DEVICE AND PROCEDURE FOR J _J OADING TESTS

ROYAL S \\" EDISH

GEOTEClfNlCAL INSTITUTE

PROCEEDJNGS

~\'. o. 3

DEVICE AND PROCEDURE FOR J _J OADING TESTS

ON PILES

\X'. t-:JELL\IA~ and Y. LJL.IEDAirL

S TOCKTIOL:\I 19 :31

lvar Ilceggstrii111s

llOKTltYCKEIII A.TI, STOCKllOJ,'.\I l!J,il

Contents

· Preface . . .. . . .. . .. . . .. . . . .. . . .. .. .. .. .. .. .. . . .. .. .. .. .. . . .. .. .. .. . . 5

§ 1. Importance of Loading Tests on Piles . . . 7

§ 2. Various Types of Loading Devices . . . 8

§ 3. Old Swedish Loading Device . . . O § 4. The SGI Loading Device . . . 9

§ 4a. Requirements for Design . . . n § 4 b. Force-Transmitting System . . . 15

§ 4c. Connection to Anchor Piles . . . 16

§ 4d. Hydraulic .Tack with Pump, T\Iotor, and Load Maintainer . . 22

§ 4 e. l\feasuring lVIovements of Test Pile . . . 23

§ 5. Statics of Loading Device . . . 24

§ 5 a. General Considerations . . . 24

§ 5 b. Case I . . . 24

§ 5 c. Case II . . . 26

§ 5 d. Case III ... 26

§ 5 e. Case IV ... 26

§ 5 f. Case V ... 27

§ 5 g. Other cases . . . 27

§ 6. Loading Procedure . . . 28

§ 6 a. Procedure A ... 28

§ 6 b. Procedure B ... 28

§ 6 c. Procedure C

...

§ 7. Present Applications . . . 33

§ 8. Summary . . . 34

Bibliography . . . 36 3

Preface

The loading device described in this report has been developed and used b?

the Geotcehnical Section of the Royal Swedish Board of Roads and Waterways and its successor since 1944, the undersigned Institute. The work has been done mainly by Mr W. Kjellman and j\fr Y. Liljedahl, and partly by Mr B. ,Jakobson and Mr I<. Hillberg, members of the staff of the Section and the Institute.

The test procedure hitherto used by the Section and the Institute, designated as Procedure B in the report, had been developed by the Bridge Department, under Mr Rudolf Kolm, of the above Board, before the Section ,rns established in 1936.

The loading device and the test procedure are rather different from those used in other countries, and they seem advantageous in certain respects. Therefore, it was deemed appropriate to make them known to the international public.

The report was prepared by Mr Kjellman.

Stockholm, January 1951.

ROYAL S,vEDISH GEOTECHNICAL lNSTI'l'UTE

5

Device and Procedure for Loading Tests on Piles

§ 1. Importance of Loading Tests on Piles.

The bearing capacity of a point-bearing pile is usually computed fr01n its resistance to driving by means of some pile formula. liowever, several authors (1, p 142; 2, p 660)' haYc shown that the result obtained in this way must be regarded merely as a rough estimate. The onl;v 1ncthod of finding the bearing capacity fairly accurately is to make a fulJ-size pile loading test on the site.

Friction piles surrounded mainly by cohr-sionless soils arc rare. As to their bearing capacity, the statement made above regarding point-bearing piles holds true here also.

The bearing capacit:v of a friction pile surrounded mainly by cohesive soils² is usually computed on the basis of the shear strength of these soils as determined by laboratory tests on so-calJcd undisturbed samples. However, the soil mass close to the pile is remoulded during the driving so that it loses a great deal of its strength, and this strength is restored only partly by thixotro)>y during the subsequent weeks. On the other hand, the driving of the piles increases the total pressure in the soil surrounding the piles. Both the remoulding and the increase of total pressure cause an increase of the pore water pressure, so that a consoli­

dation process is initiated, and the shearing strength is increased again (2, p 650; 3). The final result af alJ these phenomena can be estimated only roughly on the basis of the shear strength of the ground in an undisturbed statc:^1• A loading test on the pile is the only n1cans to get a more dependable Yaluc for its bearing capacity.

Thus, the loading test is the only way to get a fairly accurate value for the bearing capacity of any kind of pile. It is true that the bearing capacity of a pile in a pile group is not always equal to that of a single pile of the same dimensions in the same soil, but, at any rate, the former can be estimated only on the basis of the latter. Therefore, when designing a pile foundation, a lower

1 The numbers in parentheses refer lo the bibliography at the end of this report.

2 Such a pile should rightly be ca1led a cohesion pile or an adhesion pile, but lhc traditional designation is retained in this report.

3 It should also be observed that, according to Swedish experience, the shear strength of the ground in an undisturbed state can be determined only roughly by sampling and laboratory testing, the reason being that the samples normaily lose a great deal of their strength while being withdrawn from the ground (4). A more accurate dete1;nination can be made with a vane borer (4),

7

safety factor can be used if a loading test is made. As the cost of a loading test is small compared with the cost of an;y pile foundation except the smallest, it is profitable in most cases to make a loading test. It can safely be stated that pile loading tests ought to be made much more frequently than hitherto, for purely economic reasons.

§ 2. Various Types of Loading Devices.

Some loading tests on piles have been made simply by constructing a strong platform on top of the pile and loading this platform with sand, pig-iron, railway rails, or the like. The platform is prevented from overturning by being loosely supported at its corners. Lest these supports deprive the pile of an appreciable part of the load, one must take a great deal of trouble to place each load increment in such a ,vay that its centre of gntYity comes close to the centre-line of the pile. It is also troublesome to measure accurately the load placed at each step. A third disadvantage is that repeated loadings and un­

loadings of the pile can hardly be done on account of the excessive work required to shift the dead weight.

These inconveniences can be avoided by inserting a hydraulic jack between the pile and the platform. The platform is firmly supported at its corners, and the whole dead weight is resting on it throughout the test. The pile can con­

Yeniently be subjected to any desired sequence of loadings and unloadings b;v means of the hydraulic jack, the platform with the dead weight merely providing the reaction. This second method is handier and gives better results than the first 1nethod mentioned above, and nowadays it is generally preferred to the first method.

I-10\vever, constructing the strong platform and assembling the great amount of dead weight are expensive and time-consuming operations. As a hydraulic jack ought to be used anyway, it is natural to try some other type of reaction.

An existing structure can occasionally be used for this purpose, but an anchorage to the ground is normally provided, generally by n1cans of anchor piles.

The simplest possible loading device with anchor piles consists of a horizontal heavy steel beam attached at each end to an anchor pile by means of steel straps. The test pile is located half-way between the two anchor piles, and the h;ydraulic jack is inserted between the head of the test pile and the under-side of the beam. This type of loading device seems to be frequently used [see(5)].

I-Iowever, this type of loading device cannot be employed if the bearing capacity of the test pile largely consists of point resistance, because in this case two anchor piles arc not able to take the reaction. Another disadvantage is the risk of large lateral 1novements and perhaps overturn, if the centre-lines of the three piles and the jack are not exactly in one plane coinciding with the web of the beam.

In order to increase the anchoring capacity, the number of anchor piles can

be increased. T'hus, four anchor piles arc often used, and a clcYice with not less than eight anchor piles has been employed in Germany (6). In these cases the reaction is usually transmitted by a whole system of beams.

If use is made of three or more anchor piles ,vhich arc not in the same plane, the lateral stability can, as a rule (see § 5), be easily secured by constructing in an adequate way the system that transmits the reaction from the jack to the anchor piles. Such a loading cleYicc is described in § 3.

§ 3. Old Swedish Loading Device.

The loading device designed by E. \Vendcl (7) and shown in Fig. 1 has frequently been used in Sweden since 1901. Four anchor piles are employed in this device, and the lateral stability is ver,\· good. The reaction is transmitted by a system of stays and props, which has the aclYantagc of being much lighter than a system of beams.

The anchor piles arc wooden, and arc located in the corners of a rccb.mglc with the sides of about l.2 and 2.4 m, respectiYcly. The test pile is at the centre of this rectangle.

The test pile is lengthened upwards by a short steel column, and the hydraulic jack is placed aboYe this column. A steel piece reminiscent of a saddle is placed on top of the jack. The "saddle"' is connected to each anchor pile b;v an inclined sta_\'. The reaction due to the jack equals the sum of the Yertical components of the pulling forces in the sta_\'s, and is transmitted by the stays to the anchor piles. The horizontal components of the sbt,\· forces acting on the piles are taken by props inserted between the piles.

Each stay is made from steel and 1n·0Yided with a turnbuckle so that its length can easily be adjusted. It is attached to the anchor pile by means of a steel clamp which fits into a recess in the pile and is tightened with bolts.

The props are wooden, and arc made on the site to fit exactly between the anchor piles. They arc not attached to the piles, but merely rest in grooYcs dug in the soil. They abut against the piles just below the clamps.

The oil pressure in the hydraulic jack is produced with a hand-pump, the jack and the pump being integral. The pressure applied to the test pile is recorded by a manometer, and can be increased to 60 tons.

§ 4. The SGI Loading Device.

§ 4 a. Requirements for Design.

The Wendel loading dcYicc (sec § 3), characterized by its stays and props, was deemed to be, in its principal features, the best existing type. l\Iany of its details, howeYer, were considered unsatisfactory. Therefore it was decided to build a new deYice of the same type but improYed as described below.

9

,i. , ~ - - - # , . ""-~,..

~

.~·r

·----·--·--- --- ·---- ____,_

i .. r--::- -

· - - - · - - -

l?ig. 1. The TVendel loading device.

10

[. e

1 !i Ii !ii

'

,,;'! . I"

I I

I

,~

I

I I

-:--+, , -.c ,; "-·': ;;r> -,-,,(,; -r I I I

Fig. 2. The SGI loading device. Total view.

11

Fig. 3. The SGI loading device. 'Total view.

To save time and money in the loading test, piles belonging t.o the actual structure should be used, so that special piles need not be rammed. Therefore the device ought to be very flexible, so as to fit pile groups in which the ar­

rangement of the piles varies widely.

It was decided to make the props from steel and to provide them with a screw for easy adjustment of their length. Thus, the force-transmitting systc1n would be more rigid, and time would be saved.

In the Wendel device the pulling force exerted by each stay acts on the corresponding clamp eccentrically, so that the effectivity of the clamp is reduced, and the corresponding anchor pile is subjected to bending. It "·as decided that, in the new device, the centre-lines of the stay and the two props should meet above the pile in a point on the axis of the pile, so that the resulting force could attack the clamp and the pile centrically. In this way the function of the clamp is facilitated, and the rigidity of the system is increased.

In order to save time and to avoid damage to the piles, it was decided that the clamps in the new device should not be set into the piles. It was also found advisable to develop better types o! clamps which could easily be adapted to piles of different materials, shapes and dimensions.

The maximum pressure to be applied to the test pile was set at 80 tons. It was also decided that the hydraulic jack should be of the very best quality,

Fiy. !,. Prop.

Fiy. U. Stay.

so that the error in the applied pressure would not exceed I per cent of the manometer reading. In order to make such a high accuracy possible, the jack had to be placed between two ball-joints. Finally, a load maintainer was required to keep the applied pressure constant during a period of many hours irrespective of the yielding of the test pile.

13

\

\

\

.... - Socket

Screw

__:(j)__

·-/-·

Fig. G. Upper ball-joint and steel column.

Fig. 7. Upper ball-joint and steel column.

§ 4 b. Force-TransmiUing System.

The force-transmitting system of the new loading dcYicc is shown in Figs. 2 and 3.

A ball-joint socket, made from chrome nickel steel, is attacl1cd to the head of each anchor pile in the manner described in § 4 c. It has an outer and an inner spherical surface, which are concentric. The two props abut against the outer surface, and their ends are concave with the same radius as this surface.

The stay passes through a hole in the wall of the socket, and its end is a ball with the same radius as the inner surface of the socket. Thus the three forces acting on the pile are centred in a point on its axis.

Each prop (sec Fig. 4) consists of a steel tube and, at each end of the .tube, a steel head with a shaft inserted into the tube. At one end of the prop there 15

is a steel socket between the head and the tube end, and this socket is threaded to the shaft; the length of the prop can be accurately adjusted by turning the head. At the other end of the prop there is also a socket, but this one is not threaded; by exchanging it for a longer or shorter one, the length of the prop can be roughly adjusted.

Each stay (see Fig. 5) consists of a steel tnbe, a steel ball with a shaft threaded to the lower encl of the tube, and a strong steel hook with a shaft threaded to the upper encl of the tube, The length of the stay can be accurate];•

adjusted by turning the ball.

The upper ends of the stays meet in a ball-joint on the extension of the axis of the test pile above the hydraulic jack. The strange shape of this ball-joint (in vertical section it looks like an admiral's hat-see Figs. 6 and 7) makes it possible to apply the stays in any desired direction, and at the same time it ensures stability.

This upper ball-joint rests on a steel column attached on top of the hydraulic jack (see Figs. 6 and 7). The steel column consists of a tube, a screw and a socket, and its length can be accurately adjusted by turning the socket.

The hydraulic jack is mounted on the head of the test pile by means of the lower ball-joint (see Fig. 8). The bottom plate of this ball-joint has two steel arms which protrude in the horizontal plane at right angles to each other. Two similar arms are attached to the hydraulic jack directly above the lower arms. Each upper arm is coupled to the corresponding lower arm by n1eans of a turnbucklc.

By manipulating the two turnbuckles, the hydraulic jack can easily be lined up so that its axis accurately coincides with the axis of the test pile. The jack is kept in this position, while the column, the stays and the props arc erected, and the turnbucklc connections are then removed.

The lengths of the ele1nents of the force-transmitting system can be varied arbitrarily within the following limits reckoned in 1nillimctres between the centres of the ball-joints:

min. max.

props 1 070 2 420

stays 1720 2 720

column (incl. jack) 1480 2 GOO

However, each element can be lengthened as desired by exchanging its tube for a longer one. ,vhen lengthening a prop or the column in this way, the risk of buckling must of course be taken into account.

§ 4 c. Connection to Anchor Piles.

The type of connection between the anchor piles and the force-transmitting system depends on whether the piles are made from wood or from concrete.

Pig. 8. Lower ball-joint and appliance for adjusting JJosition of jack during mounting.

If the anchor piles are wooden, ther arc connected to the force-transmitting system as described below.

The ball-joint socket mentioned in § 4 b is provided with a protruding foot­

plate. A nnmber of vertical steel ribs with hooks at their upper ends are placed arouud the pile, with the hooks resting on the foot-plate (see Fig. 9). The ribs 17

,,-R1b

I - I

hm, '

j Chom::.

t *

/

...t. belt

. r

, 11,.

' \ ' \

\ r ' ·1 '\ ,,

/ ,,1 I _1J/

_J,,J

\ ' ' 1 ' \

·,

i

Fig. 9. Connection to wooden anchor pile, diameter 20-/24. cm.

cover the circumference of the pile, except for a narrow space between each two ribs, They are provisionally kept in this position by a leather belt tightened around the pile outside the ribs.

After that, two strong steel chains are placed around the pile outside the ribs, and arc kept in place by shoulders on the ribs. The chains are then put under tension with a steel vice, so that the ribs are pressed tight against the pile. The ribs have a few small horizontal teeth, which help to secure a good hold on the wood.

Thus the anchor force is transmitted by shearing stresses from the pile to

Fig. 10. Connection ta 1.cooden anchor pile, diameter> 24 cm.

the ribs. It is then transmitted by the hooks from the ribs to the socket. The pile is not subjected to bending stresses, and the connection is very rigid.

If the head of the pile is 20 to 24 cm in diameter, the hooks rest on the foot-plate as described above. If the pile is thicker than 24 cm1 a steel ring is put over the socket so as lo rest on the foot-plate. The outer diameter of the ring is slightly smaller than the diameter of the pile, so that the hooks can rest on the ring (see Figs. 10 and 11). A set of such rings with graduated external diameters are part of the device.

If the anchor piles arc made from concrete, their connection to the force-trans­

mitting system is arranged as described below.

For a pile of square cross-section, the connection shown in Fig. 12 is used. A heavy steel ring with a conical inner surface is placed concentrically around the head of the pile. Rigid steel wedges with the same angle as the inner surface 19

Ring

I I I

I

J ':

/\ ^' ·11··¹ .

J'\ I\,/

1,h \ ^{/\\/'Al !{}

^"',

/ \ ' I )1

Fig. 11. Connection to icooden anchor pile, diameter> 2,~ cm.

of the ring arc driven do,Yn between the ring and the pile, one on each side of the pile. The side of the wedge that touches the concrete is plane, but the opposite side is cylindrical with a radius somewhat smaller than the smallest radius of the ring. I-Ience, as long as the pressure is low, the wedge touches the ring only along a generatrix. In this way the pressure is centred, and irregu­

larities in the shape of the pile arc neutralized. The ring is attached by means of four bolts to n steel plate resting on the foot-plate of the ball-joint, which is placed on top of the pile. As the angle of friction between the pile and the wedge is greater than the angle of the wedge plus the angle of friction between the wedge and the ring, the device is self-locking. The device in Fig. 12 is de­

signed for a 30 cm X 30 cm pile, but it can also be applied to smaller piles by using smaller rings.

For a pile of hexagonal or octagonal cross-section, a connection similar to that shown in Fig. 12 may be used, but with six or eight wedges, if needed.

For a concrete pile of circular cross-section, a connection similar to that shown in Fig. 12, but perhaps with eight wedges, can be recommended, the inner side of each wedge being cylindrical, with the same radius as the pile.

(This type is not so handy for a wooden pile owing to the compression of the wood.) As an alternatiYe, the same connection can be used as described above for a wooden pile.

· ~ - · - -

mn

~ - - - - C o n c r e t e pile

l

.,'

, - - - · - - - , .

/ I

(

I I

.

: /

- +

\ I '

I '

I .

'

L.

Fig. 12. Connection to concrete anchor pile.

2r

Oil tank

load maintainer

j t Hydraulic jack

Pump

Fig. J.]. Diagram, of hydraulic system.

~ 4 d. llydraulic Jack with Pump, 1llotor and Load 1llaintaincr.

The hydraulic jack has an inner diameter of 170 mm and a range of 100 mm.

It can produce a maximum force of 80 tons, corresponding to an internal oil pressure of 350 kg/cm:!. In order to achicYe high accuracy, leather packings arc avoided, and the piston is lapped in to fit tight in the cylinder. Between 13 and 80 tons, the error is less than

±

The oil is delivered by a pump with three pistons. The amount of oil ddin·rcd per minute can be roughly regulated by using one, two or three pistons. The weight of the pump is 128 kg. The pump is driven by a 3/4, H.P. motor.

The pump and the load maintainer arc coupled in a circuit, as shown schematically in Fig. 13, so that the oil escaping through the load mainlainL'r returns to the oil tank. On the pressure side of this circuit there is a branch line going to the hydraulic jack. The manometer is connected to this branch.

The small amount of oil leaking through the hydraulic jack is collected in a separate tank.

The load maintainer (sec Fig. 14) functions as follows. A piston in a cylinder is actuated by the oil from one side and by a coil spring from the other.

"\Vhenever the oil pressure exceeds the spring pressure, the piston is displaced a little, a vent in the cylinder wall is freed, and some oil escapes, so that the oil pressure sinks again. The piston is continuously rotated a little to and fro in order to neutralize the friction between the piston and the cylinder wall. The spring tension is set by moving the rear end of the spring with a screw and a hand wheel. The hand wheel is set while watching the manometer, so that the load maintainer is adjusted for the desired pressure.

The outgoing line from the pump has a clamping device (see Fig. 14), which eliminates the larger pressure variations caused by the pump. It also has a.

valve for quickly removing the load without operating the load maintainer.

A separate outlet leads from this valve to the oil tank of the pump (see Fig. 14).

All parts mentioned above in this section, except the hydraulic jack and the 1notor, i.e. all parts shown in Fig. 14, arc built into one unit and are enclosed in 22

t

, · · - · · - · · - · · - · · - · ·

Valve for quick unloading

I I

I

OH tank device

I

I I

I

I

I

l_ __________________________ . __ _j

Fig. V,.. S/wtch showing part of hydraulic S!fstcm built into a nnit.

a sheet-steel case, which is shmvn in Fig. 3. This unit is mounted on a wooden platform, and, when used on a site, it is connected to the hydraulic jack by copper tubing. The motor is mounted on the same platform, and its power is transmitted to the pump by a leather belt.

All parts n1entioned in this section, except the motor, were delivered by Alfred ,J. Amsler & Co., Sehaffhouse, Switzerland.

§ 4 e. 1Vleasuring 1l1ovements of Test P1'le.

The movements of the head of the test pile during the loading procedure are measured as described below.

Two wooden poles arc driYen into the ground on the opposite sides of the

23

testing device far enough from it not to be influenced by the anchor piles. A small steel beam is attached horizontally to these poles so as to pass close to the head of the test pile (see Fig. 3).

The subsidence of the test pile is measured by two dial gauges (sec Fig. 8).

Each of them is attached to a piece of L-iron fastened to the steel beam with a screw clamp. The pins of the gauges touch a shelf consisting of hYo L-irons clamped to the test pile with two bolts. The gauges have a range of 100 mm, each scale division corresponding to a subsidence of ^0.1 mm.

The horizontal displacement of the head of the test pile is measured bet"·cen the beam and the bottom plate of the lower ball-joint with a steel rule, in the direction of the beam and perpendicularly to it. This measurement sen·cs only as a check to make sure that the test is not being invalidated by a large displacement of this kind.

§ 5. Statics of Loading Device.

§ 5 a. General Considerations.

In order that the distribution of forces in the stays, props and piles be as clear as possible, the device is always mounted in such a way that all the props arc in one plane, and this plane is perpendicular to the test pile. This is also de­

sirable for practical reasons.

In order to simplify the following exposition, we denominate the test pile as vertical and, hence, the props as horizontal. This is allowable, as gra Yi ta tion does not affect the test.

Obviously, the stays can take only pull and the props only pressure. Therefore the head of the test pile must be within the polygon formed by the heads of the anchor piles. Preferably is should be near the centre of this polygon, so

that the test force is distributed about equally among the anchor piles.

It is known fr0111 statics that a, space truss with n joints connected to a rigid foundation must have 3 n bars to be rigid and statically determinate.

This principle can be applied to our loading device. In doing so we regard not only the stays and the props but also the anchor piles as bars of the trnss, the reason being that a pile is rigid in respect to an axial force but weak in respect to a transversal force. Now, if the truss formed by the loading deYice has n joints, it has, obviously, n-1 stays, n-1 props and n-1 anchor piles, i.e.

a total of 3 n-3 bars. Thus, according to the above principle, 3 bars seem to be lacking.

§ 5 b. Case I.

,vc consider first the case in which all anchor piles are vertical (sec Fig. 15 a as an example).

24

a. Case I b

c Case III d Case III

e f Case V

Legend g Case V

Stay Prop

Anchor pile, vertical Anchor pile, inclined - · - · - Plane of symmetry

FirJ. J/J. Dia,qram of space truss formed by stays, vrops and anchor piles £n some special cases.

In this case the truss can be displaced in any horizontal direction without any resistance. In order to prcYent such moYcment, three more bars would be needed, for instance the three horizontal bars indicated by dash lines in Fig.

15 a. These bars would of course be included in the total number, 3 n, of ne­

cessary bars.

However, as the only outer force affecting the truss, viz. the force due to the hydraulic jack, is vertical, there is no tendency to horizontal 1novemcnt.

Therefore, the three additional bars arc not needed.

Jlence, the loading device with its 3 n-3 bars is rigid and statically determinate in this case. The test load will cause only those very small movements in the system which arc due to axial strains in stays, props and piles.

§ i5 c. Case II.

Let us now consider the case in which one anchor pile is inclined and the others arc vertical, as shown in Fig. 15 b.

As the vertical piles can take no horizontal forces, the horizontal component of the force in the inclined pile must be zero. But then the total force in this pile, which can take no transversal force, must also be zero. The result is that the stay, and hence the two props abutting in the same joint, will get no force, and then the whole device is unusable.

In order to make it usable, one rnust insert another bar as a diagonal in the polygon formed by the props, between the two corners adjacent to that of the inclined pile~ as indicated by a dash line in Fig. 15 b (this implies, of course, that the number of anchor piles exceeds three). The inclined pile, the stay and the bvo props adjacent to it can then be left out of account, and we come back to Case I.

§ 5 cl. Case III.

We assume thaL two anchor piles are inclined in a vertical plane perpendicular to a plane of symmetry of the loading device (including the anchor piles), whereas the others are Yertical (sec Figs. 15 c and cl).

There will be no tendency to horizontal movement of the whole device, thanks to the symmetry. The horizontal components of tlw forces iu the two inclined piles will be transmitted through the device and will balance each other. Jience the device is rigid and statically determinate in this case.

§ 5 e. Case IV.

We assume that there is one plane of symmetry and two or more pairs of anchor piles inclined as in Case III, whereas the other piles arc vertical (sec Fig. 15 e).

As the horizontal components of the pile forces balance one another in pairs, the device is rigid and statically determinate.

26

B C

;,---;;;--m---:-:x.;m

A

Fir;. JG. Diagram. of space truss formed by stays, props and anchor piles ·in the general case.

§ 5

f.

The loading device (including the anchor piles) has two or more vertical planes of symmetry (see Figs. 15 f and g).

The components of the pile forces perpendicular to one plane of symmetry balance one another. The same is true of the components perpendicular to the other plane of symmetry. Hence, the device is rigid and statically determinate in this case.

§ 5 g. Other cases.

If none of the cases described above obtains, the space truss formed by the loading device (including the anchor piles) is not rigid, as a rule, and is therefore unusable. Only by a lucky chance will the horizontal components of the pile forces balance one another so that equilibrium is secured.

To find out ,vhcther this is so in an actual instance₁ it is necessary to make a statical analysis of the whole space truss. This can be done as described below

(see Fig. 16).

The force in prop a-d, induced by a given vertical load in joint 0, is designated here by m. The forces in pile A-a, prop a-b and stay a-0 can then be computed in terms of m by means of the three equations of equilibrium of joint a (this resolution of 1n into three components ,vith given directions can also be worked out graphically according to Culman's method). We then make similar com­

putations in turn for joints b, c, and d. From the computation of joint cl we get a value for the force in prop a-cl, which we designate by x·m.

27

As a rule, we shall find that x=:= 1. implying that the space truss is not rigid.

It may happen, however, that we get x == 1, and one requirement for the rigidity of the truss is then satisfied. Two more requirements are that the forces in the stays abutting on point O must balance one another in two hori­

zontal directions. (The fact that three requirements must he satisfied cor­

responds to the circumstance that three bars arc lacking.)

§ 6. Loading Procedure.

According to Scandinavian experience (8), the bearing capacity of a friction pile surrounded by cohesive soil increases after driving, at a decreasing rate, so that, as a rule, it docs not approach its final value before 3 or 4 weeks have elapsed. For this reason, pile loading tests in Sweden are made more than 3 or 4 weeks after driving. if possible.

§ 6 a. Procedure A.

The simplest possible loading test procedure is to increase the load on the test pile steadily in steps, until the subsidence of the pile reaches a certain prescribed value. The test is then finished, and the allowable load is obtained b,v dividing the applied maximum load by the safety factor. This factor is normally 2.0.¹

This procedure is sometimes used in Sweden, and it seems always to be used in other countries (2, p G65). The procedure seems to be appropriate provided that the load on the piles after completion of the structure is always nearly the same. However, the load will fluctuate considerably in most cases, and the influence of these fluctuations on the settlement of the piles is not made clear by a test performed in this manner.

§ 6 b. Procedure B.

The test procedure used by the Geoteehnical Institute tries to take the future load fluctuations into account. It may he described here by a numerical example.

The load is first increased from O to 20 tons. It is then decreased to 10 tons and increased to 20 tons again a number of times. The settlement increases, of course, every time, but it is found that these increments become smaller and smaller. ,ve conclude from this fact) rightly or wrongly, that eYen an infinite number of such load fluctuations would not induce an excessive settlement.

The load is then increased to 30 tons, and a series of fluctuations between 30 and 15 tons is performed. Also now the settlement increments become smaller and smaller. A series of fluctuations between 40 and 20 tons giYes the same result, and also a series between 45 and 22.~, tons. A series between 50 and 25

1 It would be logical, of course, to vary the safety factor according to the local conditions, i. e.

the homogeneity of the ground, the sensitivity of the structure to settlements, etc.

28

----

ton~

30

u

w

.3

0 ~ ~

10 10

\ I \ I I

10 10 30 40 ,0 min

Time

n

-

~

\.. ^_J L J

I

~ rr

~ ]1[

, - 0

4

~

~

mm

Fig. 17'. Test procedure B. Load and settlement as functions of time.

tons, on the contrary, gives larger and larger settlement increments. Thus, a great number of such fluctuations would probably induce an excessive settlement.

The test is now finished.

The loading test procedure is illustrated in Fig. 17, showing the load and the settlement as functions of the time in the first series of loading cycles. A more practical way to show the test result, however, is used in Fig. 18. In this figure a cun·e is plotted for each series, showing the final settlement of each cycle as a function of the number of the cycle.

According to this test, the critical load, i.e. the greatest fluctuating load which

29

l!vmner cf the cycle

m rr r n DI till

'"::: I I I

s

" ,45 ils,,,.,J j

j --,_

l i I i

" ^_, I ---J..I

--~+=- ="=·-Cf·'==1I---IL__

]

,0 ,-

I i ! I r-- ⁱ

.;: i I' I I I f ... !

i ! I i ^I I

>--~,-~,--+1--',----,-.I- - i -

>---,-~1--+1-~,---~-~,- - , - - 7

,,

r:Hn

Fig. 18. Test procedure B. Pinal settlement hi each cycle as Junction of the number of the cycle.

docs not cause cxccssiYc settlement of Lhc pile, is 45 tons (sec Fig. 18). If a.

safety factor of 2.o is required, the allowable load on the piles will be 22.a tons.

At the beginning of the test, when the load is n10derate, a few loading c~·clcs (say 3 to 5) at each loading step arc sufficient in many cases to show that the settlement increments become smaller and smaller (see Fig. 18). Later, hmn:-Ycr, when the load is great, more cycles (normally 5 to 10) must be performed at each loading step in order to find out whether these increments decrease or increase.

It is desired, on the one hand, that the loading test be performed as quickly as possible, and on the other, that the critical load be determined as accurately as possible. Therefore it is best to use large loading steps, say 20 or 10 tons, at the beginning of the test, and small steps, say 5 or 2.5 tons, when approaching the estimated critical load.

,¥henever the load has been increased or decreased, the gauges are read off at time intervals of 2 minutes until the reading becomes constant. I-IoweYe1-, the operator docs not wait for settlement of the pile caused by consolidation of the clay.

In the example given above the lower load of each loading cycle was 50 per cent of the upper load. If a pile loading test is made for a structure whose permanent weight constitutes a very great part of its 1naximum weight, a higher ratio than 50 per cent may of course be used in the test. On the other hand, if the permanent weight of the structure is only a small fraction of its maximum weight, a lower ratio than 50 per cent should be used in the test.

A loading test made according to Procedure B normally takes 12 to 24 hours.

30

OH tank

-c. Load maintoiner I

- Hydraulic jack Pump

Fig. 19. Diagram,, of prospccih'e hydrmdic system.

In Procedure B the load maintainer is operated twice in every loading cycle;

this takes a lot of time, and it is difficult to get exactly the right prcssme at all times. Therefore the curves in Fig. 18 sometimes become uneven and hard to judge, so that one needs a comparatively great number (greater than stated aboYe) of loading cycles in each series in order to be able to discern whether the curYe is concave upwards or dmvnwards. Finally, even if a curve is concave u1n,·arcls₁ this is no absolute proof that the load in question, repeated a great many times, will not cause excessive settlement.

§ 6 c. Procedure C.

On account of the described inconveniences of Procedure B, the Geotechnical Institute will now change it in the following way, the modified syste1n being callccl Procedure C.

In this procedure use is made of two load maintainers, I and II, connected in parallel as shown in Fig. 19. Load maintainer I, which is supplied with a shut­

off vah-e, is set at the lower load of the desired loading cycle, and load nrnin­

tainer II at the upper load. When the valve is open, the load is regulated by maintainer I, for obvious reasons. "\¥hen the valve is shut, the load is regulated by maintainer II. Thus, the transitions between the lower and the upper load are done simply by operating the valve.

In this way we save time ,vhich can be used, instead, for making more loading cycles. Further, as the two load maintainers are not touched during the whole series of cycles, the lower and the upper load v;rill have the same values every time. Therefore, the curve of the series in question will become longer and smoother than if Procedure B were used. Hence, we hope to obtain with Pro­

cedure C curves such as those suggested in Fig. 20. For each of these curves the position of the horizontal asymptote, if any, may be fairly well estimated.

Finally, the values of the asymptotes of the different curves of the fictitious

31

tlvmber of the cycle

'

'

^{I I}

"

I

I

' , _{I I}

Sl-151,;n;

I

I

'

'

'

I r ^-

---

I I I

- L

---....

"

mm

"

Pig. 20. Test procedure C. Fictitious diagram shmci11!! final settlement i11 each cycle as function of the n1mtbcr

of the cycle.

Lood

,a 20 .

---- ^r---_ _{r---__}

'---

~

~

' _\ I

I I

!

mm

"

Pig. 21. Test procedure C. Fictitious diagram, showing settlement after an infinite nmnber of load fluctuations as function of the load.

Fig. 20 are plotted in Fig. 21 against the npper load of the series. From the fictitious Fig. 21 the critical load is estimated, i.e. that load which, repeated an infinite number of times, would cause the maximum allowable settlement. This notion of critical load is more precise than the one used in § 6 b.

Fig. :22. Loading device at 1cork on site.

In order to rationalize the loading test still more, the Geotechnical Institute is now developing an instrument for automatic measuring and recording of the settlement o! the test pile during the loading procedure. The settlement will be continuously recorded on the scale of 50 to 1 upon a slip of paper actuated by a clockwork. The diagram obtained in this way will appear as the lower part of Fig. 17.

§ 7, Present Applications.

The SGI loading device and the loading procedure B have been applied since 1943 for testing piles on many building sites in Sweden. They have been proved under varying local conditions. The range of experience now available is evident from Table I.

33

Table 1. Loading tests on

Aaverage

shear friction Ground

Time Site Structure strength¹ For

conditions of soil bearingpoint-

t/m2 p

June 10,13

I

I

I SE of Gothen- burg

soft clay

March-April Ume.''t, 520 km N Highway bridge Thick layer of 2.0 F

1945 of Stockholm sort clay and

Ditto Ditto Ditto

silt

Ditlo 3.1

I

I F

Ditto I i Ditto Ditto Diu.o 2.1 F

:May 19.fi _I I Siklcrhamn, 220 km N of Stock- Storehouse / 2 m filling on clay on firm soil 7 m

'

holm

February 10-1,$

!

! km N of Stock- holm

m clay on silt

May 194S j :i.\Icllb,v, 150 km S of Gothcnburg

Highway bridge

13 m sand on and silt

clay 2.ii F

i\fay 194-D Smme, 240 km N Highway bridge Thick layer of O.o 1-0.5 F

of Gothenburg clay and silt

July-August ₁₉₅₀ Gih·le, Hl,J km N _{of Stockholm Oil tank} 5 m clay and silt

'

on stony moraine I

If not otherwise slated, four anchor piles, parallel to the lest pile, luwe been used.

1 According lo unconfined compression test,

§ 8. Sununary.

In this report a description is given of a device and a procedure for loading tests on piles, ,vhich offer advantages that are not gained othcrvise.

The device makes it possible to use piles belonging to the structure as anchorage, fairly irrespective of their mutual positions. This is achieved by means of a force-transmitting system in the form of a space truss consisting of light stays and props of adjustable length connected to one another by ball­

joints.

T,vo new types of connection to anchor piles are described, which require no incision in the piles and ensure a centric.transmission of force. One is intended for wooden piles and the other for concrete piles. They can be applied to piles of varying sizes.

---- - - - - --- --- - - piles made 'UJJ to now.

'I' est p i I e Test result

surface

period

length widLh of _{in- critical allow-}

contacl of

material in soil of with elina- load able

m point soil tion rest

t load

clnys m:?

wood 17.7 o5 in. 1 'l.5 5:1 12 40 l.i -~ t/m²

--

!

wood 11.25 <16 in. I 7.1 oo:J 6 I 38 2.5 t/m⁹ Tesl made from lh~

iee.

wood 1L:!5 07 in. 8.i 00:] 10

i

pre-slr. 13.o 20 >< 20 10.-1 00: 1 ? 48 28 t ·1 nnC'hor piles+ dead

eoncrele em ^! weight of sacks of

sand.:?

wood 18.-1 o5 in. 12.t 00: 1 65 45 1.8 t/m² - -

wood 11.5 ^{(j ()} in. 7.:! 7: 1 30

!

wood IG.o 06 in. 10.G 00: 1 22 ^I I 30 l A t_'m²

--

'

concrete i 4.H i5 >~ 25 4.8 00: 1 ? 35 15 l 0 anehot· piles + dead

cm weight of concrcle

piles.²

i i

'

A hydraulic jack of high quality is used in combination "·ith an automatic load maintainer. The latter keeps the applied pressure constant irrespective of the yielding of the test pile.

The equilibrium of the loading deYicc is analyzed₁ when anchor piles of Yarying positions and inclinations are used. It is shown that the deYicc is always stable, when all anchor piles are parallel to the test pile. Rules arc given for a number of other cases.

,vith the described loading procedure, \Ye tr~· to take into account the fulure fluctuations of the load. It is characterized by the performance of a number of loading cycles at each loading step.

Finall~·, an account is giYen of the experience hitherto gained by using the dcYicc and the procedure in practice.

35

Bibliography

1. TERZAGHI, K., Theoretical soil mechanics. ~cw York 1943.

2. TAYLOR, D.

,v.,

3. Cm.rMINGS, A. E., KERKJIOFF, G. 0. and PECK, R. B., Effect of driving piles into soft clay. Proc. Amer. Soc. Civ. Engrs. 1948 Vol. 74 Nr 10 pp.

1553-1563. Disc. 1949 Vol. 75 Nr 5-10.

4. CADLIKG, L. and ODENS'l'AD, S., The vane borer. Stockholm 1950. (Royal Swedish Gcotechnical Institute Proc. Kr 2.)

5. BRODE, C. D., Novel pile-test loading. Eng. :Xcws-Rcc. 1940 Vol. 124 Xr 5 p. 183.

6. DIE'l'RICII, H., ~cues Hilfsmittel zur Untcrsuchung dcr Tragfiihigkci t von Rammpfahlen. Bautechn. 11 (1933) H. 23 pp. 301-302.

7. WEXDEL, E., Orn profbclastning pa palar mcd tillampning daraf pa grun<l­

laggningsforhallandcna i Goteborg. Forh. 3. allm. Svenska tckn.-motct.

Gcflc 1901.

8. SKAVEN-HAUG, S., SvreYcncle prelevrerkcrs brerccvnc og staalpclcr til fj dcl.

Geotcknik. Forcdrag fra Kursus i Dansk Ingeni9lrforcning 2-6. April 1946.

Kf)bcnhavn 1948.