D. Wanatowski 1 and J. Chu 2 Effect of Specimen Preparation Method on the Stress-Strain Behavior of Sand in Plane-Strain Compression Tests ABSTRACT: Experimental results are presented in this paper to study the effect of specimen preparation method on the stress-strain behavior of sand in plane-strain compression tests. The data obtained from K 0 consolidation, drained, undrained and strain path tests conducted on medium loose specimens prepared by the moist-tamping (MT) and the water sedimentation (WS) methods are compared. The test data show that the plane-strain compression behavior of medium loose sand under K 0 , drained and strain-path controlled (including undrained) conditions is affected by the speci- men preparation method. Under K 0 conditions, the K 0 values obtained from the MT specimens are generally lower than those obtained from the WS specimens. Under drained conditions, more contractive behavior was observed for the MT sand. However, the failure stress ratio (or the failure friction angle) was not affected by the specimen preparation method. The data presented in this paper also illustrate that the compression behavior of medium loose sand in strain-path testing can be affected by the specimen preparation method. However, the differences in the stress-strain behavior will also depend on the strain increment ratio (d v /d 1 ) imposed on the specimens. In general, different behaviors of the moist-tamped and water- deposited specimens reflect the influence of soil fabrics on the stress-strain behavior of sand. KEYWORDS: sand, plane-strain, stress-strain behavior, sand fabric, pluviation, moist tamping Introduction Several preparation methods of granular soil specimens can be used in soil mechanics laboratories. Moist tamping (MT), water sedimentation (WS) (also known as water pluviation), and air plu- viation (AP) are among the most popular techniques. In the MT method, moist granular soil is deposited into a mold in a few layers and each layer is compacted using a small tamper. In the other two preparation methods (i.e., WS and AP), a granular material is poured into the mold, which is either empty (AP method) or par- tially filled with water (WS method). If necessary, the density of the specimens prepared by pluviation methods can also be increased by vibration or tamping. It has long been recognized that different preparation methods result in different fabrics of granular soils and, consequently, in dif- ferent stress-strain characteristics of reconstituted specimens (Oda 1972a, 1972b; Ladd 1974, 1977; Mulilis et al. 1977; Silver et al. 1980; Miura and Toki 1982; Kuo and Frost 1996; Frost and Park 2003; Yamamuro andWood 2004). Owing to this, a number of ex- perimental studies discussing the various effects of sample prepa- ration methods on the stress-strain behavior of granular soils have been reported in the past. However, the majority of experiments have been conducted under axisymmetric conditions using triaxial cells. Studies investigating the effects of different specimen prepa- ration methods on the stress-strain behavior of sand under more generalized stress conditions, such as plane-strain, are very rare even though several field problems (e.g., slopes, embankments, or retaining walls) cannot be approximated to axisymmetric condi- tions. This is partially due to the fact that plane-strain devices are not commonly available and plane-strain tests are more compli- cated to conduct than triaxial tests. As a result, our understanding of the effects of specimen preparation methods on the stress-strain be- havior of soil under plane-strain conditions is still very limited. The main objective of this paper is to study the effects of speci- men preparation methods on the stress-strain behavior of sand under plane-strain conditions. Several K 0 consolidated plane-strain compression tests conducted on medium loose sand under various drainage conditions were carried out. The results obtained from the plane-strain tests conducted on medium loose specimens prepared by the moist tamping and the water sedimentation methods are compared. The effects of specimen preparation methods on the measurement of K 0 values and the stress-strain behavior of sand in plane-strain tests are discussed. Previous Studies In the past, several investigators have studied the effect of specimen preparation methods on stress-strain behavior of granular soil. Among the first attempts to study the effects of the specimen prepa- ration methods on the behavior of reconstituted sand were those of Ladd (1974, 1977), who observed that the method of specimen preparation could have a significant effect on the cyclic shear strength of sand. Similar observations were made by Mulilis et al. (1977). They reported that different specimen preparation proce- dures significantly affected the liquefaction characteristics of sand in undrained stress-controlled cyclic triaxial tests. Mulilis et al. (1977) also observed that the MT specimens were more nonuni- form than the others. Therefore, in order to improve the uniformity of the MT specimens Ladd (1978) proposed an undercompaction procedure. In this method, the specimen is prepared using a number Manuscript received June 18, 2007; accepted for publication December 28, 2007; published online February 2008. 1 Lecturer, Nottingham Centre for Geomechanics, School of Civil Engineer- ing, The University of Nottingham, University Park, Nottingham NG7 2RD, United Kingdom, e-mail: dariusz.wanatowski@nottingham.ac.uk (correspond- ing author) 2 Associate Professor, School of Civil and Environmental Engineering, Nan- yang Technological University, 50 Nanyang Avenue, Singapore 639798, e-mail: cjchu@ntu.edu.sg Geotechnical Testing Journal, Vol. 31, No. 4 Paper ID GTJ101307 Available online at: www.astm.org Copyright © 2008 by ASTM International, 100 Barr Harbor Drive, PO Box C700,West Conshohocken, PA 19428-2959. 1