec = 2.728 x 10.3 sec. The time at the second peak paint, 8,
to check the actual total pile length against the designed pile length.
is 272 units = 272 x 31 x 10-6 sec = 8.432 x 10-3 sec. The length of
pile from Po int A to Point B was LA•B = (tB-t A) x c/2 = (8.432 x 10.3 - Fig. 5 shows the measured velocity vs. time of a typical precast prestressed
2.7 28 X 10.3 ) x 3590/ 2 = 1O.24m. This reflection at Point B has the concrete pile (00 = SOOmm and thickness = 125mm) in Hong Kong.
same sign (phase) as the initial velocity, indicating a weak impedance The designed pile length was 40m. Top segment of the pile as marked
in the pile (possible necking). The calculated pile length is 10.24m, which on the visible pile's side surface obselVed at the site was Sm. The velocity
Is smaller than the pile length of 20m so that it is considered that the magnitudes in Fig. 5 were amplified using an exponential amplification
reflected wave at Point 8 is not from the pile toe. The calculated length factor of 45 in order to identify weak reflections from pile toe. Based
of 10.24m is the location of ·possible necking" (smaller impedance) on the wave reflection from the joint between the fi rst and the second
of the pile. The time at Point C is 446 units = 446 x 31 x 10.6 sec = segment on the same pile and the reflection from the pile toe. the wave
13.826-3 sec. Assuming the wave speed is the same for the whole pile, velocity was estimated to be 4850m/s. The wave reflection from the
the total length of pile shall be L ~ (t, -t,)(e/ 2 ~ (13.826 x 10·' - 2.728 first welded joint was relatively stronger than the reflections from the
x 10" ) x 3590/2 ~ 19.92m, which is very close to the design pile length following welded joints but was much weaker than the initial velocity
of 20m. It is seen in Fig.4 that there are a number of reflected secondary value. This indicates that the connections at the first joint and the following
waves making major reflected waves more difficult to interpret than the joints were gOOd. Based on the wave reflection data, it is considered
ideal case in Flg.1. that the pile consisted of four pile segments of 8m, 12m, 12m and 8m
The above results show that (a) the stress wave method can be used each in length. The field pile installation records confirm that these are
for identification of the defects inside a pile and (b) the RSM-PRT system all true segment lengths. As the Interpretation involves an element of
can adequately measure and process the testing data. However, the subjective judgement, it is critically Important that suitable qualified and
measured time-history is more complicated than the ideal one as shown experienced personnel be engaged to carry out the assessment.
in Fig.1. The complication is due to two factors: (a) damping effects causing
rapid attenuation of the stress wave and (b) soH friction assistance which RS-Rsn-Z1rD PILE-nO.1Z7B fInl: 19/1Z/~6 15:96:17
generates confusing secondary waves (Paquet, 1992).

(a) RSM

\ .L 9 Z 1 6 8 19 lZ 11 16 18 Z9 ZZ (fts)

A B C
t
o '0 120 180 210 3 00 360 '120 ()(31).1s ) 8m I 12m I 12m I 8m I t- Pile segments

(b) PIT Figure 5 - Results Of testing on a professional concrete pile (4 segments)

o 2 4 6 8 10 12 14 16 18 20 r< (m)
in Hong Kong
The length to diameter ratio of this tested pile was 40m/0.5m = 80.
v Paquet (1992) suggested a maximum length for pile integrity testing
based on the a theoretical law of wave attenuation for a pile/soil system.
The length to diameter ratio may be in the range from 25 to 90, depending
on the soil types and the diameter of the pile. Ellway (19B7) found that
a length of 50 pile diameter was observed by son ic echowtests for piles
6

(e) TNO in soft soils. However, he suggested that the effective length for pile
integrity testing be limited to 30 pile diameters for the technology in
1980·s. It is found that the RSM-PRT system is capable of measuring
reflection In a pile with length up to SO pile diameters.

Factors Affecting Test Results and limitations

of The Test
Factors which may affect the interpretation of pile integrity testing results
" 3 G 9 12 15 18 21 2.... 27 em (m)
are briefly discussed here. Influencing factors, consequences, possible
corrective measures and suggested further studies are presented in
Figure 4 - Pile No.9 - (a) RSM results, (b) PIT results, and (e) TNO results Table 1.

TRANSACT/ONS • Volume 6 Number 1

 Factors Consequences Proposed Corrective measures
Incorrect hammering Generation of poor longitudinal waves. Striking perpendicular to the top surface of pile head.
Incorrect hammers Too small a hammer with a stiff head generates high Using an adequate hammer. If defects are in the shallow
frequency waves which attenuates too fast and cannot reach part of the pile, a small and very hard tipped hammer
the deeper part of the pile. Too large a hammer with a soft should be tried. The correct selection and use of
head generates a wave with a large contents of low hammers depend on the e)(perience and judgment of
frequencies and a large pulse period which may mi)( up with the operator.
reflections from small defects in shallow depth of the pile.
Soil resistance and (a) Abrupt change in soil stiffn ess may induce a reflected Differentiating between the secondary waves from the soil
damping wave. (b) Soil laye rs may induce secondary waves. (c) Soil 's layers and the reflections from pile toe and anomalies.
damping (radial and viscous) may reduce the strength of the
wave propagation in the pile by attenuation.
Wave velocity of pi les Wrongly assumed wave velocity lead s a wrong estimation of Calibrating the wave velocity on a pile (known length) of
assumed in analysis pile length (or defect location). the same material as the pile under testing.
Large diameter pile On a large diameter pile, there is a problem in generating plan e The wave length should be larger than the pile's diameter.
waves propagating longitudinally along the pile's a)(is. If the pile diameter is too large, use other alternative
methods.
Long pile The wave reflection may be too weak or there is no reflection (a) Increase the hammer impact energy. (b) Increase the
at all due to attenuation of the wave for a long pile. resolution of sampling and signal-to-no ise ratio of the
measurement. (c) Use other alternative methods.
Defects near the pile (a) Defects near the pile head may lead to disperSion of the (a) To overcome the effects of defects near the pile head
head and multiple initial impact wave and a distortion of echo wave. (b) Multi- and to detect defects in the deep part of the pile, try to use
defects reflections reduce the wave energy transmitted to the deep heavy and less stiff tipped hammer. (b) Experienced
part of the pile and lead to wave superposition. interpretation is needed. (c) Development Of more
sophisticated programs for modeling and interpretation.
Instrument Poor resolution cannot differentiate the reflections from two The instrument shall have adequate resolution (minimum
close reflectors. l ow signal-to-no ise ratio may not pick up 50(s) and signal-to-noise ratio and other advanced features
week signals from the toe of a long pile. such as signal filtering, auto-ranging, auto-averaging, and
auto-correlation.

Table 1 - Fadors affeding pife integrity testing results

limitations of th e pile integrity test ing may be summarized as the changes may Induce strong waves that confuse the interpretation. In this
followings: case, knowing soil layer conditions in advance will help the identification
and separation of waves Induced by the soil layering. Development of
(a) There is no detection or Identification of defects along the axis of
a sophisticated computer model and program Incorporating the effects
a pile;
of soli layering may assist In the Interpretation of test results, provided
(b) No direct Indication is given to variations of pile diameter;
soil properties and layer positions are correctly recorded .
(c) Defects in the deep part of a long pile may not be deteded;
(d) The test is poor in detecting a gradual change of pile cross-sections; In spite of the above limitations, pile integrity testing based on the stress
(e) The length and anomalies estimated from the test are acoustical length wave equation has merits in that it is (a) quick and economi cal and (b)
and acoustic anomalies, not all of which correspond to physical length reli able when used correctly and Interpreted by experienced personnel.
and structural anomalies; and
(f) A pile toe socketed or grouted in rock with a Similar impedance may Summary and Conclusions
not be identified. The stress wave equation and equations for the calculation of reflected
The limitations discussed above and the factors listed in Table 1 are and transm itted waves are derived for modelling pile Integrity testing.
based on the authors' knowledge and experience. TIle summary and An Idealized pile with a necking section Is analyzed to assist in the
the factors listed in the table are by no means exhaustive. interpretation of test data on real piles. It is found that the magnitudes
of the wave velocity refleded from the toe of a pile with defects Is smaller
The stress wave propagation In a pile may not be the assumed one-
than that for a pile without any defect.
dimensional (longitudinal) propagation, especially when the pile diameter
is too large and hammer striking is not perpendicular to the surface of Original resu lts from a field verification exercise in Beijing, China and a
pile head. Theoretica lly, a one-dimensional wave may be formed in a test on a prestressed concrete pile in Hong Kong are presented and
pile under the condition: 0 < A. and 0 < l , where 0 and l are pile diameter di scussed within the framework of stress wave theory. It Is found that
and length respectively. The parameter Als wave length and A ~f, where the instrument developed at Institute of Rock and 5011 Mechanics (lRSM)
c Is wave velocity and f is the basic frequency ranging from 10Hz to has similar capabilities as other instruments developed in U.S.A. and
300HL The condition of L > D < A ~ clf should be used to judge whether Holland. With these instruments and associated computer programs, pile
or not the pile diameter is too large for the assumption of a one- anomalies may be identified and the actual pile length may be checked
dimensional stress wave. The damping from the so il layers and the pile as a means for quality control. However, wave propagation In a real pile
may have a significant influence on test results and interpretation. The is more comp licated than that assumed In thi s stUdy. Factors which may
damping effects are normally Ignored In the analysis except that signals affect testing results and their Interpretation are summarized. Possible
are amplified. Resistance from SOil layers can generate waves which mix corrective measures and further studies are suggested. limitations of
with reflected wave from pile anomalies. This wi ll make interpretation the method are presented In th is paper. However. in spite of these
of test data difficult. Small stiffness changes in soils may generate weak limitations, the correct use of the pile integrity testing method and the
secondary waves which may be Identified and separated from major interpretation of results by experienced personnel can provide a quick
reflected waves caused by pile anomalies. However, abrupt soH stiffness and economic solution to the quality control of pile construction.

The Hong Kong Institution Of Engineers

Acknowledgement
.
Financial support (AIC No.A-P089) from the Hong Kong Polytechnic
University is acknowledged. cr,
S

References
~
r 1 ~
1. Ellway, K. (1987). Practical guidance on the use of integrity tests for the quality T
control of cast-in-situ piles. Ground Engineering, Vol.20, No.7, pp.8-13. cr,+(8cr/ax)dX
2. Forde, M,e (han, H.F., and Batchelor, AJ. (198S). Interpretation of non-destructive
tests on piles. Proceedings of the Second International Conference on Structural
Faults and Repair, London, pp.333-348.
3. GEO (1996). ' Plle Design and Construction ", GEO Publication NO.l /96.
4. Kolsky, H. (1953). ' Stress Waves in Solids", Oxford Press. x
5. Uu, M.G. (1995). The principle and design of transient, binary and floating
data acquisition system, In 'Techniques for the Measurement of Foundation Plies
and Construction Sile", edited by Uu, M.G., Cai, Z.L and She S.C., Hubei Science Figure A-1 - Pile and pile element
a Technology PUbliShing House. pp.259-265. (In Chinese)
6. Paquet, J. (1992) . Pile Integrity testing - the CEB ITP reflectogram. Proceedings
Of the Conference on Piling: European Pradice and Worldwide Trends, London, The general solution to Eqn. (A-21 Is
pp.l00-110.
7. PIT Instrument (1993). Goble Rausche Likins and Associates, Inc., Cleveland, u (x, tl ~ 'P(x - etl + $ (x + etl (A-31
Ohio, USA.
8. Smith, EAL. (1955). Impact and Longitudinal Wave Transmission. Trans ASME, where \fI and $ are to arbitrary functions of (x-ct) and (x+ct) respectively.
pp.963-973.
Eqn.(A~5) is known as d'Alembert's soiution. The Item \fI (x-ct) represents
9. Tharn, LG., Chu, C.K. and LeI, zx. (1996). AnalYSIS of the transient response
of vertically loaded single plles by tlme-domaln BEM, Computers and Geotechnics, a downward travelin g wave and q, (x+ct) represents upward traveling,
Vol.19, No.2, pp.117-136. both at a speed of c.
10. Tijou, J.c. (1984). Integrity and dynamic testing of deep foundations - recent
experiences in Hong Kong (1981-1983). Hong Kong Engineer. Vo1.12, No.9, The downward velocity of a mass point of the pile, denoted as vJ" can
pp.1S-22. be obtained from Eqn. (A-31
11 . Tlmoshenko, S.P. and Goodier, IN. (1970). Theory of Elasticity (3rd edn.). McGraw-
Hill, New York. v.l ~ "'P(x - dl ~ _ e'P. (A-41
12. TNO Instrument (1993). The Institute TNO for Building Materials and BuHdlng at
Structures, The Netherlands. The downward force (or stress) wave is
13. Weltman, A1. (1977). Integrity Testing of Piles: A Review. ((tRIA Report NO.PG4.)
Construction Industry Research 8' Information Associatlon, London, 68p. rl'P(x - etl
14. Yuan, lX. (1988a). On the Problems of Dynamic Pile Testing. Journal oj Rock P .l ~dAE~I- IAE~-'P ' AE (A-5)
ax
and Soil Mechanics, VoLl. No.2, pp.2-18. (In Chinese)
15. Yuan, Jx. (1988b). Non-destructive Testing Methods for Piles. Journal oj Rock Using Eqn. (A-41, Eqn. IA-5) can be written
and Soil Mechanics. voU, No.], pp.83-94. (In Chinese)
P.l~Zv.l (A-61

where Z = AE/c, the impedance of the pile.

Using the same method as for Eqn. (A-4) to (A-6) for the upward wave
Appendix 1 pi~-Zvi 1A-7)
Wave Equation, Wave Reflection and Wave From (A-51, the total velocity is
Transmission
au o'P(x - ct) a~(x + ct)
v lx, t) ~ at'~ ilt + at ~v.l+vi (A-B)
(a) Wave equation, velocity wave and force wave
The total force is
Referring to Fig.A-l, according to Newton's second law, the stress wave
equation can be derived as au a'P(x - etl + a~(x + ct) ~ P.l + pi
P(x ,tl~ -AE at - (A-g)
ax ax
a'u
(A-l I
at' (bl Free end and fixed end of pile
where u = displacement, positive for x-direction; t = time; c = "(E/p),
When the toe of the pile is Ideally free, the totai force at the toe shall
the speed of wave propagation; E = Young's modu lus; p = density of
be zero. Therefore
pile; x = coordinate along pile axis; r = fridion force per unit pile length
from soil; A = cross-section area of pile. In Eqn. (A-1), for simplicity,
P (x, tl ~ P.l + pi ~ 0 (A-lOl
damping is not considered.
pi ~ -p .l
For low-strain testing, since the Young's modulus E of the pile is several
orders higher than 5011'5 Young's modulus, the wave propagation is mostly
Substituting Eqn. (A-6) and Eqn. IA-7) into Eqn. (A-lOl leads to
traveling along the pile and is less affeded by soli's friction on the pile,
compared to the high-strain dynamic testing. In most low-strain analyses vi ~ v.l IA-lli
and for simplicity, the friction force, r, in Eqn. (A-l ) Is considered zero. From Eqn . lA-B) and using 1A-1l)
Assuming r=O, which means no soil resistance considered, Eqn. (A~1)
1A-12)
becomes
a'u Eqn. (A~lO) means that a downward force (or stress) wave becomes a
(A-21 reflected wave at the free end of the pile. The reflected wave is of the
at'
same magnitude but opposite sign (changing the phasel. If the downward

TRANSACTIONS. Volume 6 Number 7

force wave is compressive (considered positive, say 100kN/sec), the P, I p,l P,I p, l
reflected force wave will be tensile force wave (negative, - lOOkN/ sec). + = +
I1 Zl Z2 I1
Eqn. (A-l1) means that a downward velocity (say 20m/s) wave is reflected
at the free end with the same magnitude (20m/s) and the same phase Solving the two equations in (A-17) for two unknowns Pli and P),
(also downward). The total velocity at the pile toe is doubled (40m/ s) 2Z ,
as shown in Eqn. (A-12). The duration (time-history) of the doubled velocity ,
PI~ P,t + Z,+ Z, P,i
is dependent on the time-history of downward velocity wave. (A- 18)
If the toe of the pile is ideally fixed , the displacement at the toe shall p) = Z2Z2Z P1J- + ~1 -~2 P21
2' + 1 1 +1
be zero. This means that the velocity at the toe shall be zero. From Eqn.
(A-B) The first equation In Eqn. (A-18) shows that the upward wave P, I is equal
v ~ VI + v1 ~ 0 (A-l3) to the reflected upward wave due to the downward wave P plUS the J
transmitted upward wave due to the upward wave P21. The second
vi ~ - v1 equation in Eqn. (A-1S) shows that the downward wave P2i is equal to
the transmitted upward wave due to the downward wave PI t pIUS the
substituting Eqn. (A-6) and (A-?) into (A-B) leads to reflected transmitted downward wave due to the upward wave P2 J..
PI ~ p1 (A-14) If there is only a downward force wave P, .1., from Eqn. (A-1S), the reflected
And from Eqn. (A-9), using Eqn. (A-14) upward wave and transmitted downward wave are

P ~ PI + p1 ~ 2PI (A-1S) Z2 - Z1
,
P I~
Z2 + Z1
p,l
Eqn. (A-13) means that a downward velocity wave (say compressive and (A-19)
2Z,
considered positive in x-direction) is reflected at the pile toe and becomes P, t
a reflected (negative) wave of the same magnitude. The total velocity 22 + Z1
at the toe is zero. Eqn. (A-14) means that the downward fo rce (say If there is only upward force wave p), from Eqn . (A-1S)
compressive) wave become an upward wave of the same magnitude in
2Z,
the same phase (also compressive). The total force at the pile toe is doubled ,
P I ~
I 2 + Zl
P,i
as in Eqn. (A-lS). The duration (time-history) of the doubled force wave
(A-20)
is dependent on the time-history of downward force wave.
Z1 - Z2
,
p t~
Z2 + Z1
P,I
(c) The change in piTe impedance

Consider the change of the impedance of a pile at a cross-section as Using Eqn. (A-6) and (A-?), Eqn. (A- 18) can be written in terms of velocities,
shown in Fig. (A-2). The impedance in the upper part is Z1 = (Al E )/C1 = 2Z,
A1-..J (E 1 pJ The impedance in the lower part, I 2 = (A 2 E)lcz = A)(E 2P2)' v,I ~ - v,J, + Z2 + I , v,i
Th e change in impedance may be due to the change in section area A, (A-21)
Young's modulus E, andl or the wave velocity c (or densily, p, of pile). 2Z, 2, - 22
The continuity conditions at the section in Fig. (A-2) are that (a) the total
I , + I , v 1J.- Z1 + Z1 v,1
forces on the both sides of the sedion shall be equal and (b) the velocities If there is only a downward wave, from Eqn .(A-21)
on the both sides shall be equal , that is,
VI 1= - Zz - Z, v t
P,I + p,l~ P,I + p,l (A-16) I 2 + Z1 1
(A-22)
v1i+ v 1J. = v) + v)

substituting Eqn. (A-6) and (A-?) into the second equation In (A· 16), and
or only an upward wave
rearranging Eqn. (A-16). the following is obtained
2Z, I
(A-1?) V11 = Z~Z
,. , v,
(A-23)

Eqn.(A-18) to Eqn.(A-23) can be used to study the propagation and

ZI superposition of waves and to identify the position of section changes.

Appendix 2
CalCUlation of Wave Reflection and Wave
TI'ansmission
Z,
Referring to Fig.l, the wave reflection and transmission in a pile with a
change in impedance are calculated and summarized in this appendix.
For an initial narrow rectangular compressive (N+") force pulse of magnitude
x P, the corresponding velocity is v = P/ ZA' In the pile section with an
impedance ZA and a length LA' The impedance, Ze in the pile section
of length Lc' are assumed to be Zc = ZA' The pile impedance ZB in the
Figure A-2 - Wave reflection and transmission at the pile section Of pile section of length LB are Zn = IA/ 4. Since Z = AE/ c = A-..J(Ep), the
Impedance change smaller I s may be due to (a) a smaller cross-section area and/or (b) the

The Hong Kong Institution Of Engineers

weak pile material (the smaller Ep). The numbering for different reflected The above velocity arrives at the pUe head at time t4 and is doubled
waves and transmitted waves is shown in Fig.1. The first reflected waves as
from Section A-S are calculated in detail below.
vtop = 2 x (-0 • 230v) = - 0460v
. at pIle head at time t4
No. (1): Using Eqn. (A-19) for a downward wave only, noting Z/Z, = Z/ which is plotted o n Fig.1(fj.
Zs = 4 , the reflected force wave is
Other reflected and transmitted waves can be calculated using the
Z -Z 1-4
P1t = Z2 ZI P1 ..t. = 1 4 P = - 0.6 pi passing Section k-k procedure above. As the waves are transmitted and reflected many times,
l + ! + the magnitudes of the force and velocity passing Section k<k and arriving
The reflected velocity wave is, using Eqn. (A-22) at the pile head are getting smaller and smaller.

v 1 1' = -
Z -Z
z: + z~ v 1 J., = -
1-4
1 +4 v = 0.6 vi passing Section k-k

The above force wave and all other force waves passing Section k-k
upward should be zero at the pile head as discussed in Appendix 1. Jlan-hua Vln
But all velocity waves should be doubled which may be measured at
Jlan"hua Yin was awarded a BEng from Chongqing
the pile head. tn this case,
Institute of Communications and Chongqing Institute
v10p = 2 x O.6v = 1.2v at pile head at time t1 of Architecture and Engineering, China 1983, a MSc
from the Institute of Rock and Soil Mechanics, the
which is plotted in Fig.l (fj. Chinese Academy of Sclences In 1984 and a PhD from
the University of Manitoba, Canada 1990. He worked
The calculation of force and velocity waves passing Section k-k and the
as an engineer in a major consulting firm in Halifax.
velocity at the pile head are summarized below.
Nova Scotia, (1990 to 1992), in C-CORE, St John 's
No. (4): The transmitted waves due to upward waves in No.(3) through Newfoundland, Canada (1993 to 1994) and in Binnie
Section A-B, are Consultants Ltd In Hong Kong (1994 to 1995). He joined
the Department of Civil and Strudural Engineering. the Hong Kong pOlytechnic
2x4 University in 1995 and is now an Associate Professor In Geotechnical Engineering.
(0.24P) = 0.384PI passing Section k-k His research Interests include (a) soft sot! behavior and improvement, (b) laboratory
1+4
testing study and constitutive modeling of soil behaVior, (c) geotechnical numerical
2Z2 2 analyses, (d) soil/strudural/geo-syntheti cs interaction and modeling and (e) 3-D
V,I = ""z,-+"zc- V,I = 1 + 4 (-0.96v) = -0.384vI passing Section k-k
, , slope stability modeling.

The velocity wave passing Section k-k will be doubled at the pile head.
v top = 2 x (-0.3 B4v) = - 0.7B6v at pile head at time t2
No. (8): Transmitted waves through Section A-S due to upward waves
in No. (7) are Jlan-xln Yuan
Mr Jlan-Xin Yuan is a professor in the Institute of Rock
2Z
P t = -oc-_ ""c- P 1 = 2 x 4 (-0.256P) = - 0.410pi and Soil MechaniCS, the Chinese Academy of Sciences
, Z+Z , ' ,
4+1 (former Director of the Institute). Professor Yuan was
passing Section k-k the Chairman of 9th InternatIonal Conference of the
International ASSOciation for Computer Methods and
2Z, 2
= ""z- Advances in Geomechanlcs held In Wuhan, China, Nov
V,I
, +'=z'--, v,i = 4 + 1 (1.024v) = 0.41OvI passing Section k-k
2-7, 1997. He holds scholarly positions as a Member
in Editorial Board, International Journal of Computers
The above velocity wave is the wave reflected trom the toe of the pile,
and Geotechnics, Director, Board of Directors,
arriving at the pile head the first time at time ts' The above velocity International Association for Computer Methods and
should be doubled at the pile head. Advances in Geomechanlcs, USA, Member of Editorial Board, Ada Mechanics Salida
Sl nl ca, Wuhan, China, Vice-chairman of Editorial Board, Chinese Journal of
v top = 2 x 0.4096v = 0.B20v at pile head at time t ,~
Geotechnical Engineering, Nanjing-Wuhan, China.
No. (11 ): Transmitted waves through Section A-S due to upward waves
in No. PO} are
2Z1 2x4
P, 1 = -oc---'oc-P,i = -"4'-c+'--1;-(0.0864P) = 0.138Pt
Z2 + ZI
passing Section k-k Ming-gul Liu
2Z, i 2 Mr M!ng-Gul lIu obtained a MSc in Automation and
v, i = - -oc---'=c- v, = -,------,- (-0.3456v) = -0.138vI Electronic from the Central China polytechnic University.
Z2 + ZI 4 +1
He is Chief Engineer and leader of Division No 4 of
passing Section k-k
the Institute of Rock and Soli Mechanics, the Ch inese
These are the waves which are reflected twice in the layer LA' The above Academy of Sciences. He hods scholarly positions as
velocity arrives at the pile head at time t3 and is doubled. a Committee Member of the Instrument and ObselVation
Division of the Chinese Society of GeophYSics,
v,,,, = 2 x (-0.138v) = - 0.276v at pile head at time t, Comm ittee Member of the Acoustic Measurement
Division of the Chinese Society of Acoustics, and
No. (13}: The reflected waves from Section A-B due to downward waves
President of the Association of Acoustics of Hubei
in No. (12) are
Province, China.
1 - 4
,
PI = P,J. = 1 + 4 (-0.384P) = 0.230Pi
passing Section k-k
1- 4
,
VJ.=- 1 + 4 (-0.384v) = -0.230vI
passing Section k-k

TRANSACTIONS· Volume 6 Number 1