Prediction of unconfined compressive strength of rock surrounding a roadway using artificial neural network

摘要: The unconfined compressive strength (UCS) of rocks is an important design parameter in rock engineering and geotechnics, which required determined for mechanical studies mining civil projects. This usually through a laboratory UCS test. Since the preparation high-quality samples difficult, expensive time consuming tests, development predictive models determining properties seems to be essential engineering. In this study, attempt was made develop artificial neural network (ANN) multivariable regression analysis (MVRA) order predict surrounding roadway. For this, database tests prepared, includes type, Schmidt hardness, density porosity as input parameters output parameter. To make (including 93 datasets), different samples, ranging from weak very strong types, are used. compare performance developed models, determination coefficient (R2), variance account (VAF), mean absolute error (Ea) relative (Er) indices between predicted measured values were calculated. Based on comparison, it concluded that ANN model considerably better than MVRA model. Further, sensitivity shows hardness recognized most effective parameters, whereas considered least study.