Indirect Estimation of Uniaxial Compressive Strength of Limestone Using Rock Index Tests Through Computational Methods

Uniaxial compressive strength (UCS) is a critical geomechanical property of rocks that is frequently required during the preliminary stage of civil engineering design. To obtain the UCS value needs a time consuming and costly process of samples collection and preparation. There are alternate methods for determining UCS that can be conducted in situ. In this study, an attempt has been made to predict the UCS of limestone from some simple and inexpensive rock index tests such as block punch index (BPI), ultrasonic wave velocity test (Vp), Schmidt's hammer rebound number (SHR), and point load index tests (I_s50). For indirect estimation of the UCS as a function of BPI, Vp, SHR, and I_s50, block samples of limestone were collected from a quarry site in Birjand, the center of Southern Khorasan province in Iran. Then, the number of 70 core samples and 210 bulk samples were prepared and tested based on available standards. According to extensive experimental results, a database was established for estimation of the UCS via three computational methods such as support vector machine (SVM), adaptive neuro-fuzzy inference system (ANFIS), and multi layer perceptron (MLP). After developing the models and considering several performance indices including the coefficient of determination R^2, variance account for (VAF), root mean squared error (RMSE), and using simple ranking method, the predictive models were applied to obtain the best model. Consequently, SVM approach predicted the UCS of limestone with higher accuracy in comparison to other studied computational methods.