Options
Development of Statistical Models to Predict the Mechanical Properties of Some Metamorphic Rocks from P-Wave Velocity and Certain Physical Properties
Journal
Geotechnical and Geological Engineering
ISSN
09603182
Date Issued
2022-08-01
Author(s)
Singh, Hari Om
Ansari, Tariq Anwar
Singh, T. N.
Singh, K. H.
Abstract
Strength and elastic properties of rocks have vital importance for the design of structures in several rock engineering projects. Additionally, these parameters are extensively used in numerical modelling to envisage the mechanical behavior of rock before failure and their relevance to geotechnical characteristics in slope stability. This study was intended to predict the geomechanical properties of three different metamorphic rocks using simple and multiple regression analysis. The predictive models and their correlation coefficients were established and well discussed for P-wave velocity and different physico-mechanical parameters. Initially, simple regression analysis was performed and then the authenticity of the result was also validated by student’s ‘t’ and ‘F’ test. Further, 13 different empirical models were developed to predict the strength and elastic properties of these rocks through multiple regression analysis. The predicted values estimated from these models show a deterministic and closely associated result with the observed values. Analysis of variance of these rock-mechanics data was also compared to check the confidence of multiple regression models. The coefficient of determination (R2), the root mean square error (RMSE), the mean absolute percentage error (MAPE), and the variance account for (VAF) of these empirical equations were estimated and compared to quantify the performance of the models. The comparative result of statistical indices (R2, RMSE, MAPE, VAF) revealed negligible error in developed models with high prediction performance and may be accepted for empirical models to predict the mechanical properties.
Subjects