DISCUSSION

RESISTANCE O F SOIL STRUCTURE TO CONSOLIDATION*

It was a primary purpose of the paper under discussion to explore the influence of time or rate of strain on the effective stress-deformation characteristics of an ondisturbed sensitive clay. It had been argued previously (Crawford, 1964) that it is difficult and unrealistic to assess time effects from incremental loading tests unless, of course, the increments are very small and provide an almost continuous loading. It is not possible therefore to explain adequately the conflicting results quoted by Mr. 1.0, but some comments are in order.

Langer (1936) tested three soils ranging in consistency from soft to stiff. He found that the pressure-compression of the stiff clay only was affected by loading rate and that the compression curve was displaced to the right as stated by Mr. Lo. This stiff clay had a liquidity index of about 0.1, compared to a value of about 4.0 for the Lcda clay described in the present p1 LP p er. Langer's slowest test lasted for six months; the slowest test quoted by the author (94-19-4) lasted for five days. It would be surprising if the test results in these two cascs were in agreement.

Review of the test results of Hairmilton and Crawford ( 1959 ) shows that the total test time for twelve tests ranged from 3.2 to 16.5 days. The test of longest duration was loaded with a load increment ratio of 1/10. Up to about 2/3 of the preconsolidation pressure, loads were applied in increments of 10 minutes. r611e next increment was allowed to act for three days. Beginning just below the preconsolidation pressure, further load increments were applied daily, except that one load was allowed to remain for five days at a pressure a little greater than the preconsolidation pressure. Pieither of the long term loads on this specimen had much effect on the pressure-void ratio curve. This ability of the soil structure to resist a small load incrcn-nent for several days without appreciably affecting the pressmc-void ratio curve and the relatively minor influence of load increment ratio, together with the similar results of Casa- grancle and Fadum (1944), led the authors to the conclusion that the size of load increment u7as not a critical factor. This is in agreement with two of the three soils tested by Lnnger (1936). The limitations of incremental loading must now be explored.

Mr. Lo notes that if tllc available test results arc extrapolated to the very low field rates of strain the precoilsolidation pressure is very low. This illns- trates the danger of extrapolation and emphasizes the author's closing para- graph. Mr. Lo's final question is currently under study. In general, it can be said that for a particular test procedure specimens cut from blocks have a higher test preconsolidation pressure than those cut from tube samples, and it may be more than twice as high when the tube samples are not taken with qreat care.

"By C. M. CRAWFOHD, flak JOURNAL 2 , 2 ( May 1965 ) : 90-1 15. f Head, Soil hlechanics Section, Division of Building Research, National Research Council2

Canada.

98

Canadiar~ Geotechnical Journal, vol. 111, no. 2. Printed in Canada.

Can

. Geo

tech

. J. D

ownl

oade

d fr

om w

ww

.nrc

rese

arch

pres

s.co

m b

y U

nive

rsity

of

P.E

.I. o

n 11

/23/

14Fo

r pe

rson

al u

se o

nly.

DISCUSSION 99

Mr. Raymond accepts the major conclusion of the paper, that "the soil structure has an important time-dependent resistance to compression," but he resists the inference that such an effect would influence the computed per- meability. Perhaps the best proof that it does influence the computed perme- ability is provided by his Figure 2, which shows, at an average pressure of about 3.6 kg/cm2, a coefficient of permeability of 0.6 X 10-%m/sec., 3.2 X lo-' cm/sec., and 31X cm/sec., a variation of fifty-fold, depending on the load increinent ratio. Clearly the actual permeability would not vary greatly among three specimens of the same soil at almost the same void ratio, and it follows that the computed values are a function of the test procedure.

Care should be taken to avoid confusion between "computed" and "actual" values of permeability. The tabled values from Hamilton and Crawford (1959 j used by klr. Raymond in his Figures 1, 2 and 3 were quoted to illustrate the difficulty of computing satisfactory values. The merit of publishing further detailed results of this nature is questionable. It can be said, however, that more consistent results are obtained from tests at one load increment ratio. In the incremental tests reported in Table I of the paper the computed per- meability ranged from 0.4 x 10-Vo 2.6 x cm/sec.

The apparent contradiction between test results given in the paper and previously published results is questioned by Mr. Raymond. The previously published results are "computed" coefficients of permeability and compressI- bility and these are obviously quite unreliable. It has been adequately demonstrated that as tlie load increment ratio is decreased the amount of secondary consolidation becomes proportionately greater, and that at very low increments (or under continuous loading) there is virtually no hydrodynamic influence. As this coildition is approached it is not realistic to compute per- meability or related coefficients from tlle Terzaghi classical theory. The author prefers not to draw conclusions from the manipulation of these numbers.

The independent test results reported by Mr. Raymond supplement rather than discuss the paper. It would have been more useful to have had effective stresses rather than total stresses plotted, because his Figures 7, 8, 9, and 10 might be confused with the usual c - log p curves. In particular, his assump- tion that preconsolidation pressure is related to the total axial stress is open to question.

Mr. Raymond's belief that "secondary consolidation occurs during priinary consolidation and the two processes are to some extent interrelated" is a common belief. Ally suggestion that the resistance of the soil structure is not acting during primary consolidation would excite great debate.

It is for this reason that the author suggests more research on compression of clay at low or negligible pore pressures. Leonards and hltschaeifl (1964) went to a great deal of trouble to freeze-dry clays so that structural deformation could be studied without hydrostatic influences. At field rates of con~pression this can be studied much more easily by slow continuous loading of natural soils without changing the physico-chemical nature of the soil.

For many clays the author would postulate that the greatest strtucturnl resistance to effective stresses (i.e., the highest preconsolidation pressure) will occur with relatively rapid loading. This is based on results quoted in the

Can

. Geo

tech

. J. D

ownl

oade

d fr

om w

ww

.nrc

rese

arch

pres

s.co

m b

y U

nive

rsity

of

P.E

.I. o

n 11

/23/

14Fo

r pe

rson

al u

se o

nly.

100 CANADIAN GEOTECEIEjICAL JOURNAL

paper and on the generally accepted belief that the compressibility of the soil skeleton is time-dependent. It must be admitted that Leonards and AltschaeH (1964) did not observe this on freeze-dried clay. At very slow rates of loading another factor has to be considered: the Terzaghi (1941) hypothesis that the bond between particles will sustain a low compression index until broken by rapid loading or by disturbance of the structure. This has been demonstrated by Leonards and Altschaeffl (1964). At intermediate rates of loading the lowest measured preconsolidation pressure may be expected owing -to a decrease in both influences. Further research on the deformation of natural clays under known effective stresses at various rates of application is required to reveal the appropriate values for actual cases.

CASAGRANDE, A., and FABUM, R. E., 1944. Applications sf soil mechanics in designing building foundations. Tram. ASCE 109 : 383.

CRAUTOBB, C. Be, 1984. Interpretation sf the consolidation test. 1. Sail Mech. and Found. Biu. ASCE 90, SM 5: 87-1Q2.

HAMILTON, J. J., and CRAWFORD, 6. B., 1959. Improved detern~ination of precorisslidation pressure of a sensitive clay. ASTM STP 254: 254-71.

EANGER, K., 1938. Influence of speed of loading increment on the ressure void ratio diagram of undisturbed soil samples. h e . 1st Int. Conf. Soil MecK. and Found. Eng. ( Cambridge, Mass. ) 2: 116-20.

LEONARDS, G. A., and ALTSCHAEFFL, A. G., 1964. The compressibility of clay. 3. Soil 1Mec72.. and Found. Div. ASCE 90: 133-55.

TERZAGKI, K., 11941. Undisturbed clay samples and undisturbed clays. J. Boston Soc. Civil Eng. 28,3: 45-65.

SETTLEMENT O F A MAT FOUNDATION ON A THICK STRATUM O F SENSITIVE CLAY"

The authors have presented a well-documented case record comparing calcu- lated versus observed settlements for a major structure on a thick stratum of sensitive clay. The following comments offer some suggestions that may aid the assessment of field conditions in future studies of this nature. An alternate explanation for the observed settlements is also proposed.

1. Owing perhaps to air coming out of solution during unloading, the magnitude of the recompression index (C,) obtained from the rebound portion of the pressure versus void ratio curve may be overestimated. Tests on a glacial lake clay using standard and back-pressure oedomcters gave values of C, equal to 0.03 and 0.02, respectively. The latter value appeared to be more com- patible with the observed settlements. If settlemen& due to recompression are significant, tests using the back-pressure oedometer (Lowe et a.l., 1964) should be considered.

2. The use of small load increments is helpful whenever the determination

*L. Gasagrmde, B. Firing, G. Sckoof, and E. 1%'. John Turcke, this JOWWAL 2, 4 (Nov. - - 1965 ) .

+Professor of Soil Mechanics, Purdue University, Lafayette, Ind.

Can

. Geo

tech

. J. D

ownl

oade

d fr

om w

ww

.nrc

rese

arch

pres

s.co

m b

y U

nive

rsity

of

P.E

.I. o

n 11

/23/

14Fo

r pe

rson

al u

se o

nly.