大型洞室群稳定性与优化的并行进化神经网络有限元方法研究──第二部分:实例研究

STUDY ON PARALLEL EVOLUTIONARY NEURAL NETWORK FEM ON STABILITY AND OPTIMIZATION FOR LARGE CAVERN GROUPS──PART Ⅱ:CASE STUDY

利用作者提出的并行进化神经网络有限元方法对水布垭地下厂房进行了软岩置换方案的优化及稳定性分析。结果表明,该方法具有极大的全局搜索和快速收敛的优势。实例给出了最优的置换方案和置换步骤,通过最优方案的有限元与神经网络的计算比较表明,该方法是合理的。同时,提出了施工的合理化建议。

A case history study on replacement scheme optimization and stability analysis of soft rock mass at a Shuibuya underground power house is presented using the proposed parallel evolutionary neural network FEM. The results indicate that the presented methodology is superior in global searching and quick convergence. The case history study gives the optimum replacement scheme and replacement steps. Through the comparison between the FEM and neural network calculation of the optimum scheme,the methodology is tested to be reasonable. Meanwhile,the rational suggestion is proposed to guide the construction.