Applying soft computing methods to predict the uniaxial compressive strength of rocks from schmidt hammer rebound values

The uniaxial compressive strength (UCS) of rock is widely used in designing underground and surface rock structures. The testing procedure of this rock strength is expensive and time consuming. In addition, it requires well-prepared rock cores. Therefore, indirect tests are often used to estimate the UCS, such as the Schmidt hammer test. This test is very easy to carry out because it necessitates less or no sample preparation and the testing equipment is less sophisticated. In addition, it can be used easily in the field. As a result, comparing with uniaxial compression test, indirect test is simpler, faster, and more economical. In this paper, the application of soft computing methods for data analysis named support vector regression (SVR) optimized by artificial bee colony algorithm (ABC) and adaptive neuro-fuzzy inference system-subtractive clustering method (ANFIS-SCM) to estimate the UCS of rocks from Schmidt hammer rebound values is demonstrated. The estimation abilities offered using SVR-ABC and ANFIS-SCM were presented by using experimental data given in open-source literatures. In these models, the Schmidt hammer rebound values (T1-T3, R1-R4) were utilized as the input parameters, while the UCS was the output parameter. Various statistical performance indexes were utilized to compare the performance of those estimation models. The results achieved indicate that the ANFIS-SCM model has strong potential to indirect estimation of the UCS of rocks from the Schmidt hammer rebound values with high degree of accuracy and robustness.