FOILS SOIL

(1)

(2)

(3)

ROYAL S\VEDISH

GEOTECHNICAL INSTITUTE

PROCEEDINGS No. 1

SOIL SAMPLER

\VITI-I i\1ETAL FOILS

Device for Taking Undisturbed Samples of V cry Great Length

By

W. KJELLi\lAN, T. KALLSTENIUS, nnd 0. WAGE!\

STOCKHOLi\I 1950

(4)

Ivar liceggstrOms

IJOKTIIYCKERI A. Il. STOCKIIOJ.M 1950

(5)

Preface

The series of publications initiated by this report will contain the main results ol the research conducted by the Royal Swedish Geotechnical Institute.

The new type of soil sampler described in the report was invented by J\Ir Walter Kjellman, Head of the Institute, and Mr Torsten Kallstenius, Head of its J\Iechanical Department. The field experiments were made by Mr Olcg '\Vager, Research Department Engineer. l\1ost of the equipment ·was designed by the Mechanical Department. The report was prepared by J\Ir Kjellman.

T-Irn report covers a very great amount of theoretical and experimental ,York extending over many years, but is confined to be a summary of information bearing on the final result.

Stockholm, February, 1950.

RovAL SwEorsn GEOTECHNICAL INSTITUTE

(8)

(9)

Soil Sampler with lVIetal Foils

§ 1. Some Disadvantages of Existing Samplers.

§ 1 a. Introduction.

The samplers hitherto used in soil exploration for civil engineering purposes are capable of taking only rather short undisturbed samples. Sampling is generally performed in vertical bore holes.

Normally, the properties of the soil change much more rapidly in the vertical direction than in both horizontal directions. Therefore, it is quite logical that the spacing of the samples in each bore hole is normally made very much smaller than the spacing of the bore holes.

Another reason for making the vertical spacing of the samples very small is the fact that even a very thin soil layer may be of great importance. For instance, a slightly inclined thin layer of soft clay in the ground under a highway embankment may under certain topographical conditions cause a landslide.

A thin layer of sand under a dam or behind its abutment may cause disastrous piping.

For these reasons, it is evident that, where important structures are involved, sampling should be done in such a way as to furnish, for each bore hole, a complete and correct representation of all layers from the soil surface to the firm ground. In less important projects, sampling should be carried out in the same way, if the local conditions are such that a thin layer may be important; if this is not the case, it is sufficient to take representative samples from the ma.in layers of the ground.

§ 1 b. Intermittent Sampling.

In most cases, at least in Sweden, sampling is intermittent. This means that representative samples are taken from the main layers of the ground according to the results of a previous sounding. If the ground is found to be fairly homogeneous, samples are taken with a constant spacing of, say, 1 m.

If the ground is firm, so that driving is a great part of the work, it may be advisable to space the samples closer than otherwise. Driving, which must be performed anyway, involves nearly the same amount of ,vork, whether few or many samples are taken, provided the bore 'hole is more or less open between samplings.

(10)

This is true especially where the bore hole is eased. In that case, to take a sample does not require much more work than to remove the same length of soil by other means. IIvorslef (1, p. 65)¹mentions a case where "continuous"

sampling ,vas found to be even faster and less expensive than intermittent sampling.

§ 1 c. Nearly Continuous Sampling.

Theoretically speaking, each sample could be taken immediately below the previous one. According to Hvorslef (1, p. 64), this procedure is called continuous sampling, and is extensively used in detailed and special explorations for the foundation of important structures.

In practice, however, such sampling ,vill never be quite continuous. Firstly, the length of the same bore rod may vary a little on different occasions, depending on how strongly the join.ls are screwed together. Secondly, during the sampling process, the bore rod and the driving device and its anchorage in the soil undergo elastic deformations. And thirdly, as conditions on the site usually arc, no great accuracy can be expected in the measurement of the depths, and the recovery ratio may also be different from what is presumed. For these reasons, there may easily be an error of a fe,v centimetres in the levels from which the samples are believed to be taken. Another source of discontinuity is that the lowest part of the sample may fall out and get lost during the withdrawal of the sampler. Thus, every soil layer that comes in the partition bet,veen two samples may appear in the boring record to be a few centimetres thicker or thinner than it really is. And, still ,verse, a thin layer may under unfavourable conditions entirely escape attention. Therefore, the procedure in question ought to be called "nearly continuous sampling".

The requisite number of samples per metre of depth in this procedure can be estimated by determining the "safe length of sample", as defined by Hvorslef (1, p. 109). This length, L,, depends upon the diameter, D,, of the sample, the character of the soil, the depth belo-,v ground surface, and the design and operation of the sampler. For a properly designed and operated drive sampler, I-Ivorslef recommends

in cohesionless soils L,

= (

5 to 10) D, m cohesive soils L,

=

⁽¹⁰^{to 20)}^D,

This means that, for a sample diameter of, say, 6 cm, nearly continuous sampling will require

m cohesionless soils 3.3 to l.1 in cohesive soils l.1 to O.s

samples per metre of depth. In Fig. 1 this method is illustrated for 1.7 samples per metre.

1 The numbers in parentheses refer to the bibbiography at the end of this Ieport. Thanks to the extremely complete and valuable book by Mr Ihorslef (1), only very few other publications had to be consulted when this report was prepared.

(11)

0

a

b

C

d e

.,

'-' 0

~

~ 2 ::,

"'

·-₀

"'

~

.,

0 .D

4

5

m Number of operations

34

i i I I

I i

I i I

(-samples) for 20 m of depth :

20

₄₈ 56 1

Fig. 1. Different sampling methods schematically illustrated (sample diameter Ds

=

6 cm; shaded part of sample= undisturbed).

Existing samplers c. nearly continuous undisturbed

(safe length of sample= 10 D5 = 60 cm) d. continuous undisturbed

a. intermittent New sampler

b. nearly continuous e. continuous undisturbed

(12)

The diameter of 6 cm was chosen because it is standard at the Institute. By using a greater diameter the requisite number of samples can be reduced, but then the equipment will be more difficult to handle.

For each sample, the sampler with its bore rod must first be screwed together and lowered into the bore hole. After that it must be driven into the bottom of the hole, and then withdrawn and unscrewed. Finally it must be emptied, cleaned, and fitted with a new liner. This operation takes a lot of time, especially at great depths, where screwing together and unscre,ving of the rod involve much work.

As nearly continuous sampling requires such a great number of laborious operations, it is a very time-wasting and expensive procedure. To its credit can be said that the samples obtained, when put together one after another, form a string of soil which is a nearly complete representation of the ground.

But it must be borne in mind that only parts of this soil string are undisturbed.

The upper and the lower part of each sample are always disturbed, and only the middle part is undisturbed. The upper part of each sample is injured, firstly, by torsion or tension when the previous sample is separated from it, secondly, by swelling as long as it fotims the bottom of the hole, thirdly, when the dirt on top of it is removed from the hole or pushed aside, fourthly, by friction against, and adhesion to, the inside of the sampler when sliding through tJhe whole length of the sampler. The lower part of the sample is damaged, firstly, by torsion or tension while being separated fron1 the subsoil, and, secondly, by swelling during the withdrawal of the sampler.

§ 1 d. Nearly Continuous Undisturbed Sampling.

Better results than those given by the procedure described in § 1 c can be obtained in the following way.

Two bore holes are sunk as close to each other as possible without risk of mutual disturbance. Samples are taken alternately in one hole and in the other from such levels that the undisturbed parts of the samples in each hole only just cover the disturbed parts of the samples in the other hole. The disturbed parts are then cut off from eadh sample. After that the remaining undisturbed parts of the samples from both holes are pnt together in a sequence corresponding to their Ievel,s, so as to form a single string of undisturbed soil. This procedure should be called "nearly continuous undisturbed sampling".

The length of the disturbed top part of the sample is assumed to be 2 D,.

The length of the disturbed bottom part may be D,, which is a low estimate, provided no special device is used to cut off the sample from the subsoil. Then the rema.ining undisturbecl part of the "safe length of sample" (see § 1 e) is

in cohesionless soils L,u

=

⁽²^to ⁷⁾ ^D,

in cohesive soils L,,

=

(7 to 17) D,

(13)

Thus, for a sample diameter of 6 cm, nearly continuous undisturbed sampling will require

in cohesionless soils 8.a to 2..1

in cohesive soils 2.4 to I.o

samples per metre of depth. In Fig. I the method is illustrated for 2., samples per metre.

§ 1 e. Continuous Undisturbed Sampling.

Under favourable circumstances it may even be possible to eliminate the uncertainty in the levels from which the samples are supposed to be taken, see § I e. For this purpose, the samples must be still closer spaced in both bore holes, so that the undisturbed parts of samples from one hole overlap those from tlhe other. Now if the soil layers are thin, clear, and horizontal, some layers in the bottom part of one sample can always be identified in the top part of the next one. Thanks to these correspondences, the real levels from which the samples were taken can be deduced irrespective of the uncertain level figures in the boring record. Therefore, this procedure should rightly be called "continuous undisturbed sampling".

If the undisturbed part of the sample ,has the same length as that mentioned in § I d, and if the length of overlap is D,, the effective part of the "safe length of sample" is

in cohcsionless soils Lse

=

(I to 6) D, in cohesive soils L"

=

⁽⁶^{to 16)}^D,

Thus, for a sample diameter of 6 cm, continuous undisturbed sampling will reqmre

m cohesionlcss soils 16., to 2.s in co11esive soils 2.s to l.o

samples per metre of depth. In Fig. I the method is illustrated for 2.s samples per metre.

Continuous undisturbed sampling gives the best results that can be obtained by means of existing samplers. For instance, it is the only method by means of which these samplers can be used in geochronology, where it is imperative to get complete and correct profiles of the varved clay. On the other hand, this method is very expensive and time-vrnsting, still more so than the metlhod in § I c.

§ 2. Atte1npts to Improve Existing Samplers.

The main disadvantages of existing samplers can be attributed to the fact that they can take only rather short samples.¹If the length of sample could be materially increased, the diameter being kept constant, great advantages would be achieved.

1 In oceanography cores up to 2·0 m in. length are taken b,r piston samplers. Thjs can be achie\·ed by using the enormous hydrostatic pressure at the sea bottom to force the soil into the sampler.

Of course, such cores are disturbed.

(14)

Firstly, the requisite number of sampling operations ,vould decrease. In nearly continuous sampling it would decrease in inverse proportion to the length of sample. In nearly continuous undisturbed sampling and in continuous undisturbed sampling it would decrease still more. Consequently, the sampling process would be more rapid and less expensive.

Secondly, the result of sampling would be better. In intermittent sampling and in nearly continuous sampling a relatively shorter part of each sample would be disturbed.

The fact that existing samplers are capable of taking only rather short samples is chiefly due to the sliding resistance which is caused between the sample and the inside of the sampler when the latter is pushed down around the former.

Whith the increase in length of the sample that has entered the sampler, the sliding resistance becomes greater, and causes an increasing excess pressure in the mouth of the sampler. ,vhen the "safe length of sample" is reached, the excess pressure is so high that part of the soil under the mouth is pushed aside instead of being caught by the sampler. Then it is useless to drive the sampler any deeper.

Attempts have been made to improve existing samplers by reducing their inside -sliding resistance. For this purpose, Hvorslef (I, p. 95) recommends that the inside of the sampler be polished or lacquered, and also oiled. Polishing and lacquering may reduce the friction of coarse-grained soils, but it is uncertain whether they can reduce the adhesion of fine-grained soils. Oiling, on the other hand, will ,be effective in fine-grained soils, but probably not in coarse-grained soils. Evidently, these measures can reduce the sliding resistance only to a certain degree. Furthermore, it is to be remembered that the- sliding resistance is useful inasmuch as it helps to retain the sample in the sampler during the withdrawal.

According to I-Ivorslef (1, p. 135) the sampler should be forced down by rapid continuous pushing. Then it should be at rest for ten or twenty minutes before starting the separation and withdrawal. In this way the inside sliding resistance is reduced during pus11ing, and then partly restored before withdrawal. In fine-grained soils, these effects are mostly due to thixotropy, in coarse-grained soils, they can chiefly be attributed to the fact that the coefficient of friction is usually smaller in motion than at rest. However, the amount of reduction and the amount of restoration vary considerably with the type of soil. In Scandinavian

"quick clays", the reduction is extremely high, and the restoration is very small.

In normal clays, both the reduction and the restoration are fairly good. In coarse-grained soils, both effects are rather slight.

In order to help the sample entering the sampler to overcome the sliding resistance, a stationary tight-fitting piston can be inserted above the sample.

Thus t.he pressure is reduced above the sample, whenever it tends to separate from the piston. This suction not only secures a specific recovery ratio (1, p. 101) of 100 per cent during the driving, so that the sample is less disturbed than in other samplers, but also helps the sampler to •keep the sample during the processes of separation and withdrawal. This sampler with a stationary piston

(15)

invented by John Olsson in 1923 has been found to be the best existing sampler for 11se in ordinary soils (1, p. 130). Only in very coarse-grained soils, troubles 1nay occur because pore water may be sucked out of the sample and the subsoil, and gather between the piston and the sample.

Outside Scandinavia, samplers are usually given a certain inside clearance, i. e.

the inner diameter is a little smaller in the lower mouth than in the rest of the sampler. This clearance is said to reduce the sliding resistance by reducing the pressme in the sample (1, p. 107). However, it must be kept in mind that the total pressure, which in the first place is reduced, has no appreciable influence on the ·sliding resistance. Only if the effective pressure, too, is reduced on account of rapid swelling of the sample, will the sliding resistance decrease. This happens in those samples which are very -expansive, i. e. contain much gas, and in those samples which are coarse-grained enough to suck in ,vater from the subsoil very quickly. In both cases the shearing strength of the sample is reduced at the same time as the sliding resistance. In ordinary soils, no appreciable swelling will occur, and therefore the clearance cannot reduce the sliding resistance by reducing the pressure in the sample.

In open samplers, the clearance may induce water or air from the bore hole to form a film between the sample and the sampler wall, so that the sliding resistance is reduced during the driving operation. If the fihn consists of water, the sample will absorb this water and swell, so that the resistance will be increased again during the withdra,val, while the shearing strength will decrease.

These phenomena may occur in most soils, and they are probably the most important result of the clearance.

It should be observed that the last-mentioned effect cannot develop in piston samplers. Nevertheless, piston samplers with a clearance seem to be frequent in the U.S. When such a sampler is used in gas-free fine-grained soils, the volume corresponding to the clearance is generally filled, owing to entrance of excess soil. The intended reduction of the sliding resistance will not be achieved.

Owing to the measures described in this section, the safe length of sample has reached the values given in §§ 1 c to e. Perhaps these values could be ,somewhat increased by making full use of the advantages brought about by the tight

fitting piston. I-Iowever, it does not seem possible to go much further on these lines.

§ 3. Purpose of New Sampler.

At the outset of the experimental work on the new sampler described in this report, the purpose in view vms to increase the length of sample as much as possible. Later, when the principle adopted for the sampler as described in § 4 proved very successful, ,ve concentrated on the 1nore precise aim of attaining a length of sample equal to the depth to the firm ground. Thus only a single operation ,vould be required in each bore hole.

The result obtained by means of such a sampler would equal the result obtained in continuous undisturbed sampling by existing samplers. It would,

(16)

of course, be better than the result obtained in intermittent sampling, in nearly continuous sampling, and in nearly continuous undisturbed sampling.

It was foreseen that the new sampler would work very rapidly and cheaply in a great many cases. Only ,vhere intermittent sampling with wide spacing of samples is sufficient, the latter method was believed to be faster and cheaper.

§ 4. Principle of New Sampler.

The principle of the new sampler consists in completely eliminating the sliding resistance between the sample-which ,vill be called "the core" owing to its great length-and the inside of the sampler. This is achieved, as shown schemati

cally in Fig. 2, by insulating the core from the inside of the sampler by means of a number of thin axial metal strips or foils. Their upper ends are attached to a piston placed in the saanpler immediately above the core. The piston, which is not tight-fitting, is attached by means of a chain or a rod to a driving scaffold on the soil surface. Therefore, the piston, together with the foils, is kept on a constant level, while the sampler is driven down. During the driving, the inside of the sampler slides against the foils. The resulting friction causes a pulling force in the foils and their anchorage but does not affect the core. There is no sliding between the core and the foils.

If, for some reason or other, the core should tend to move downwards or upwards, this movement is immediately prevented by friction and adhesion between the foils and the core. Thus, the recovery ratio as defined by Hvorslef (1, p. 100), is automatically kept equal to 100 per cent. The friction against, and adhesion to, the surface of the core, which are the main cause of all diffi

culties in other samplers, have in this way been turned into useful forces which keep the original thickness and structure of each soil layer unchanged on its way into the sampler and upvi'ards inside the sampler.

During the withdrawal, the foils follow the movement of the sampler. Each part of the core hangs on its part of the foils by friction or adhesion.

The lower end of the sampler is shaped into a sharp cutting edge. Tihe sampler is provided with a double wall, which extends through a part of its length from a level slightly above the edge, so as to form a magazine for the foils. At the bottom of the magazine there is a horizontal slot in the inner wall. The foils pass through this slot into the interior of the sampler, and run from the slot up,vards between the core and the sampler wall all the way to the piston. As the sampler cuts out the core during the driving, the core is covered by the foils which are fed continually through the slot.

§ 5. Outline of Stress Conditions in Core.

This section gives an outline of the stress conditions in the core under the influence of the foils and the clearance. These conditions, and also the influence of some other factors, will be more thoroughly discussed in §§ 8 to 11.

(17)

Fig. 2. Sketch showing pr-inciple of new sampler.

(18)

§ 5 a. Cohesive Soil.

If an ordinary tube is driven down into clay, downward shearing stresses are transmitted by adhesion from the tube to the core formed inside. Therefore.

the vertical pressure in the core increases, so that, on each level, it ,vill be greater than the vertical pressure in the soil outside the tube. This excess pressure in the core increases from zero at the soil surface to a maximum at the lower end of the core (see Fig. 3). For instance, if the core is 6 cm in diameter and I.a m in length, an adhesion as low as 0.01 kg/cm²will cause an excess pressure of not less than 1.o kg/cm²at the lower end of the core. Owing to this high excess pressure, the core would push the underlying soil aside, and the core itself would move downwards.

Vert1co1 pressure

0 0, !.O 1<q/cm¹

I '-....

^W1thour^foils

w1thou1 clearance

o~~~

,o

m

Fig. 3. Influence of foils and clearance on vertical pressure in clay core. (Diameter= G cm, unit weight= 1.o t/111,3,

adhesion= 0.01 kg/cm².)

If foils are used, the core is protected from shearing stresses due to the tube.

'l'he vertical pressure ,vill be the same in the core as in the soil outside the l11l1c (see Fig. 3). Therefore, the core cannot push the underlying soil aside and move down,vards. Consequently, the direct purpose in view has been achieved.

liowever, the horizontal pressure exerted by the core causes friction between the tube and the foils, and this friction produces a pulling force in the foils.

If the length of core is great, the foils are exposed to the risk of breaking. Jn order to eliminate this risk, the tube is given a certain clearance ,,thich is kept air-filled. When entering the wider part of the tube, the core then tends 1.o expand horizontally, and therefore tends to contract vertically and to moye downwards. Thus the core hangs on the foils by adhesion, so that the vertical pressure, and hence also the horizontal pressure, wiU be smaller in the core than in the soil outside the tube. The lower pressure in the core reduces the friction of the tube against the foils, and therefore also the pulling force in the foils, to harmless values.

(19)

If the clearance is great enough, the core will completely hang on the foils.

Then the vertical pressure in the core will be zero (see Fig. 3). Consequently, the horizontal pressure in the core and the friction between the tube and the foils will also become zero. The pulling force in the foils will then equal only the weight of the core, which is small compared to the strength of the foils.

§ 5 b. Cohesionless Soil.

If an ordinary tube is driven down in sand, downward shearing stresses are transmitted by friction from the tube to the core. Therefore, the vertical pressure, and hence also the horizontal pressure, in the core increase. The increase of the horizontal pressure causes an increase of the friction, which in its turn increases the vertical pressure, etc. This phenomenon constitutes a reversed silo effect, and it can be calculated in the following way.

"\Ve assume that the vertical pressure is constant in each cross section of the core, and further that the ratio of the horizontal and vertical pressures is constant throughout the core. Consider a N1in horizontal layer of the core, as shown in Fig. 4. The following symbols are used:

z

=

depth of layer below soil surface, dz

=

thickness of layer,

r ;;;:;::::: radius of core, y

=

unit weight of core,

/3

=

angle of friction on surface of core, a

=

vertical pressure on top of layer,

o

+

d a

=

vertical pressure on bottom of layer, au= horizontal pressure in layer.

Equilibrium of the layer in the vertical direction requires that y 1r r' dz - x r' cl a

+

a a 2 :n: r dz tg /J

=

⁰

By inserting

m=-tg/J2a r

the above equation is converted into

faz=}-,,::

which by integration becomes

. . . (1) This is the formula for the inverse silo effect.

17

(20)

rE--2r-

l

1

^cccr _dz

""

_O+do

i ₇

Fig. 4. SlLetch for calculation of inverse silo effect.

ver:1cc! pressure

0 o, \0 l<O/c"'.'

0

ft'C"'=====.c=========

W11t>0ul fo,15, w<tnout cieor-once

,o m

Fig. 5. Influence of foils and clearance on verfical pressure in sand core. (Diameter= 6 cm, unit weight= 1.o t/mfl,

~atio of horizontal to vertical pressure= 0.,;, friction between foils and core= 20°.)

(21)

Assuming, for instance, r=3 cm, y=l.s gr/cm^3, ~=20°, and a=O.a, Equation (1) gives the upper curve in Fig. 5 showing that the pressure in the core increases enormously with the depth. This implies that only very short cores, or rather samples, can be taken in thls manner.

By introducing foils the inverse silo effect is eliminated, so that the vertical pressure will be the same in the core as in the soil outside the tube (intermediate curve in Fig. 5). In this way the core is prevented from pushing the underlying soil aside and moving downwards.

In addition, when the tube is provided with a clearance, the ordinary silo effect sets in according to the well-known Janssen equation

<5=

~ (1-e-m') ...

⁽²⁾

where the symbols have the same meaning as in Equation (1). This effect reduces the vertical pressure in the core to a very low value, as shown by the lower curve in Fig. 5, which represents Equation (2) alter insertion of the same numerical values as those inserted in Equation (1). Therefore, the horizontal pressure in the core and the friction of the tube against the foils will also be low.

1.'he pulling force in the foils will then just slightly exceed the weight of the core, and therefore it will be harmless.

§ 6, Development of Sampler Head.

§ 6 a. Model I.

The method of insulating the soil core from the sampler wall by means of foils running parallel to the axis of the core anchored at their upper ends was devised in 1943.

Preliminary tests were carried out with foils made of rubber, celluloid, aluminium, and brass. A sampler head, which will be called ]\fodel It in what follows, was designed for a core diameter of 60 mm (see Fig. 6). In this model each foil is stored in a roll placed in a magazine in the sampler head. The foils are six in number and rather thick. Therefore, ii the core length is to be great, the rolls will be rather bulky. Lest the outer diameter of the sampler in a cross section through the magazine be too great, the rolls are placed one above another in such a way that each of them is concentric with the core. Each foil alter having left the roll is de!lected from the horizontal to the vertical direction by passing over a lip inclined 45° from the horizontal plane. From the lip the foil runs downwards to the slot mentioned in § 4, and passes through it into the interior of the sampler.

This sampler has never been built. A model of the interior parts was built and tested in the laboratory. During these tests, in which a wooden pole served as a core, it proved di!!icult to make the foils run as they should. When passing

1 The design and experimental work with this model was performed by Messrs. Y. Liljedahl and D. Almstedt. of the Research Department.

(22)

··-1·

1. Foils.

__j 2. Roll.

3. Lip.

4. Slot.

Fig. 6. Sampler head JY!odel I.

(23)

0 0 ITT N

1. Foil.

2. Slot.

60

Fig. 7. Sampler head Model II.

21

(24)

over its lip, a foil was easily displaced laterally, so that its border touched some other part of the apparatus and was scratched. Certain possibilities of guiding the foil were considered, but they were not put into effect. Later on, a simple way of guiding the foils was developed and tested, bnt the principle was abandoned, as simpler devices had in the meantime been invented.

§ 6 b. Model II.

It was decided to try the foil principle in the simplest possible device. There

fore, the apparatus illustrated in Fig. 7 -and called Model II was constructed in 1945.

This sampler made for a core diameter of 60 mm and a maximum core length of 2., m, was double-walled through the whole leugth above the slot. The annular space between the inner and the outer wall was divided by ridges running parallel to the axis of the sampler into six canals, each of them holding one foil. Celluloid foils were used.

It proved difficult and time-wasting to insert the foils into the sampler.

Attempts to take up soil cores from soft ground, the sampler being pushed down by hand, failed partly, because the foils were damaged and broke.

Evidently, the celluloid foils were not strong enough to withstand the pulling force produced in them. This force may partly have been caused by the fact that the foils got stuck in the slit inside or outside the ridges. It may also have been produced because the core strongly pressed the foils against the wall, which was too rough for efficient lubrication. The high pressure in the core may have been brought about by the great yield of the foils and their anchorage, or by the considerable changes in t11e cross-sectional area of the sampler.

Later on, the sampler was revised and tested with steel foils. The ·result was better, and it seems reasonable ta presume that a double-walled tube can be used in practice, altl10ugh there will be trouble when long cores are wanted.

§ 6 c. Model Ill.

In order to make a sampler work according to the adopted principle, it was evidently necessary to provide better conditions of operation for the foils and to make them as strong as possible. For this purpose, and to the detriment of some geotechnieal desiderata, the sampler illustrated in Fig. 8 and called Model III was designed and built.

This sampler was made for a core diameter of 60 mm and a maximum core length of 3 m. The uumber of foils was increased to eight, in order that the buckling which each foil undergoes when passing through the slot should be somewhat reduced. The foils made either from 0.os mm thick tempered strip steel or from mild steel were rolled on reels such as those used in a camera.

The reels were placed in the magazine of the ,sampler head with their shafts tangentially directed relative to the core, so as to avoid the 90° deflection of the foils which caused difficulties in Model I. All reels were placed on the same

(25)

A-A

I

B-B I

Fig. 8. Sampler head Model III. (Piston locl-:, not shown.) l. Foil. 2. Roll, 3. Space for core retainer mechanism.

level above the slot, so as to form a ring around the core. Small rolls ,vere placed at the slot. The inside of the sampler had an exact diameter, and was very sm-ooth. Before the sampler was driven down into th·e ground, the piston, to which the foils were attached, was locked in the lower month of the sampler.

After reaching the depth where the coring was to begin, the piston was released and kept on a constant level. Thus, it was possible to take firstly a core from the soil surface to the level -3 m, then a core from -3 to -6 m, etc. By putting these cores together, a complete representation of the ground was obtained from the surface to the firm ground.

(26)

In April, 1946, the sampler was tested for the first time in the ground on a soft clay site at Ifagalund, outside Stockholm. The sampler was pushed down by means of a lever device. The success was complete. A 3 m long core ,vithout perceivable disturbances was obtained. The experiments ,vere then continued on other sites, and a great number o.f cores, as a rule quite undisturbed, at least visibly, were obtained from clay, mud, and sand soils.

The sampler was driven d°'vn by means of simple provisional devices, and the pulling force in the anchorage of the piston in the driving scaffold, i.e. the pulling force in the foils, ,vas measured at the same time. It proved very important to keep the piston quite still, because if it was allowed to sink a little bit, the pulling force immediately increased alarmingly. The light tripod which at first was used as a boring scaffold was therefore replaced by a stronger one. Never

theless, it happened sometimes, especially at great depths and in so-called quick clay, that the pulling force in the foils became too great, evidently because the pressure in the lowest part of the core was too high. Therefore, the sampler was lengthened downwards by a thin-walled piece of tube, and a clearance was provided. The purpose of the lengthening was to remove the cutting edge from the region where the excess pressure produced by the bulky sampler head in the adjacent clay mass is deemed dangerous for the sample.

Furthermore, the lengthening created a certain "braking range" which made possible a pressure decrease in the core from the mouth of the sampler to the level of the slot.

After these modifications, Model III proved quite fit for practical purposes.

§ 6 d. Model IV.

After the foil principle had been found to be useful, it was decided to construct a new model so that its shape should be geotechnically better than that of the very bulky sampler just described. Everything was done to make the wall thickness of this new model as small as possible. On the other hand, the require

ment of good conditions for the foils was somewhat neglected.

The result was the sampler Model IV, illustrated in Fig. 9 and made for a core length of 5 m. 'rhe core diameter was increased to 68 mm. Consequently, a sample cut from the core can be freed from its thin disturbed surface layer before being put into laboratory apparatuses, which are made for a sample diameter of 60 mm. The number of foils was increased to 20. The rolls were placed in a ring around the core with their axes tangentially directed, as in the previous model. They were placed rather high up in order that the bulky torus which they formed in the outer contour of the sampler should be at a safe distance from the month of the sampler.

This sampler was used for taking a lot of cores in the field. As was expected, it gave extraordinarily good cores. However, the loading of the sampler with foils proved rather troublesome. Another disadvantage was that the foils were deformed when passing over the lip in the slot, which had a very small radius.

This deformation did not injure the core but made it unpraetical to use the foils more than once.

(27)

O·

8-B I

B

'

1,

r I

! A-A

:' I

A A

68

Fig. 9.

Sampler head Model IV.

I. Foil. 2. Roll.

25

(28)

A-A

I. Inner wall with recesses.

2. Foil roll.

3. Slot.

4. Upper guide.

5. Lower guide.

6. Outer wall.

'1. Foil fastener.

8. Lock.

68

Fig. 10. Sampler head :A-Iodel V. (Foils shoun only in rolls.)

(29)

Fig. 11. Details of sampler head.

a. Inner wall and a foil roll. c. Piston.

b. Outer wall. d. Sampler head assembled.

§ 6 e. Model V.

On the basis of the valuable experience acquired in the tests with the extremely bulky Model III and with the extremely slender Model IV, a new sampler was constrncted in 1947. In respect of wall-thickness this sampler is intermediate between the previous two models. It embodies several material improvements.

No,v the experimental stage had come to a,n end, and the sampler with metal foils was ready for regular use. The new model is called Model V, and is described in detail in the following section.

§ 7. Details of Sampler Head, Moclel V.

For the same reason as be-fo1¹e, the sampler l\Iodel V is made for a core diameter of 68 mm. The maximum core length vrns increased to as much as 40 m, and this should be sufficient in any practical case.

In order that it should be easier to load the sampler with foils, the number of foils was decreased to 16. For the same reason, and also in order to limit the 27

(30)

Fig. 12. Photo of lower part of inner 'Wall, showiny two foils r-unning from roll through upper

and lower guides into slot.

maximum outer diameter of the magazine in spite of the great length of foils, the foil rolls are placed on four levels above one another, the axis of each roll being radial with respect to the core.

As is seen from Fig. 10, the magazine consists of a thick inner wall, with arbored recesses for the rolls, and a thin outer wall which protects the foils from the soil. The top of the inner wall is threaded to the sampler tube; its lower end extends to the slot. The outer wall is attached in the middle to the inner wall by means of screws. At the top it abuts against a shoulder on the inner wall, and at the bottom it is threaded to the sampler wall below the slot.

The shape of the inner wall and the positions of the rolls in this wall are shown in Figs. 11 a and 12. In order to facilitate the loading of the sampler with

(31)

foils, each roll is kept in position by a small vertical rod. On its way from the roll downwards to the slot, each foil is directed by two guides. The upper guide consisting of a groove sa,vn in the inner ,vall forces the foil to run off correctly from the roll. The lower one directs the foil to the slot. In the gap between the two guides the foil is twisted 90°. 1'he guides space the foils evenly on the circumference of the sample.

As shown in l•'ig. 11 b, the outer wall is divided into two parts. The lower part, which is attached to the inner wall by four screws, keeps the upper part in position. By removing the screws and pu1Iing off the outer vrnll, the foils are easily and quickly made accessible.

The sampler wall below the slot is made as thin as possible without impairing its strength. It ends below in a thin, sharp edge which is kept in position by a threaded ring. By virtue of its shape, the edge can easily be adequately hardened.

It is inexpensive aud readily replaced.

The piston is shown in Fig. 11 c. The upper end of each foil is securely wedged to the piston by means of a conical ring, which is visible half way up on the piston. On top the piston is provided with a lock, so that it can be fixed in its lowest position relative to the inner wall. In this position, the lower end of the piston covers the lower mouth of the sampler, see Fig. 11 d, so that the sampler can be driven down into the ground without taking any core, if desired. After reaching the depth where coring is to begin, the chain provided with a special catch is lowered in the sampler tube until it engages the head on top of the piston, sec Figs. 10 and 11 c. This releases the piston from the inner wall and attaches it to the chain which is then clamped to the driving scaffold. In this way the piston is kept on a constm1t level, so that the lower mouth of the srnmpler is freed during the subsequent driving to admit the core cut by the edge.

Each foil is 12.5 mm wide, and the foils cover nearly the whole circumference of the core. As a rule, the foils are made of cold-rolled mild steel. Tempered steel foils have also been used, and have given good results. They are stronger but more liable to rust. The foil thickness has varied from 0.12 to O.o, mm.

The maximum length of foil that each roll in the magazine can hold varies according to the thickness, and is 40 m when O.o5 mm foils are used.

The inner \vall, which needs much machining, is made of bronze. In this way rusting is prevented, and friction between the foils and the wall is reduced.

All other parts of the sampler arc made of steel for the sake of strength.

§ 8, Analysis of Stresses in Upper and Intermediate Parts of Core.

In this section it is assumed1 that the clearance space is air-filled.

§ 8 a. Three Principal Cases.

The case shown by the lowest curve in Fig. 3, where the whole weight of the core hangs on the foils, will be called "Case A". Evidently this case requires that

(32)

Vertical pressure

0 o,

0.,..---~---~---

~ Case A

Cose B -Case C

10 m

Fig. 13. Vertical pressure in 1ipper and intermediate part of core. (diameter= 6 cm, unit weight= 1.r; t/111}!,

adhesion= 0.001 kg/cm,ll.)

Case A: ordinary clay. Case B: quick clay, Case C: sand.

the adhesion between the core and the foils should be able to transmit the shearing stress caused by the weight of the core. If, for instance, the core has a diameter of O cm and a unit weight of l.r, t/m3 , as small an adhesion as 0.002

kg/cm²will be adequate. Accordingly, in most cohesive soils Case A will occur.

llO\vever, there is one soil which possesses considerable cohesion, while its adhesion to the foils is not reliable, and that is the "quick clay", which is found in Scandinavia. When remoulded, this clay loses more than 98 per cent of its shear strength-' When taking cores in quick clay, it is probable that the outer thin layer of the core, which has been touched by the edge, is capable of only very slight adhesion to the foils. Then the pressure in the core cannot decrease to zero but only to a certain value which increases with depth at a lower rate than the pressure in the soil outside the tube. Consequently, considerable friction is produced between the tube and the foils, and, when the length of core is great, the foils ma.y break. The case now described will be called "Case B".

Finally, the case represented by the lowest curve in Fig. 5, where there is a very lov;r pressure everywhere in the core, will be called "Case C". This case occurs in those soils which completely lack cohesion but pos_sess friction.

Fig. 13 shows a comparison of the three principal cases. Cases A and C are very favourable ,for taking cores, and they represent the normal, behaviour of soils. Case B is unfavourable, but fortunately it is rather unusual. However, a soil 1nainly representing Case A and/or C, may happen to contain a layer that

1 More exactly (2, p. 3) a clay is said to be quick if its ratio H_{3 :}H_1_ e.xceeds 50 and its I-1₁is very low, H ₃and H₁being the undisturbed strength and the remoulded strength (»h1illfasthelstal») respectively according to the Swedish cone test (3, p. 51). Ratios up to 6"00 have been observed.

(33)

behaves according to Case B. Therefore, it is not sufficient that the new sampler is practical in Cases A and C; it must also be able to take at least short cores in Case B.

§ 8 b. Influence of Elastic Deformations of Sampler Tube.

It must be ascertained that the elastic deformations which the sampler tube undergoes during the coring operation do not affect the stresses in the core and foils.

During the driving of the sampler there is continuous sliding between tube and foils, and therefore the friction between them is as great as it can be, the pressure due to the core being given. Consequently, the axial compression of the tube caused by the increasing pushing force cannot affect the stress in the foils.

Afterwards, the pushing force is replaced by the pulling force, and the pulling force decreases during the withdrawal of the sampler, with the result that axial strains are caused in the tube, and these strains influence the stress in the foils.

I-Iowever, since this stress cannot become greater than it was during driving, these influences are of minor interest.

In addition, the pushing force and the pulling force produce lateral strains iu the tube. If, for instance, the axial pressure is 1 000 kg/cm^{2 ,}the modulus of elasticity is 2 000 000 kg/cm^{2 ,}Poisson's ratio is 10/3, and the radius is 3 cm, the radial dilatation will be 0.oo, mm. This value is small compared with the tolerance of the tube, which is +0.1 mm, and cannot appreciably affect the stress in the core.

Owing to the outer earth pressure, the tube undergoes a small lateral compres

sion. For instance, at a depth of 20 m in a soil having a unit weight of 1., t/m^{3 ,} the horizontal total pressure is about 3 kg/cm^{2 •}The inner pressure may be zero.

The wall thickness is 4 mm, and the modulus of elasticity is 2 000 000 kg/cm^{2 •} 'I'hen the radial compression will be 0.ooos mm. This value is small compared with the tolerance of the tube, and cannot appreciably affect the stress in the core.

§ 8 c. Influence of Swelling of Core.

In Cases A and B the soil is usually saturated with water, and usually has a very high capillarity and a very low permeability. When such a soil undergoes a pressure decrease, it swells, as a rule, only very little. The amount of s,v.elling is determined by the coefficients of compressibility of the grains and the pore water, which may be O.ooo ooa and O.ooo 050 respectively. For soft clay having a porosity of, say, 71 per cent corresponding to a unit weight of I.a t/m³ the compressibility will be 0.ooo oao. When a core is taken at a depth of, say, 20 m in such clay, the reduction of the isotropic total pressure will be 3 kg/cm²and the amount of swelling will be 0.011 per cent. This corresponds to a radial swelling of 0.002 mm. This value is small compared with the tolerance of the tube, and cannot appreciably affect the stress in the core.

(34)

In ca.ses A or B, if there are bubbles of gas in the pores of the soil, these bubbles will expand owing to the pressure decrease during the coring process, and the result will be a much greater swelling of the core than that computed above.

Even if there are no gas bubbles at the beginning, but if the content of gas dissolved in the pore water is high, gas bubbles will develop because of the pressure decrease, and the result will be the same. Great swelling may also occur in those soils whose capillarity is so small that they are able to suck in air from the clearance space. The fine-grained soils n1entioned above constitute what is usually meant by "expansive soils". According to Hvorslef (1, p. 178), a volume expansion of 7 per cent has been observed. This corresponds to a radial expansion of 1 mm, ,vhich would of course cause troubles when trying to take long cores. Fortunately, these soils seem to be rare.

In Case C, which refers to coarse-grained soils, the core will swell to a certain extent by sucking in air from the clearance space. For instance, if the unit weight is l.s t/m³ and there is no ground-water, the pressure decrease at a depth of 20 m will be nearly 2.4 kg/cm²(average in the three principal direc

tions). In the coarse-grained soils under consideration, this pressure decrease can produce a radial swelling of the core of, say, O.oa mm. This value is less than the tolerance of the tube, and it cannot appreciably affect the stress in the core.

Finally, it may be mentioned that even where the ground contains no "expan

sive soil", parts of the core may swell exceedingly for the following reason. If the ground contains both coarse-grained and .fine-grained layers, the fine-grained layers may s,vell by sucking in water frmn the coarse-grained layers, this water being replaced by air that is sucked in simultaneously from the clearance space by the coarse-grained layers. For instance, at a depth of 20 m in a ground having a unit weight of 1., t/m^3,the decrease of the isotropic effective pressure during the coring process will be about 1 kg/cm² (if the ground-water level is at the soil surface). In a fine-grained layer, this pressure decrease may cause a radial swelling of, say, O.s mm. However, this very great swelling can take place only in comparatively short parts of the core, namely in those parts of the fine

grained layers which are close to coarse-grained layers. The reason is that the pore water has time to migrate only short distances in the fine-grained layers during the coring process. Consequently, if the fine-grained layers are thin, as is the case in some varved clays,! the great swelling will ·spread to every part of the layers during the coring process. If, on the contrary, the fine-grained layers are thick, the great swelling will spread to their interior parts only if the cores are kept in the tubes for a long time.

§ 8 d. Influence of Lateral Bending of Foils.

In a stress-free state the foils are flat. In the sampler they are, as a rule, bent laterally, and must therefore exert an elastic, radial pressure on the core. This pressure will now be computed on the following assumptions, viz. inner diameter

1 In most varved clays all layers are fine-grained; the difference between adjacent layers lies mainly in the colour.

(35)

- - - -

--- ---

- ^=~~- -~-~-=~-.,,,,

~ ^{,$"- -}

~=- ~,

Fig. 14. lrlagnified part of cross-section of sampler tube and foils, slwuing influence of lateral bending of foils.

of sampler 68 mm, width of foils 12., mm, thickness of foils 0.1 mm, and modulus of elasticity of foils 2 000 000 kg/cm^{2 •}

In the short space extending from the slot up to the level where the clearance begins, the foils are pressed by the core against the sampler wall, so that their cross-section has the same curvature as the wall. This is shown by full Jines in Fig. 14, which represents a magnified part of the cross-section of the sampler.

Each foil forms a 12., mm long arc of a circle with a radius of 34 mm. The height of this are is 0.57 mm. Now a 12., mm long and 0.1 mm deep steel beam supported at its ends will bend 0.57 mm under an evenly distributed load of 0.30 kg/cm2.

Therefore, the elastic pressure of the foils against the core will be approximately

0.30 kg/cm²,where there is no clearance.

After having passed the level where the clearance begins, the foils tend to recover their flat shape as far as the clearance permits, i. e. until their edges touch the sampler wall. It is now assumed that the core does not swell at all, but that it follows without resistance the lateral deformation of the foils so as to remain in contact with them throughout their whole width. In Fig. 14 the clearance has been chosen so that the foils become quite flat when their edges only just touch the sampler wall, as shown by dash lines. This radial clearance, which amounts to 0.38 1nm, reduces the elastic pressure of Hie foils to zero.

In Fig. 15 the elastic pressure of the foils is shown approximately as a function of the clearance. According to the above considerations, the pressure should

kg/cm' 04

Standard Toleran e clearance of tube w 0.3

:, '-

"'

Q)

"' 6..

0.2

ci '-

..5 .I! 0.1

0

'---,----.--J....,-~,.,,

0 ^0.1 0.2 0.3 0.4

Radial clearance ^mm

Fig. 15. Lateral pressure of foils as function of clearance.

33