Application of artificial neural networks and multivariate statistics to predict UCS and E using physical properties of Asmari limestones

Geomechanical properties of rocks such as uniaxial compressive strength (UCS) and modulus of elasticity (E) have been essentially evaluated for rock engineering projects as well as dam sites. In this paper, in order to estimate the parameters, some mathematical methods are proposed including multiple linear regression, multiple nonlinear regression, and artificial neural networks (ANNs). These methods were employed to predict UCS and E for limestone rocks in terms of P wave velocity, density, and porosity. The data of 105 rock samples from two different dam sites (located in Asmari Formation, Karun 4, and Khersan 3 dams) were obtained and analyzed for developing predictive models. Comparison of the multiple linear and nonlinear regressions and ANNs results indicated that respective ANN models were more acceptable for predicting UCS and E than the others. Also, it observed that between multiple linear and nonlinear regressions, second case has more capability to predict UCS. It should be noted that there were no strong relationships between the predicted and measured E in the both multiple regressions.