Shear Strength of a Large Limestone Discontinuity: In Situ Pull Test and Prediction

Confidence in predictions of discontinuity shear strength at full scale is traditionally plagued by scale effects on material strength and roughness. The stochastic approach to discontinuity shear strength (here abbreviated as StADSS) bypasses scale effects on shear strength predictions of natural rock discontinuities by using full-scale discontinuity roughness data. The StADSS methodology was applied to an inclined limestone discontinuity specimen, and compared to experimental results at full scale, by in situ pull test. The limestone specimen was located on a highwall approximately 6 m above ground level. The discontinuity had two traces exposed on the surface on the specimen: one approximately parallel to the shear direction, and one approximately perpendicular to the shear direction. The traces were digitised using high-resolution images captured by a UAV drone to allow implementation of StADSS. A robust material characterisation was conducted to determine the input parameters for StADSS. Lastly, the statistics of an extracted trace from a scan of the shear surface parallel to the shear direction was utilised to determine the peak and residual shear strength of the discontinuity using the continued fraction regression model developed by one of the authors. The investigation revealed that the specimen had been displaced by approximately 20 mm in the shear direction prior to the investigation, because of blasting operations, so the peak shear strength had already been mobilised and a lower shear strength was measured during the pull test. The predictions of shear strength successfully corroborated the in situ stability of the specimen; however, it was not possible to validate the value of peak shear strength, as peak shear strength had been historically mobilised. The paper presents a full application of StADSS and highlights the importance of establishing possible prior displacements of a discontinuity when predicting shear strength using roughness data exposed on a rock face.