INTERNATIONAL SOCIETY FOR

SOIL MECHANICS AND

GEOTECHNICAL ENGINEERING

This paper was downloaded from the Online Library of

the International Society for Soil Mechanics and
Geotechnical Engineering (ISSMGE). The library is
available here:

https://www.issmge.org/publications/online-library

This is an open-access database that archives thousands

of papers published under the Auspices of the ISSMGE and
maintained by the Innovation and Development
Committee of ISSMGE.
 8/2

Interpretation of Axial Load Tests on Long Piles

Longs Pieux: Interpretation des Essais de Charge

M. APPENDINO ENEL-CIE, Torino, Italy

SYNOPSIS The paper describes some results deriving from load tests on long high-capaci­
ty end bearing driven piles. Piles consist of a concrete core cast in a closed eno pipe.
Interpretation is complex owing to concrete non-linearity and interaction with steel casing. In
addition, unavoidable limitations as to data precision come out. Instrumented piles snow that
failure at tip occurs suddenly, therefore the ultimate resistance cannot be anticipated by
extrapolation of load v/s settlement curve. Not instrumented piles do not allow the determination
of the transferred load due to the variable intensity and distribution of shaft resistance.

1. Soil properties and piles design layer of dry concrete. Sermide soil consists
of soft N.C. silty clay underlayed at 28 m
Porto Tolle site (North Italy) consists of
depth by medium dense sand. The pile tested
8-10 m of silty sand, underlayed by soft N.C.
was indentical except for a lower length and a
silty clay, containing a 5-7 m thick layer of
larger penetration into the sand.
silty sand, and medium dense sand at 37-45 m
depth, figures 1,2,3. Pile design load is 1.2
2. Load test program
MN. No consideration was given to shaft resi­
stance from superficial soft soil, therefore Tests on Porto Toile were performed using a
the design bearing capacity had to be provided slew maintained load incremental procedure,
exclusively by tip resistance. A penetration with several loac-unloac cycles. Maximum appli­
of 2-3 m in the sand was rec.uired to satisfy ed load was 6.0 MN. Reaction was provided by a
this condition. Outside pile diameter is .508 weighted platform. Loac on pile butt was deter­
m ; pipe wall thickness ranges from 5.5 mm at mined from the pressure in the jack by a
top to 9.9 mm at tip. Concrete core average calibration curve obtained experimentally.
cubic strength and elastic modulus are 4.1 Mpa Butt settlements and lateral displacements
and 35,000 Mpa at 28 days. The concrete core were measured continuously by three electrical
did not rest directly on the plate closed pipe displacement transducers. Table 1 summarizes
tip owing to the existance of a .5 m thick the main data related to all Porto Tolle
CPT rp (M Pa) (cm ) Av. st rain s (£xlo®)
10 20 30 40 50

Pile axis
v e r t ica lly

St r ain s
Fig. 1 - PT 868: Strains and loads v/s depth. Lo ad s

38 - 017111 593
 r p (M P a) (c m )
8 /2 10 2 0 30 4 0 50

P ile a x is
v e r tic a lly
te s t

-------- S t r a in s
-------- L o a d s

Fig. 2 - PT 361: Strains and loads v/s depth.

CPT rp (M p a ) (E»io )
10 20 30

CLAYEY
S ILT

FIN E
SAND
W IT H
S ILT

S ILT
W IT H
FINE
SANO

Fig. 3 - Sermide: Strains and loads v/s depth.

tests. Sermide pile was initially tested in a couples of telltales, diametrically opposed.
fast monotonio procedure and later retested All instruments were read continuously by a
according to the same slow procedure as in monitoring system throughout the test, while
Porto Tolle; finally four quick tests up to manual readings were taken before and after
failure were performed to produce an almost the test.
undrained loading condition.
4. Load-gross settlement curves
3. Pile instrumentation
Figures 4 and 5 show load v/s gross settlement
PT 868, PT 51 and Sermide piles were instru­ curves from PT 868 and PT 361 instrumented
mented with diametrically opposed couples of - piles. Accidentally, they represent well the
complete bridge, six lines connection - strain- set of results from all piles tested. Fig. 6
gages, attached on steel bars (two 90° apart shows results from Sermide tests.
couples in PT 068). This type of gages has These curves and Tab. 1 data indicate that
been largely used and tested, proving to be failure occurs suddenly, no sign of it appear­
precise and stable. PT 361 pile was instrument­ ing in advance. Fig. 7 shows piles top re­
ed with recently developed electrical extenso- bounds from 5.0 MN load stage, which is very
meters directly embedded in the concrete. PT close to the ultimate resistance of failed
868 and Sermide were also instrumented with piles. Pile top rebound is the effect both of

594
 8/2

TABLE 1 - SUMMARY OF LOAD TESTS rN PORTO TOLLE

1 Original | Soil 1 Gross Settlement Gross Residual 1

PILE Soi 1 | Properties | at (MN) 1 Settlement i
1
1 Properties| on driving | from (MN)
(1) C.P.T. v | (4)| 1 i
No. Depth Instr. D.R. .G.E. 3.75 1
! e ™
5.00 ' 6.00 1
(2) 3.75 , 5.00 6.00
rp | i !
I(-) m 1 (-) Mpa | b.p.cm. 1 mm 1 mm ¡ mm J mm 1 mm [ mm
1 H 42.0 1 - | .6 1 A 1 13.6 1 1 - 1 2.1 ! - i i
1 2 40.0 1 150 | .6 1 A 1 13.6 1 1 - 1 2.5 1 - ! -
1540 41 .0 1 130 | .8 1 A/B 1 11.5 1 1 - i 2.0 1 - i
1307 41.0 1 110 | .6 1 A/B 1 11.0 1.7
! " (7) ! I i (7) :
1361 42.0 1 C.E. 190 | 1.2 1 B/C 1 14.6 ! 25.¡(6> 1 41.7 1 2.4 1 6.1 1 17.0 i
1 51 42.0 1 S.E. - t 1.3 1 B/C 1 1 1 - 1 I : 1
i -
5 ■1 fS )
1868 41.0 1S.E.+TT 160 | 1.5 1 c 1 Ì o *!(5) 1 13.7 ! 25.2 i 2.4 1 3.9 j 5.8 i
(5) J 7
1160 42.0 1 165 | A/B 1 1 24.9 134.2 x 1 ¡ 6.9 !
■8(3) [ 1^ ‘°(5) 7 2 -3 d ! 113-:m 1
1 30 43.1 i - I A/B 1 1 20.1 i 43.7 ; 1 3-3 ! 21.1 !
■5(3) 1 *4 (5 )
1100 43.0 1 150 | A/B 1 1 16.2 ! 22.6 ! 2.0 1 2.4 ! 4.9 1
1351 45.6 1 | i'” 150 B 1 ! 16.5 1 4.1 I
22*9 (7) 6 '6 (7) j
1320 41.0 1 140 | C 1 1 29.2 ! 86.o ! 4.8 ! 12.4 63.3 1
(1) C.E.=Concrete extensometer / S.E.=Extensometer on steeibar / TT=Telltale.
(2) Average of 4 diameters over and 1 diameter below pile tip;
(3) Corrected value for lower hammer efficiency. (4) Group effect: A=single pi le; 3=Pi le a smal i group;
C=pile in a dense group. (5) at or from 4.0 MH. (6) at or from 5.5 MN'. (7) Fai lure between 5.C and 6. 0
Load (M N )
Load (M N ) 1.0 2.0 3.0 4.0 5.0

Fig. 4 - PT 868: Loads v/s gross settlements.

L o a d (M N )

tio rebound and of pile shortening. Analysis

of compaction and compression variations i,' )
on a Porto Tolle similar sane stated, Appendi-
no (1980), that pile tip displacements are
almost constant at failure. In fact, a lower
deformation, corresponding to a more resistant
soil compensates for a larger load. This was

Fig. 5 - PT 361: Loads v/s gross settlements. (') From soil original heterogeneity or driv­
ing effect.

595
 8/2

1. the average of two diametrically opposed

AZ gages is not constant;
(m m )
2. calibration curves are not linear and do
Z 20 P ile s fa ile d
not coincide, the scatter growing with the
5 361 y e s > b e t w e e n 5.5 load;
o a n d 6.0 M N
in %
E 160 J to ta l lo a d C a lib r a tio n c u r v e
320 80
8 • • -S e p a r a tio n lin e
TJ
§ 15 P iles n o t fa ile d
0 âs 100
no > a t 6 0 0 t
jQ
£ 351 to t a l lo ad
O
0)
r
10
£ 40 42 44 46 (m ) Av. load. rate(M N |h )
P ile le n g th P ile S e rm . 15
O N o t Instr. pile 0 In s tr u m e n te d pile RT. 8 6 8 .21
RT. 51 .22
Fig. Piles top rebounds from 5.0
RT. 361 .12
verified by field driving data and instrument­
ed piles (point 8). Consequently lower re­
200 400 600 800 1000 1200 1400 1600 1800 2 000
bounds in Fig.7 indicate that a larger frac­
S t r a in s ( £ » 106)
tion of total load is resisted by superficial
soil layers, as in PT 868 case. This is why jOads v/s strains at top levels.
piles having lower rebound did not fail. It A v. s t r a in s (E *1 0 6 )
derives that not-instrumented piles tests do 20 40 60 80 100 120 140

not point out variations of tip ultimate resi­

stance, but rather of distribution and inten­
sity of shaft resistance.

5. Strains analysis \. I S tr a in s
\ »
Average strains from every instrumented level
are shown in Figures 1, 2, 3 as a function of
depth. Table 2 shows separately average
strains and deviations from both couples of
gages in PT 860 pile at three levels. Fig. 8
shows the applied load as a function of
strains from all instrumented levels at piles
top (calibration curves). Finally strains
after unloading from Sermide top level are
■5 1.0 Q (M N )
plotted in Fig. 9. These and Table 2 data show
that: Fig. 9 - Sermide: After-unloading strains.

TABLE 2 - PT 868 - STRAINS CORRESPONDING TO EVERY LOAD - UNLOAD CYCLE

1 1
6 6 6
1 Q (MN) £ x 10 1 3 x 10 - x 10 1
1 5
1 1
0 î 0 0 0 0 j 0
1 o 1 !
! (1) (2) ! (1) (2) 1 (1) (2) 1 (1) (2) j (1) (2) Ì (1) (2) 1
1 .75 78 + 56 ! 111 + 39 1 26 + - 1 25 + 3 1 6 + - 1 6 + - 1
1 o 1 + 13 I 11 + 17 1 4 + - 1 4 + - 1 1 + - 1 2 + - 1
Î 1.50 168 + 72 j 239 + 11 1 72 + 3 1 70 + 4 1 23 + - 1 21 + - 1
! o 2 + 9 1 21 + 10 1 13 + - j 12 + 3 1 7 + - 1 8 + - 1
Ì 2.25 284 + 68 ! 436 + 27 1 143 + 4 1 139 + 7 1 50 + - 1 55 + - 1
1 o + 1- + 3 1 68 + 1 1 31 + - 1 28 + 6 1 20 + 3 1 25 + - 1
1 3.00 441 + 103 1 696 + 49 1 221 + 6 1 215 + 13 1 87 + - 1 99 + 7 1
1 o 11 + 6 1 134 + 7 1 45 + - 1 41 + 8 1 37 + - 1 43 + - 1
1 4.00 708 + 188 j 1156 + 90 1 349 + 8 1 338 + 22 1 160 + 5 1 186 + 13 1
! o 10 + 16 1 189 + 12 1 61 + - 1 54 + 14 1 66 + - 1 79 + 5 1
1 5.00 991 + 251 1 1674 + 100 1 500 + 9 1 482 + 35 1 265 + 14 1 317 + 21 1
1 o 4 + 46 1 334 + 163 1 92 + - 1 81 + 20 1 112 + 5 1 137 + 8 1
1 6.00 1084 +
1 2284 + 250 132
1 663 + 8 I 638 + 49 1 396 + 22 1 507 + 25 1
1 o 41 90 1 329 * 260 1 114 + - 1 100 + 29 1 156 + 7 1 215 + 19 1
(1) average of two diametrically opposed instruments - (2) deviation.

596
 8/2

3. average strains measured by PT 868 two 90° on unavoidable and undeterminable causes,
apart couples of gages do not coincide; such as incorrect jack positioning, misaline-
4. residual strains - growing with load - ap­ ment of force and pile axis and pile curvatu­
pear after unloading; they are generally re. Load v/s depth curves were also connected
positive (compression), but may become nega­ in order to fit with soil stratigraphy; pile
tive (traction) at higher loading stages; tip load was obtained by linear extrapolation.
5. residual strains at pile top level cannot P.T. 361 strains elaboration was limited to
be related to any external load; they par­ lower levels on account of large positive resi­
tially recover with time converging, 24 dual strains in upper levels appearing as the
hours after load removal, to a common va­ load exceeds 2.5 MN, Figures 1,2,3.
lue, independent on previous load intensity
and duration; 7. Residual loads
6. after-unloading residual strains grow with
pile depth. Fig. 9 shows that 24 hours after-unloading
This behaviour is common to all instrumented residual strains at pile top converge to a
piles, except PT 361, where positive residual unique value independent on previous loading
strains begin to appear as the load exceeds intensity and duration. This finding enables
2,5 MN. It is not possible to establish to separate external load dependent strains
whether they are the consequence of incorrect from those due to creep and internal stresses
extensometers performance or whether represent in lower pile levels. Service pile residual
residual concrete tensile strains. After­ load v/s depth curves are shown in Fig. 10 at
unloading strains at pile top cannot be relat­ different times from unloading. A residual
ed to any external load, they consequently load of .5 MN, increasing slightly with time,
must correspond to creep phenomena and pile is applied at its tip (circled numbers).
internal stresses. After-unloading strains in Q u ic k lo a d in g ( 3 d c l.)
: xio
lower instrumented levels may be the result of Q u ic k lo a d in g (4t h c l.)
200
the same effects in addition to residual load
(3.0 MISTl (5.0 M N )
from skin friction. Present understanding of
150 U / M a n t . lo a d in g
experimental data allows to state that:
1. concrete non-linear deformability prevents
100
elimination of force excentricity by averag­
ing two diametrically opposed gages; 50-
2. concrete non-linear deformability causes
also calibration curves nonlinearity; their
curvature being very sensitive to small 10 " ’ 103
tim e ( m in )
load excentricity;
3. Residual strains are partly originated by Fig. 10 - Sermide: After-unloading strains at
concrete creep and complex interaction bet­ pile top v/s time.
ween it and pipe (slippage plus pile bend­
ing?) ; 8. Shaft load transfer curves
4. residual loads are generated when the load
is removed. Only P.T. 868 shaft load transfer curves are
shown, Figures 11, 12, 13. Only a few comments
are allowed for limitation of space. All
6. Load v/s depth distribution
curves are qualitatively good but doubtless
Load v/s strains curves from t o D ?tvel were they may be affected by errors from load deter­
used to determine loads applied at lower le­ mination, in levels close to top. This is
vels. To do this, scatter from load excentrici­ clear from Fig. 12 where shaft resistance is
ty was eliminated using the external envelope much smaller than undisturbed soil resistance,
in Fig. 8. Non linearity and pile cross sec­
tion variability were considered assuming the
load at level i proportional to the homoge­
neous pile cross section. It follows that

(1) Q, - Qb ^
I ,b
where Q. is the load on section i, l , Q. is the
i o
load from calibration curve corresponding
resi to
strains £. in section l, AT . and A. A are
i I .i .I.b
respectively pile homogeneous section in i and
at the pile section b where the calibration
curve was obtained. The load is exactly de­
termined only if force in section i is cente­
red, otherwise it is affected by a positive
error. This error cannot be eliminated and its
magnitude cannot be estimated as it depends Fig. 11 PT 868 Shaft load transfer curve.

597
8/2

although soil loss of resistance from driving ultimate tip resistance is slightly larger
- not entirely recovered - might also lower than PT 361, Appendino (1980). However, if
shaft resistance. Fig. 11 shows load transfer pile butt settlements were only available,
curve at upper pile segment. A very high hori­ only a larger shaft resistance on upper pile
could have been established, figures 1 and 7.
zontal pressure in the superficial silty sand
layer must be admitted to explain the high 9, .
skin resistance. This finding must be accepted
K -3 .0 9 M N /c m
because errors implied in determining load
transfer can only lead to underevaluate it.
(6.0 )
The large skin resistance can be the consequen­
ce of heavy compaction from driving. This was
proved by vertical soil deformations observed,
Appendino et al (1977); PT 868 in particular
was driven when all surrounding pipes had al­
ready been driven. A much lower skin resistan­
ce was measured in PT 361 upper pile segment,
figures 1 and 2. These results prevent the use X K = 2.47 M N /c m
of procedures based on a known shaft resistan­
ce to determine load transfer from not instru­
( ) lo a d o n p ile to p
mented piles.
1.6 X d
RT. 8 6 8
( D = p ll e d ia m e t e r )

.013= - 2 6 .0 m d e p th Z (m m )
tip d is p la c e m e n ts
(D T m a x f ° ,s -1 4 .5 m d e p th 'ig. 14 PT 868-361 : Tip load transfer curves.
\ Und. D ist.

7 m ax .015 .015 10 . Comments on load tests interpretation

© Cu .035 .017

^m ax .013 .013
The failure of end bearing piles driven in
Cu .060 .025 sand occurs suddenly and consequently the
ultimate resistance cannot be anticipated from
S ilty c la y
( D - p ile d ia m e t e r ) extrapolation of load v/s settlement curves.
-f- Interpretation of load tests on end bearing
10 Z (m m ) high-capacity long piles may lead to erroneous
results when piles are not instrumented, even
Fig. 12 - PT 868: Shaft load transfer curve. if the ultimate gross resistance is obtained.
This derives from the large influence of varia­
T
tions of skin unit resistance and distribu­
(M P a ) tion. After-unloading residual loads were ob­
served. Strains from concrete creep are to be
° PT. 8 6 8 -3 8 .5 m d e p th eliminated for their correct determination.
• RT. 8 6 8 - 3 3 .5 m d e p th
Sermide pile test indicates that a fast monoto­
200 nie loading procedure minimizes interpretation
S a n d a n d s ilty s a n d
V difficulties. These difficulties and the cost
u I
§ T m a x =-15 of high intensity load tests as well as of
» '50 kJ m a x _
piles instrumentation encourage efforts in de­
(A
su K £Tv tg(5 75 veloping alternative procedures for determin­
£ . 100
<5= 3 0* ing pile tip ultimate resistance such as those
(0 based on driving measurements.
£
1%D
g, .050 11. Note
s?
s Tests were carried on by Enel CTN - Milano,
$ o Italy.
10 (mm)
D is p la c e m e n ts

Fig. 13 - PT 868: Shaft load transfer curve. 12. References

- Appendino, M., Bogetti, F., Jamiolkowski,
9. Pile tip load transfer curves
M.B. (1977) - Geotechnical problems in Porto
Load transfer curves are shown in fig. 14. It Tolle. Power Plant design (in Italian),
is interesting to note the linearity of pile Enel, DSR, Q.V., No. 294, June, pp. 28+48.
tip settlement up to the ultimate resistance,
- Appendino, M. (1980) - Prediction of static
reached at 1.6 diameters penetration. Analo­
ultimate resistance from driving data, Se­
gous result was obtained in Sermide. This beha­
minar on appi. of stress-wave theory on
viour is typical of an elasto-plastic mate­
rial. Soil stiffness is sligthly higher below piles, 4-5 June, Stockholm.
pile 868 tip. This allows to suppose that its

598