Expansive soil has been well recognised as a problematic soil due to its swelling and shrinkage behaviour during wetting and drying cycles. The swelling pressure developed upon wetting is often investigated by placing an initially unsaturated soil specimen into an oedometric cell modified to restrain the specimen’s volume change. Water is then added to the specimen and the increase in vertical stress due to wetting is then measured. In this study a K0 triaxial cell, originally designed for saturated soil testing, is modified and used to investigate the development of swelling pressure. The cell allows not only vertical but also radial stresses to be measured during the wetting process. An initially unsaturated residual soil specimen is assembled into the K0 cell where wetting of the specimen is allowed from the bottom drainage line. Allowing neither vertical nor radial strain during the wetting, vertical and radial swelling stresses developed during the process were recorded. Important experimental issues related to the tests are discussed aiming to improve the result quality.
An expansive soil swells upon wetting and shrinks when subject to drying. This type of geomaterial has been well recognised as problematic. Lightly loaded structures built on expansive soils heaved when subject to wetting (Jennings & Kerrich, 1962; Fredlund & Rahardjo, 1991). Pender (1996) reported that retaining walls supporting expansive soils could tilt noticeably when subject to seasonal wetting and drying. In extreme cases, the walls could even collapse upon prolonged raining (Thomas, 2008; Ozer et al. 2012). Furthermore, deformation caused by soil swelling was identified as an important failure mechanism in expansive soil slopes (Ng et al. 2003; Qi & Vanapalli 2016). Conventionally, swelling pressure is often evaluated using oedometric cells while specimens are inundated under constant volume. It is anticipated that swelling stress measured using this approach would have been underestimated due to side friction of the oedometric ring (Al-Shamrani & Dhowian, 2003; ASTM D4546-03). This paper presents a modified triaxial cell which allows wetting of soil specimen under K0 condition. Preliminary results are presented and important experimental issues related to the testing are highlighted and discussed.
2 STUDIED SOIL
Block soil samples were retrieved from a site near Newmarket, Auckland. The orange-brown residual soil is a weathering product of sandstones belonging to the East Coast Bays Formation (Kermode 1992). Natural water content of the soil was 60±2%. Table 1 summarises index properties of the soil obtained in accordance with NZS 4402:1986. The soil contains significant amount of clay fraction, exhibits a high liquid limit, and is referred to as CH according to USCS. The soil is considered to be highly expansive based on its liquid limit, plasticity index and activity according to Chen (2012) and van der Merwe (1964). X-ray diffraction (XRD) and X-ray fluorescence (XRF) analyses were carried out to determine the mineralogy of the studied soils. The soil was found to contain 64% quartz, 21% montmorillonite, 12% goethite and 3% of oxides with various cations.
Table 1: Index properties of the studied soil
3 TEST EQUIPMENT AND SPECIMEN PREPARATION
Soil specimen in the oedometer cell has diameter and height of 76 mm and 19 mm respectively. The oedometer ring with specimen was carefully placed into the cell. Then, a vertical stress of 25 kPa was applied. The specimen was inundated under constant volume and the development of vertical swelling stress was recorded. It was found that the change of swelling pressure became negligible 24 hours after the inundation.
3.2 Triaxial Cell
A K0 triaxial cell (Campanella & Vaid 1972; Hettiaratchi et al., 1992; Meyer, 1997) was used to investigate the development of vertical and radial swelling pressures of an initially unsaturated triaxial specimen when subject to wetting. The specimen has dimensions of 120 mm height and 60 mm in diameter. Upper drainage line of the specimen was connected to an open-end burette tube partially filled with water maintaining at atmospheric pressure. Saturation of the specimen could be easily visualised when the out-coming fluid changed from air bubbles to yellowish water containing fines of the studied soil. The K0 triaxial cell has a rigid chamber wall. By restraining the flow of cell fluid, a K0 condition could be imposed onto the specimen. By measuring the pressure of the cell fluid, the change of radial stress of the specimen could be monitored. Vertical deformation of the specimen was not allowed and the development of vertical swelling pressure was recorded by an internal load cell.
After the specimen had been assembled into the triaxial cell, vertical and radial pressures of 25 kPa and 10 kPa respectively were applied to the specimen. A small back pressure was then applied at the bottom drainage line which aimed to wet the specimen from bottom to top. Fig. 1 shows a schematic diagram and a photo of the setup of K0 triaxial cell. Note that all the K0 triaxial wetting tests were carried out at a temperature-controlled room with temperature varying within 1 °C.
Figure 1: Setup of the K0 triaxial cell: (a) schematic diagram; (b) photo
3.3 Specimen Preparation
This paper only presents preliminary results obtained from compacted specimens. Soil was compacted at 22% water content in a standard proctor compaction mould to a target density of 1.6 Mg/m3. Each proctor sample was then trimmed into two specimens; one for oedometer and another one for K0 triaxial testing.
4.1 Swelling in oedometric cell
Constant volume oedometer test provided a relatively simple experimental means to evaluate the development of vertical swelling pressure. Fig. 2 shows the test result. It was found that the vertical swelling pressure was rapidly developed within the first 3 hours and became essentially stable after 10 hours. The oedometer specimen showed a final vertical swelling pressure of about 370 kPa upon inundation.
Figure 2: Development of vertical swelling pressure in constant volume oedometer testing
4.2 Swelling in K0 cell
Two wetting tests, denoted by K1 and K2, were conducted in the K0 triaxial cell. Fig. 3 shows the development of vertical and radial swelling pressure with time. It can be seen that both vertical and radial stresses in K1 developed very slowly during the first 5 hours yet the rate increased noticeably afterwards. It is because the back pressure first adopted in K1 was too small to wet the specimen effectively. The back pressure was then increased to 7 kPa to improve the effectiveness of soil wetting. Since then a more noticeable increase in both the vertical and radial stresses can be seen. The test was terminated after 6 days of testing though a mild increasing in the vertical swelling stress was observed. It was because K1 only served as a preliminary test to investigate the effect of back pressure on the rate of wetting. In K2, a 7 kPa back pressure was applied since the beginning of the test and the development of swelling pressure resembled a better trend and a faster rate as compared to K1. Test K2 was terminated after 3.5 days as both the vertical and radial swelling stresses became essentially steady. Besides, water was found coming out from the top drainage line which appeared to be indicting that the soil has reached a very high degree of saturation. At the end of the K2, the increase in vertical and radial stress due to wetting induced swelling was about 500 kPa and 100 kPa, respectively. The vertical swelling stress was larger than that obtained using the oedometer (which was 370 kPa). It is anticipated that friction along the oedometer ring may account for the difference in vertical swelling stress.
Figure 3: Development of vertical and radial swelling pressure in the K0 triaxial cell: (a) Test K1; (b) Test K2
4.3 Observations and lessons learned
At the end of the K2 test, the specimen was cut into 3 slices representing soils at the top, middle and bottom of the specimen. Oven drying of the slices provided clear evidence that the water content at the end of the wetting test varied along the height of the specimen. Soil at the upper part of the specimen was not yet fully saturated. Indeed, fully saturating the specimen was not a trivial task and required a good control of back pressure and time. Furthermore, whether the cell chamber has been fully filled up with de-aired water or not is crucial to the determination of swelling pressure. Even a small air void inside the cell fluid chamber allows a specimen to increase its volume upon wetting and therefore gives noticeably lower swelling pressures. To sum up, the swelling pressures reported in K1 and K2 should have underestimated the swelling pressure that the soil can actually develop.
Experimental studies were undertaken to investigate the development of swelling pressure of a residual soil found in Auckland when subject to wetting. Key findings of the studies were:
- The proposed K0 triaxial setup can be used to investigate the development of swelling pressure in both vertical and radial directions when an initially unsaturated soil is subject to wetting.
- Higher vertical swelling stress was reported from the K0 triaxial cell as compared to results obtained based on conventionally used oedometer setup. Side friction along the oedometer ring probably attributed to the difference.
- The swelling pressures reported in the two preliminary K0 triaxial tests should have been underestimated due to the difficulties in fully saturating the soil and the cell fluid chamber of the K0 triaxial cell.
Al-Shamrani, M.A. & Dhowian, A.W. (2003). Experimental study of lateral restraint effects on the potential heave of expansive soils. Engineering Geology, 69(1), 63-81.
ASTM (2003). Standard Test Methods for One-Dimensional Swell or Collapse of Soils, D4546-03, ASTM International, West Conshohocken, PA.
Campanella, R.G., & Vaid, Y.P. (1972). A simple K0 triaxial cell. Canadian Geotechnical Journal, 9(3), 249-260.
Chen, F.H. (2012). Foundations on expansive soils. Elsevier.
Fredlund, D.G. & Rahardjo, H. (1991). Stress state variable approach to the prediction of heave. International Workshop on Clay Swelling and Expansive Soils, Cornell University, Ithaca, New York, USA, 12-16.
Hettiaratchi, D.R.P., O’Sullivan, M.F. & Campbell, D.J. (1992). A constant cell volume triaxial testing technique for evaluating critical state parameters of unsaturated soils. European Journal of Soil Science, 43(4), 791-806.
Jennings, J.E.B. & Kerrich, J.E. (1962). The heaving of buildings and the associated economic consequences, with particular reference to the Orange Free State Goldfields, The Civil Engineer in South Africa, 4(11), 221-248.
Kermode, L.O. (1992). Geology of the Auckland urban area. Institute of Geological & Nuclear Sciences Ltd, Lower Hutt, New Zealand
Meyer, V. (1997). Stress-strain and strength properties of an Auckland residual soil. Doctoral Thesis. The University of Auckland, New Zealand.
Ng, C.W.W., Zhan, L.T., Bao, C.G., Fredlund, D.G. & Gong, B.W. (2003). Performance of an unsaturated expansive soil slope subjected to artificial rainfall infiltration. Géotechnique, 53(2), 143-157.
Ozer, M., Ulusay, R. & Isik, N. (2012). Evaluation of damage to light structures erected on a fill material rich in expansive soil. Bulletin of Engineering Geology and the Environment, 71(1), 21-36.
Pender, M.J. (1996). Aspects of Geotechnical Behaviour of Some NZ Materials. Proceedings of the 7th Australia New Zealand Conference on Geomechanics: Geomechanics in a Changing World. Adelaide, Australia. 21-39.
Qi, S. & Vanapalli, S.K. (2016). Influence of swelling behavior on the stability of an infinite unsaturated expansive soil slope. Computers and Geotechnics, 76(6), 154-169.
Thomas, M.G. (2008). Impact of lateral swell pressure on retaining structure design using expansive cohesive backfill. Master Thesis, University of Texas, Arlington, USA.
Van Der Merwe, D.H. (1964) The prediction of heave from the plasticity index and percentage clay fraction of soil. South African Institute of Civil Engineers, 6, 103-107.