Neural identification of rock parameters using fuzzy adaptive learning parameters

An improved fuzzy adaptive back-propagation algorithm is proposed and applied to the identification of mechanical parameters of the surrounding rock of underground caverns in stratified layers. The improved algorithm adopts some advanced ideas and techniques from computational intelligence, and combines the fuzzy theory with artificial neural network techniques and the mutation strategy in genetic algorithms. The improved algorithm extends the effectiveness and the adaptivity of the Fuzzy BP algorithm. The successful estimates of mechanical parameters and the initial stresses in the surrounding rock show that a feasible method of identification is provided, which can be used to identify parameters in rock engineering quickly and effectively.