Back analysis for mechanical parameters of surrounding rock for underground roadways based on new neural network

摘要

For its complicated depositing environment of surrounding rock mass for underground roadways, it is a very important work to back-calculate the mechanical parameters of surrounding rock mass by measurement displacements. To overcome the shortcomings of the traditional neural networks, a new neural network based on black hole algorithm has been proposed. Then, a new back analysis method based on new neural network has been studied. Using this new back analysis method, the mechanical parameters of surrounding rock mass for two deep roadways in Huainan coal mine of China have been back-calculated based on the measurement convergence displacements. Moreover, the good performance of the new back analysis method has been compared with those by back propagation network, neural networks based on genetic algorithm and immunized evolutionary programming proposed in previous studies. The results show that, using the back-calculated parameters, the computed displacements agree with the measured ones. And, considering the computing effect and efficiency comprehensively, the new back analysis method is the good method to determine the suitable mechanical parameters of surrounding rock mass for underground roadways.