Академический Документы
Профессиональный Документы
Культура Документы
COMMON GROUND 07
Identification of possible new failure mechanisms for the collapse of tailings dams
Jian Chu Nanyang Technological University, Blk N1, 50 Nanyang Ave, Singapore 639798 Wing Kai Leong Victor Li & Associates Ltd, Rm 1103, Kowloon Investment Co. Building, 2-12 Bute Street, Hong Kong A. Balasubramaniam School of Engineering, Griffith University Gold Coast Campus, PMB 50 Gold Coast Mail Centre, Queensland 9726, Australia Sik-Cheung Robert Lo University College, University of New South Wales, Australian Defence Force Academy, ACT 2600, Australia
Keywords: granular soil, failure, laboratory test, liquefaction, stability, tailings ABSTRACT Static liquefaction has been considered as one of the most common failure mechanisms for tailings dams. However, for tailings with relatively high permeability, the assumption of an undrained condition becomes questionable under static loading conditions. In fact, many granular soil or tailings at its in-situ density may not liquefy under static, undrained conditions. In this paper, the problems with the current design approaches for tailings dams are identified. Laboratory test results are presented to demonstrate that granular soil can become unstable under drained or nonundrained conditions. Based on the testing data, two possible new failure mechanisms for tailings dams are proposed. 1 INTRODUCTION
Slope failure or landslide is still one of the common geotechnical hazards. This includes failures of tailing dams of mine waste, mineral sands, or municipal solid waste. There have been a number of failure cases of tailing dams in recent years. Examples in Australia and elsewhere have been given by Eckersley (1985), Davies et al. (2002), Blight and Fourie (2005) and Wright (2006). As explained by Davies et al. (2002), the failure of loose granular soil slopes or cohesionless tailing dams is often considered to be triggered by instability or static liquefaction occurring under undrained conditions, as shown in Fig. 1. Two approaches, the effective stress analysis (ESA) and the undrained strength analysis (USA), have been suggested (Martin and McRoberts, 1998).
F
(a)
(b)
Figure 1 Response of loose, saturated cohesionless tailings under monotonic and cyclic loading (after Davis et al. 2002) According to Martin and McRoberts (1998), in the ESA method, effective stresses during shear are assumed unchanged from those that existed immediately prior to the onset of shear. In other words, failure is calculated as the failure shear stress corresponds to the in-situ effective stresses using the effective failure envelope, at point F, as marked in Fig. 1a. This method may be applicable to dense, dilative soil where the excess pore pressure generated during shear is very small or negative.
590
COMMON GROUND PROCEEDINGS 10TH AUSTRaLIa NEW ZEaLaND CONFERENCE ON GEOMECHaNICS BRISbaNE
However, the use of post-liquefaction strength may be problematic. Firstly, granular soil or tailings have to be very loose to exhibit a contractive behavior as shown in Fig. 1. Many granular soil or tailings at its in-situ density may not liquefy under static, undrained conditions. Secondly, the value Sus cannot be defined properly. As the post-liquefaction strength, Sus, should be measured from a test where the soil liquefies. However, once a soil liquefies, the sample will collapse suddenly from the point where liquefaction is initiated, e.g., the peak point P in Fig. 1 (Chu and Leong, 2001). Thus the stresses and strains in the post-peak region cannot be measured properly. The so-called post-liquefaction strength is, in fact, measured as the post-strain softening behaviour, where the specimen does not collapse and only the shear stress is reduced gradually till the ultimate state, which may or may not be constant. It has been demonstrated experimentally by Chu and Leong (2001) that the post-peak behavior during strain softening and instability (or liquefaction) is quite different. Therefore, the Sus obtained from tests with strain softening may not be relevant to soil or tailings that liquefy. Thirdly, the stress state after strain softening is on the critical state line (CSL), as shown in Fig. 1. The study of Chu et al (2003) have shown that the stress ratio at which instability occurs is generally smaller than the stress ratio at critical state for loose sand. Furthermore, for granular soil or tailings with relatively large permeability, the assumption of undrained under static loading conditions may not be reasonable. There are cases where instability or static liquefaction occurred under essentially drained conditions. In a recent reanalysis of the Wachusett Dam failure in 1907, Olson et al. (2000) concluded that the failure was mainly triggered by static liquefaction that occurred under completely drained conditions. Through laboratory model tests, Eckersley (1990) observed that the pore water pressure increase in the gentle granular soil slope was a result, rather than the cause of flowslide. In other words, the flowslide took place under a drained condition. In these cases, failures appear to be triggered by a mechanism similar to static liquefaction although the drainage condition prior to failure is not undrained. If this is the case, equilibrium analysis using the peak friction angle may not be applicable either.
(a) (b) Figure 2 Failure mechanisms identified by NRC (1985): (a). Mechanism B for the situation where void redistributes within a globally undrained sand layer; (b). Mechanism C for the situation where failure is induced by spreading of excess pore pressure with global volume changes. Instability or liquefaction of sand may also occur under other non-undrained conditions (Chu et al. 1993; Chu and Leong 2001). Failure mechanisms related to a redistribution of void ratio within a globally undrained sand layer (Fig. 2) and spreading of excess pore pressure with global volume changes (Fig. 2) have been envisaged by NRC (1985) as Mechanisms B and C, respectively. The possibility of dilating behaviour of soil masses prior to slope collapse is also observed in several case studies. Been et al. (1988) argued that the Nerlerk berm failure case might have occurred for dilative sand which lies below the steady state line. Several other cases of flow slide failure in dilative sand have also been presented by Been et al. (1988). Therefore, in addition to liquefaction under undrained conditions, there are other types of failure mechanisms that control the stability of granular slope. However, instability (as defined here as a behaviour in which large plastic strains are generated rapidly due to the inability of a soil element to
SLOPE AND STABILITY WALLs
591
COMMON GROUND 07
However, for loose, contractive soil where positive excess pore pressure is generated, this method is unconservative, as failure occurs at point Sus or P, not at point F, as shown in Fig. 1a. In the USA method, the undrained shear strength is defined as the post-liquefaction strength, Sus, as shown in Fig. 1a. As elaborated by Martin and McRoberts (1998), for contractive materials, design analyses must include both undrained strength analysis (USA) and effective stress analysis (ESA), with design controlled by the analysis type giving the lowest factor of safety. For dilative or fully drained materials, only ESA is required.
COMMON GROUND PROCEEDINGS 10TH AUSTRaLIa NEW ZEaLaND CONFERENCE ON GEOMECHaNICS BRISbaNE
COMMON GROUND 07
sustain a given stress) under other than undrained conditions, such as fully drained conditions, has seldom been studied. The mechanisms of instability of dilating sand have not been properly investigated either. A research program on the instability behaviour of granular soil has been carried out at the Nanyang Technological University, Singapore, in the past years. The objective of this paper is to illustrate two new failure mechanisms using experimental data, that is, the instability of granular soil under drained condition and the instability of granular soil under nonundrained (i.e., other than undrained) conditions. 2 INSTABILITY OF GRANULAR SOIL UNDER DRAINED CONDITIONS
It is well known that when a loose sand specimen is sheared along an undrained path, an effective stress path as typically shown in Fig. 3a will be obtained. Point A is the peak of effective stress path. If the test is conducted under a deformation-controlled condition, strain-softening behaviour that is characterized by a reduction in deviator stress will manifest. On the other hand, if the test is conducted under a load-controlled condition, the specimen will become unstable at point A. This behaviour has often been referred to as static liquefaction. The line, which connects the top of the effective stress paths, has been called the instability line. The zone between the instability line and the failure line (or the critical state line) is called the zone of potential instability.
500
180
400
drained test
300
200
Instability line
100
A
0 100 200
(a) (b) Figure 3 Instability of loose sand under undrained conditions It needs to be pointed out that the instability line is not unique, but varies with the void ratio of sand and the applied effective mean stress. Fig. 3b shows the effective stress paths obtained from a series of isotropic consolidated undrained (CIU) tests conducted on specimens with different consolidation void ratios, ec, but under the same mean effective stress. It can be seen from Fig. 3b that the smaller the ec, the higher the instability line. It has also been established that instability will not occur for loose sand under a drained condition if the stress state imposed at a point (for example, at point A in Fig. 3a) does not change (Chu and Leong, 2001). However, the stress state of a soil element along a slope is changing. For example, during water infiltration, a soil element along a slope may follow a constant shear drained (CSD) effective stress path, as suggested by Brand (1981). Along this stress path, the shear stress is considered constant but the mean effective stress is reducing. The instability behaviour of sand along a CSD path has been examined experimentally. The results of a typical test, DR7, are shown in Fig. 4. The loose sand specimen (with a void ratio ec = 0.945) was firstly sheared to point A along a drained path (Fig. 4a). The deviator stress at point A is q = 150 kPa. On the q = constant path, the confining stress was reduced at a rate of 1 kPa/min, resulting a stress path moving from point A to point B (Fig. 4a). There were little axial and volumetric strain developments until point B where both axial and volumetric strains started to develop at a faster rate, as shown in Figs. 4b and 4c. This can be seen more obviously from Fig. 4d which shows that the axial strain rate shot up at point B, indicating an unstable behaviour. Using point B, the instability line can be determined as shown in Fig. 4a. With further reduction in the confining stress, the stress path moved further toward the CSL. However, at this stage the axial and volumetric strain rates had increased to such an extent that the testing system could not catch up to maintain q to be constant. It needs to be pointed out
592
COMMON GROUND PROCEEDINGS 10TH AUSTRaLIa NEW ZEaLaND CONFERENCE ON GEOMECHaNICS BRISbaNE
300 250 q, Deviator Stress (kPa) 200 150 100 50 0 0 100 200 300 p', Mean Effective Stress (kPa)
160
DR7, ec = 0.945
Confining Stress (kPa)
140
A B