A tolerance-based framework for optimizing profile sampling intervals in joint infilling assessment and shear strength prediction

The accurate assessment of joint infilling is critical for reliably estimating the shear strength of rock masses. While common practice often follows the International Society for Rock Mechanics (ISRM) suggestion of using a sampling interval ( ∆L ) equal to 2% of the total sample length, the empirical basis for this specific guideline remains limited. This study systematically investigates the influence of this key methodological parameter—the profile sampling interval ∆L —on the assessment of the infilling degree ( λ ) and its consequential impact on shear strength prediction. Through direct shear testing on filled joint replicas and a comprehensive analysis of 37 natural joint profiles across 500 different ∆L values, a defined “interval effect” for the coupled parameter λ was identified, characterized by a stable phase and a subsequent fluctuating phase. Statistical analysis revealed a strong linear correlation between a predefined maximum permissible error in λ and the corresponding maximum allowable sampling interval ( ∆L _max). A conservative lower-bound linear model was established to determine ∆L _max for practical application. Validation by predicting shear strength demonstrated that using these optimized intervals yielded accuracy comparable to the conventional ISRM 2% rule, while potentially reducing the number of required profile lines by approximately 55% in a representative case. This indicates that the fixed 2% interval can be unnecessarily conservative for some joint morphologies. The proposed framework provides a data-informed method for selecting ∆L based on a performance criterion, thereby offering a practical approach to enhance the efficiency of infilling assessment.