INTERNATIONAL SOCIETY FOR SOIL MECHANICS AND GEOTECHNICAL ENGINEERING This paper was downloaded from the Online Library of the International Society for Soil Mechanics and Geotechnical Engineering (ISSMGE). The library is available here: https://www.issmge.org/publications/online-library This is an open-access database that archives thousands of papers published under the Auspices of the ISSMGE and maintained by the Innovation and Development Committee of ISSMGE. 4/ 19 CONFINED COMPRESSION OF LOESS COMPRESSION AVEC ETREINTE LATERALE DU LOESS C1KATHE J1ECCA BE3 BOKOBOrO PAC111MPEHMH HARRISON KANT Ph. 0., Professor of Civil Engineerin g, the University of Iowa, Iowa City (USA!) SYNOPSIS. Effective stress paths for a loessial soil subject to collapse during confined compression have been determined from the results of a testing program consisting of (1) confined compression tests on natural samples of loess with initial water contents ranging from air-dry to saturation, (2) negative pore-water pressure meas­ urements to -300 psi during these tests, and (3) Retests in which the lateral stress ratio was measured for one-dimensional strain. Before collapse, K was found to average 0.23, an extremely low value for a loose soil, whereas after collapse, Kq increased to 0.5?, which is consistent with values for other soils. From the stress path interpretation of the results, it is demonstrated that the collapse mechanism of loess in confined compres­ sion and during wetting Is a shear phenomenon and subject to analysis in terms of effective stresses. INTRODUCTION pends also on the water content before wetting. Studies by Bally (1961) demonstrated large compressive strains This research has been undertaken to study and de­ upon wetting of air-dry loess but small strains for scribe, quantitatively, the mechanisms involved in the the same soil at an initial water content of 24%. collapse phenomenon in loess. Supporting data incor­ Similar observations have been made by others. porates two quantities which have not previously been measured in this context: the negative pore-water To provide the data for interpreting the behavior of pressure and the lateral stress ratio, both measured the soil, two special types of test6 were run: (1) during confined compression. confined compression tests in which the negative pore- water pressure was measured during the course of the Loess is a wind-deposited sediment transported from tests utilizing a specially constructed cell, and (2) the flood plains of glacial rivers. The natural, un­ a special series of tests in which the lateral stress disturbed loess is a loose, open-structured soil com­ ratio (the ratio of the horizontal to the vertical posed of silt particles separated by clay coatings or stress) was measured in a triaxial compression cell aggregates of clay particles (Larionov, 1965; Gibbs under a zero-lateral-strain condition (KQ-tests). and Holland, 1960). A typical midwestem loess has a clay content of 10% to 30%, with water contents from SOIL INDEX PROPERTIES 5% to 30% and densities from 70 pcf to 90 pcf (Sheeler, 1968). The significant properties in this study are The undisturbed soil samples were obtained from a road (1) the low natural density which permits the occur­ cut in a loess deposit on the Oakdale campus of the rence of large volume changes, (2) the bond strength University of Iowa, just west of Iowa City. Index provided by the clay coatings, and (3) the changes in properties for the Oakdale loess are listed in Table this bond strength that occur with changes in water 1. The natural dry density of the Oakdale loess is content. relatively high for loessial soils in Iowa. Densities in east-central Iowa from 80 pcf to 90 pcf have been The collapse of loess occurs when the stress between reported by Lyon, Handy, and Davidson (1954) and den­ particles exceeds the bond strength provided by the sities as low as 66 pcf have been measured. On the clay cdatings. This may be caused by an increase in basis of the properties in Table 1, the Oakdale loess stress due to an applied load, or by a decrease in is described as a relatively dense, silty loess of low strength due to swelling and softening of the clay plasticity. binder after wetting. The loss of strength on wetting has been recognized for some time (e.g. Holz and Gibbs, SAMPLING AND PREPARATION OF TEST SPECIMENS 1951). A recent description of this behavior has been given by V. G. Berenzantzev, et al. (1969). In this Hand-carved blocks of soil, 8 in. by 10 in. by 10 in. paper the subsidence deformation below a foundation is were removed from the test pit, wrapped in plastic, shown to occur in a zone where the shear strength has and transported directly to the laboratory. Upon ar­ decreased as a result of wetting to the level of the rival at the laboratory, the blocks were divided into existing shear stresses. The amount of settlement de­ smaller samples, sealed and stored in a moist room 115 4/ 19 TABLE 1 SOI L I NDEX PROPERTI ES in the p a rtia lly saturated specimens of loess w hile the s o il was undergoing compression under standard consolidation test loading. O a k d a le L o e s s The confined compression ce ll is shown in Fig. 1. T h e three main parts are the base, the cylinder, and th e Liquid Lim it 2 7 top. The base and top are machined from stainless P lastic Lim it 2 3 steel and the cylinder Is a section of 5-inch diame t e r P la sticity Index 4 aluminum pipe. A fine-grained porous stone w ith a S pecific G ravity 2 . 7 2 rated bubbling pressure of 225 psi is sealed into t h e Percentage Clay base. Two sm all-diam eter ports enter through the bo t ­ L e s s t h a n 0 . 0 0 5 Dim 1 7 tom of the base to the lower surface of the stone. L e s s t h a n 0 . 0 0 2 mm 1 3 One port is fitte d w ith a valve and a supply of dea i r ­ A c t i v i t y 3 0 . 5 0 ed water confined in a control cylinder capable of N atural Dry Density pcf forcing measured quantities of water through the st o n e R a n g e 90.4 to 92.5 into the so il. A 150-psi pressure transducer is mou n t ­ A v e r a g e 9 1 ed at the other port. The ce ll top encloses the top o f N atural Void Ratio the cylinder. A 1/2-inch diameter loading ram is R a n g e 0.800 to 0.940 guided through the top by a ball-bushing and teflon A v e r a g e 0 . 8 6 1 seals provide a low -friction seal. The top is con­ N atural Water Content, % nected to the base by four bolts extending from the R a n g e 21.2 to 22.9 base to the top outside the cylinder. Accurate alig n ­ A v e r a g e 2 2 ment of the loading ram is assured by the machined cylinder ends. aA ctivity = p la sticity index/(% 0.002 ram clay - 5%) The so il specimen is carved in to a standard consoli d a ­ (Seed, H.B. et a l, 1962) tion ring, 2.5 in. in diameter by 0.75 in. thick. T h e ring is located centrally in the hase by three lugs u n til needed for testing. and held in place by a co lla r. The load from the lo a d ­ ing ram is transm itted to the so il by the loading c a p The confined compression testing program required and stone. The ring, co lla r, and loading cap are stru ctu ra lly undisturbed specimens w ith a range in standard consolidation c e ll components. water contents, nom inally 4%, 8%, 12%, 16%, 20%, 24 % , and 28%. Since the natural water content was about 22%, the specimens had to be wetted or dried to achieve the desired water contents before trim m ing t o test specimen size. The a lteratio n of the water content was accomplishe d in the follow ing manner. To achieve a water content o f 28%, the sample as stored at its natural water cont e n t was unwrapped and wetted by spraying the surface w i t h a measured quantity of water. The sample was then r e ­ wrapped and placed in a m oist chamber fo r several d a y s to perm it the dispersal of the water throughout the so il. This process was repeated u n til the water con ­ tent of the sample, estim ated from the sample's wet weight and its orig in al weight and water content, w a s the desired 28%. The sample was then trimmed into t h e consolidation ring for testing. The water contents o f the specimen and the trim m ings were compared to che c k the uniform ity of water distribution in the sample a n d in a ll cases the differences were less than 1%. Water contents below the natural water content were achieved by perm itting the surface of the sample to air-dry for several hours, during which tim e the sa m ­ ple was weighed periodically to determine the weigh t of water evaporated.