Civil Engineering April 2022 | Vol 30 No 3

Civil Engineering April 2022 15 confinement has resulted in critical state conditions being reached. It would be concluded that a loss of confinement of this magnitude is a trigger which will cause post-peak strengths and is also thus a significant risk to the facility that would need to be prevented through appropriate risk management. INVESTIGATION INTO EFFECT OF RAISED PHREATIC SURFACE Investigations into this potential trigger have been performed by elevating the phreatic surface by some degree in a steady-state seepage analysis. As an example, Figure 9 provides the plot of stress paths in p’:q’ space, where solid lines indicate the portion of the simulation where the phreatic surface was raised. The increase in phreatic surface level results in a minimal decrease of mean effective stress with little change of deviatoric stress. Stress paths remain far from intersection with the CSL, for this scenario. It may be concluded that an increase in elevation of the phreatic surface of this magnitude is not a trigger for liquefaction in itself. DYNAMIC ANALYSIS The historic approach to assessing the effect of seismicity has involved selection of an appropriate event return period (Global Tailings Initiative and ICOLD advocate a specific return period for various consequence ratings) and then consulting seismic or natural hazard assessment records of past events to statistically predict the magnitude of various return period events. The event magnitudes are expressed in terms of peak horizontal ground acceleration. Traditional stability analyses then apply the horizontal acceleration as a static load in the stability model. This approach is flawed in that real seismic events are of relatively short duration and impose dynamic/cyclic loading. Furthermore, the mechanism by which seismic actions destabilise TSFs is not related as much to the presence of global destabilising forces as it is to a decrease in internal material strengths due to the seismic actions. The CSSM analytical approach allows cyclic load to be applied as a realistic dynamic transient external stress (load) to the NorSand controlled tailings FEM model based upon the time histories of actual recorded seismic events scaled Deviator stress 800 700 600 500 400 300 200 100 0 Mean effective stress 1 000 800 600 400 200 Figure 9 Stress paths for queried points Point 6 Point 7 Point 8 Point 9 Point 10 Point 15 Point 11 Point 12 Point 13 Point 14 Point 16 Point 17 Point 18 Point 19 Point 20 Point 25 Point 21 Point 22 Point 23 Point 24 Point 2 CSL Point 1 Point 1 Point 3 Point 4 Point 5 Point 2 Deviator stress (q (kPa)) 600 500 400 300 200 100 0 Mean effective stress (p’ (kPA)) 1 200 800 600 400 0 Figure 10 Stress paths for queried points in a facility submitted to seismic loading Point 1 Point 2 Point 3 Point 4 Point 5 Point 6 Point 7 Point 8 Point 9 Point 10 Point 15 Point 11 Point 12 Point 13 Point 14 Point 16 Point 17 Point 18 Point 19 Failure envelope Point 20 Point 21 Point 22 Point 23 200 1 000