Estimation of rock mass rating system with an artificial neural network

Abstract

The geo-mechanical classification - rock mass rating (RMR) - is used for categorizing rock mass. Assessing RMR is an important factor for successful accomplishment of a tunneling project. In the rock mechanics and mining literatures, some empirical methods exist between rock mass and other rock properties, such as using characteristic of the rock, geological structure etc. However, those means have some limitations by special rock types. After analyzed the information to identify RMR, a new parameter as one of the input neurons was used to develop predictive relations. There are eight parameters as the input parameters are presented based on artificial neural networks (ANN). The situ-test data of the tunnel face were measured and the experimental results indicate the proposed method was effective. © 2009 Springer Berlin Heidelberg.