Horpibulsuk, S., Bergado, D. T. and Lorenzo, G. A. (2004). Géotechnique 54, No. 2, 151–154 TECHNICAL NOTE Compressibility of cement-admixed clays at high water content S . H O R P I B U L S U K * , D. T. B E R G A D O † a n d G . A . L O R E N Z O † KEYWORDS: compressibility; laboratory tests; soil stabilisation INTRODUCTION The compressibility characteristics of cement-admixed clay are an important issue for the deformation analysis of composite soft clay. This paper attempts to analyse and assess the compressibility characteristics of cement-admixed clay. It has been found that cement content is the prime parameter governing the compression curve at post-yield state. The generalised compression line is newly developed for uncemented and induced cemented clays. The compression curve of cement-admixed clay can simply be predicted by employing this generalised compression line together with initial void ratio and yield stress. The prediction method is useful not only for quick determination of compression curve with acceptable error, but also for examination of the test results. COMPRESSION CHARASTERICTICS Nagaraj & Srinivasa Murthy (1986) analysed the compression lines of uncemented clays. They showed that the compression lines can be normalised by the void ratio at liquid limit, eL . The normalised compression line was called the intrinsic state line (ISL), represented by the following equation: e ¼ 1:23
0:276 log ó v (1) eL Based on extensive data on uncemented (remoulded) clays, Burland (1990) has proposed the intrinsic compression line (ICL) for assessing the in-situ state of natural clays. The ICL is expressed by the following equation: where x ¼ log óv . The void index is defined as e
e100 Iv ¼ Cc (2) 3000 (3) where e100 is the void ratio corresponding to ó9v ¼ 100 kPa of uncemented clays, and Cc is the compression index. e100 and Cc can be approximated in terms of eL using the following equations: e100 ¼ 0:190 þ 0:679eL
0:089e2L þ 0:016e3L Cc ¼ 0:256eL
0:04 (4) (5) It has been found that the ISL proposed by Nagaraj & Manuscript received 28 February 2002; revised manuscript accepted 16 December 2003. Discussion on this paper closes on 1 September 2004, for further details see p. ii. * School of Civil Engineering, Suranaree University of Technology, Thailand. † School of Civil Engineering, Asian Institute of Technology, Thailand. 151 LABORATORY INVESTIGATION This investigation is in two parts: studying the compression behaviour of cement-admixed clay, and assessing its compression line. The first part is investigated by testing two clays: Bangkok and Ariake clays. Bangkok clay was collected at the campus of the Asian Institute of Technology, Bangkok, Thailand. Ariake clay was collected in Fukudomi town, Saga, Japan. The basic properties of both clays are shown in Table 1. The water contents of both clays varied, with a liquidity index of about 1.0–3.0. The clays were thoroughly mixed with ordinary Portland cement at cement content, Aw , of 3–15%, in a soil mixer for 10 min, as recommended by Horpibulsuk (2001) and Miura et al. (2001). The uniform paste was then transferred to oedometer rings, taking care to prevent air entrapment. The total sample preparation time was less than 45 min, which is less than the initial setting time of the cement (about 2 h). All samples along with the rings were wrapped in vinyl and stored in a humidity-controlled room at constant temperature Yield stress in K0-consolidation, σy: kPa I v ¼ 2:45
1:285x þ 0:015x 3 Srinisava Murthy (1986) and the ICL proposed by Burland (1990) basically reflect the same response. Because of the cementation, cemented clays can be stable at high water content compared with uncemented clays at the same consolidation pressure, and they can exhibit higher yield stress and unconfined compressive strength. As the magnitude of the yield stress, óy , and unconfined compressive strength, qu , depend on the degree of cementation (bond strength), it appears logical to relate these two parameters. The relationship óy ¼ Cqu is generated, based on the test results for the following cement-admixed clays: Bangkok clay (Bergado & Lorenzo, 2001), Tokyo clay (Terashi et al., 1979) and Ariake clay (Horpibulsuk, 2001), as shown in Fig. 1, where C is constant. It has been found that C is dependent upon the clay type, and varies from 1.4 to 2.2. 2500 2000 σy ⫽ 2·2qu σy ⫽ 1·70qu 1500 |r | ⫽ 0·90 1000 Bangkok clay σy ⫽ 1·4qu 500 Tokyo clay Ariake clay 0 0 500 1000 1500 Unconfined compressive strength, qu: kPa 2000 Fig. 1. Relationship between unconfined compressive strength and yield stress in K0 -consolidation of cemented clays HORPIBULSUK, BERGADO AND LORENZO 152 Table 1. Physical properties of soft Bangkok and Ariake clays Properties Liquid limit: % Plastic limit: % Plasticity index Natural water content Grain size distribution: % Clay Silt Sand Total unit weight: kN/m3 Bangkok clay Ariake clay 103 43 60 76–84 125 60 65 120–130 69 28 3 14.3 55 44 1 14.1 account. However, there are two unknown parameters required to obtain this line, which are Cc and e100 . As e100 can be written in terms of Cc (equations (4) and (5)), it is possible to plot e/e100 against vertical pressure, óv , for uncemented clays, as shown in Fig. 3. The data for uncemented and cemented clays are based from the results of Burland (1990) and Nagaraj et al. (1998). This line is designated as the generalised compression line (GCL), and determined by the following equation: e e100 (20  28C). Oedometer tests were carried out after 28 days of curing. In addition to the above-mentioned laboratory test, test results for cement-admixed Tokyo clay (Porbaha et al., 2000) have been taken to verify the proposed method of predicting the compression line. TEST RESULTS AND DISCUSSION Typical compression behaviour is shown in Fig. 2 for cement-admixed Bangkok and Ariake clays under the highpressure oedometer. The compression is negligible pre-yield (of the order of 1% strain). This is due to the resistance from the cementation bond. Beyond the yield stress, óy , there is sudden compression of relatively high magnitude. At this stage, the compression index, Cc , is practically linear even as the vertical pressure increases. The figure also shows the influence of cement content on the compression behaviour at post-yield state. The post-yield compression behaviours of all mixtures with the same Aw converge on a single straight line on a plot of e against log óv , irrespective of clay water content. ASSESSMENT OF COMPRESSION LINE From the test results, it is notable that, for a particular clay admixed with cement, the compression index is practically constant, depending only on cement content, irrespective of the clay water content. Thus it is possible that the ICL can take the effect of clay type and cement content into ¼ 1:848
0:422 log ó v (6) with a correlation coefficient of 0.989, indicating the excellent correlation for 10 kPa , óv , 3000 kPa. In the absence of laboratory compression test data on cement-admixed clay, e100 can still be predicted if a value of vertical pressure and its corresponding void ratio are known. The method to determine e100 and to assess the (e, log óv ) of cemented admixed clays is as follows: (a) Determine the unconfined compressive strength, qu , and the void ratio, e0 . (b) Based on the relationship between yield stress and unconfined compressive strength (óy ¼ Cqu ), determine the yield stress. (c) Determine e100 by substituting void ratio and yield stress in equation (6). (d ) Draw (e, log óv ) at post-yield state by employing the GCL. (e) Assuming that there is no compression at pre-yield state, draw the horizontal line from e0 to the (e, log óv ) line at post-yield state. The predicted curves (Fig. 4) were drawn following steps (c)–(e), as the yield stress, óy , and initial void ratio, e0 , are available. It is found that the predicted and experimental curves are in good agreement within acceptable error. Furthermore, once the compression line of a particular cement content is drawn, the yield stress corresponding to any clay water content of the cement-admixed clays can be determined by drawing a horizontal line from the initial void ratio to the compression line at post-yield state. 6 6 5 wn 5 Void ratio, e Void ratio, e 4 3 2 180% 4 wn ⫽ 130% 150% wn ⫽ 160% 1 250% 3 wn ⫽ 190% 120% wn ⫽ 200% 0 100 101 102 103 Vertical pressure, σv: kPa (a) 104 2 100 101 102 103 Vertical pressure, σv: kPa (b) Fig. 2. Compression characteristics of cement-admixed clays cured for 28 days: (a) Bangkok clay, Aw clay, Aw 15% 104 5%; (b) Ariake COMPRESSIBILITY OF CEMENT-ADMIXED CLAYS AT HIGH WATER CONTENT 4·0 Uncemented clays Sail Argile Plastique Bangkok clay Vienna Magnus clay 3·0 Void ratio, e 153 Kleinbelt Ton Black cotton London clay Wiener Tegel Lower Cromer Till 2·0 1·0 0 5·0 Void ratio, e 4·0 3·0 Cemented Bangkok clay wn ⫽ 200%, Aw ⫽ 3% 2·0 wn ⫽ 200%, Aw ⫽ 5% 1·0 wn ⫽ 200%, Aw ⫽ 10% wn ⫽ 200%, Aw ⫽ 15% Cemented Ariake clay 0 wn ⫽ 180%, Aw ⫽ 15% 1·6 e/e100 1·2 0·8 e/e100 ⫽ 1·848 ⫹ 0·422 log σv 0·4 |r| ⫽ 0·989 0 100 101 102 Vertical stress, σv: kPa 103 104 Fig. 3. Generalised compression line for uncemented clays and induced cemented Bangkok clay CONCLUSION This paper deals with the analysis and assessment of the compressibility of cement-admixed clays. It is found that, for a given clay admixed with cement, the compressibility at post-yield state is governed mainly by the cement content, irrespective of clay water content. In order to take the effect of the clay type into account, the generalised compression line is introduced, based on the results for uncemented and cemented clays. This line is useful for quick determination of the compression curves of cement-admixed clays at any cement content and water content immediately after the unconfined compression test is completed. 3·0 Void ratio, e 2·5 2·0 1·5 5% cement 10% cement 1·0 15% cement Predicted 0·5 100 101 REFERENCES 102 103 Vertical stress, σv: kPa 104 105 Fig. 4. Predicted and experimental (e–log óv ) curves for cementadmixed Tokyo clay Bergado, D. T. & Lorenzo, G. A. (2001). Recent developments of ground improvement in soft Bangkok clay. Proceedings of the international symposium on lowland technology, Saga, Vol. 1, pp. 17–26. Burland, J. B. (1990). On the compressibility and shear strength of natural clays. Géotechnique 40, No. 3, 329–373. 154 HORPIBULSUK, BERGADO AND LORENZO Horpibulsuk, S. (2001). Analysis and assessment of engineering behavior of cement stabilized clays. PhD dissertation, Saga University, Japan. Miura, N., Horpibulsuk, S. & Nagaraj, T. S. (2001). Engineering behavior of cement stabilized clay at high water content. Soils Found. 41, No. 5, 33–45. Nagaraj, T. S. & Srinivasa Murthy, B. R. (1986). A critical reappraisal of compression index equations. Géotechnique 36, No. 1, 27–32. Nagaraj, T. S., Pandain, N. S. & Narasimha Raju, P. S. R. (1998). Compressibility behaviour of soft cemented soils. Géotechnique, 48, No. 2, 281–287. Porbaha, A., Shibuya, S. & Kishida, T. (2000). State of the art in deep mixing technology. Part II: Geomaterial characterization. Ground Improvement 3, 91–100. Terashi, M., Tanaka, H., Mitsumoto, T., Niidome, Y. and Honma, S. (1979). Fundamental properties of lime and cement treated soils (2nd report). Report of Port and Harbour Research Institute 19, No. 1, 33–62 (in Japanese).