This paper describes the application of the maximum likelihood statistical technique to back calculation of shear strength parameters of joints. The approach adopted assumes that the factors of safety against shear failure along discontinuity surfaces are log-normally distributed and it seeks the values of cohesion and friction angle that maximize the frequency distribution of factors of safety around one. Using this method, the cohesion and friction angle values of the two dominant joint sets in an open pit mine are estimated from 48 failed and 46 unfailed cases of plane failure geometries. Further shear strength estimation studies using conventional back-analysis and laboratory testing are also carried out and these show that the results obtained from the three approaches are all in good agreement.
A realistic determination of shear strength of discontinuities is the prime importance in designing rock slopes. Laboratory determination of discontinuity cohesion and friction angle involves uncertainties due to the insufficient sample size to represent in-situ conditions. Back-analysis of actual failed cases provides more realistic results and thus is the preferred method over laboratory testing when there is sufficient data available for statistical analysis. This paper describes the application of maximum likelihood statistical technique as a back-analysis tool for estimating the cohesion and friction angle of joints. The approach is similar to that used by Salamon and Munro (1967) for estimating strength of coal pillars. The in-situ data used comes from the Chuquicamata open pit mine in Chile, where the two main joint sets are known to play a significant role in the overall stability of the mine''s west wall.
