Engineers Australia

GSI Adjustments for Directional Hoek-Brown Strength Calibrated by Step-Path Case Studies

Norbert R. P. Baczynski


Step-Path methods aim to quantify the negative rock mass shear strength impact of geological defects co-aligned with failure paths through rock slopes and the positive strength impact of intact rock ‘bridges’ between some or all of the co- aligned defects. In effect, Step-Path considers directional strength in rock masses.

Conceptually Step-Path may be reconciled with the Hoek-Brown Method by adjusting the Geological Strength Index (GSI) input to Hoek-Brown equations so that both Step-Path and Hoek-Brown yield the same shear strength outcomes. The GSI adjustment is a two-step process. GSI is first negatively adjusted for the relative portion of the failure path that is defined by geological defects co-aligned with the failure path. The GSI is then positively adjusted for the relative portion of the failure path that is defined by intact rock ‘bridges’ between co-aligned defects. For a ‘general’ rock mass, the adjusted GSI is derived via the following equations:

GSI design = GSI general rock mass – GSI defects adjustment + GSI rock bridges adjustment


GSI general rock mass = GSI rating that would have been conventionally estimated for the ‘general’ rock mass

GSI defect adjustment = 0.4 x co-aligned defect occurrence (%)

GSI rock bridge adjustment = 1.2 x intact rock ‘bridge’ occurrence (%)

The above GSI adjustment factors of 0.4 and 1.2, respectively, are calibrated by reviewing 230 Step-Path case study models developed on projects over the last 20+ years. Key considerations impacting Hoek-Brown rock mass and Barton geological defect shear strengths and their influence on Step-Path strength are flagged. Challenges facing the reconciliation task are identified and discussed. In addition to GSI adjustment factors for the ‘general’ rock mass condition, adjustment factors are also suggested for rock masses partitioned by mi grouping and rock type. Limiting conditions are identified.