Predicting the Young's Modulus of granites using the Bayesian model selection approach

The value of Young's modulus (E) is critical to the design of geotechnical engineering projects. Although E can be directly measured by laboratory tests, high-quality core samples and expensive sophisticated instruments are required. Therefore, a method for the indirect estimation of E is an appealing possibility. This study develops a model for predicting the E of intact granite based on the Bayesian model class selection approach. An experimental database of granite rock properties that includes the value E, point load strength index (I-s50), L-type Schmidt hammer rebound number (R-L), P-wave velocity (V-p), porosity (), and uniaxial compressive strength, is applied to develop the most suitable model. The proposed model is then compared to existing approaches. The results indicate that the proposed models provide satisfactory predictions and good practicality in application.