基于松动圈–位移增量监测信息的高地应力下洞室群岩体力学参数的智能反分析

INTELLIGENT BACK ANALYSIS OF ROCK MASS PARAMETERS FOR LARGE UNDERGROUND CAVERNS UNDER HIGH EARTH STRESS BASED ON EDZ AND INCREMENT DISPLACEMENT

考虑到高地应力下洞群围岩力学行为独特性和深部地下工程研究需要,提出大型洞室群岩体参数的智能反演新方法。该方法是采用弹脆塑性本构CWFS模型,以分步开挖引起的松动圈和位移增量监测信息为输入,首先通过参数敏感性分析确定待反演的参数,再建立位移增量–松动圈深度的联合适应度函数,用进化神经网络–遗传算法求得待反演参数的数值,并进行后续开挖引起的位移演化的灰色关联度分析和松动圈的实测值与计算值对比对反演结果进行检验的反分析方法。用该方法对具有高地应力特征的拉西瓦水电站花岗岩地下洞室群围岩的力学参数进行了反演,通过地下洞室群的第2～6层开挖引起的位移增量和松动圈的测试值,反演求出相关5个岩体力学参数的值。用反演所得的参数值对第7步开挖引起的位移增量和松动圈变化进行计算分析,结果表明该方法的正确性。

A new intelligent back analysis method based on excavation damage zone（EDZ） thickness,increment displacement of rock mass and evolutional artificial neural network is presented for understanding the particular mechanical activity of rock mass under high earth stress and for estimating surrounding rock stability of deep underground rock engineering.The method is applied to stability analysis for the large caverns of Laxiwa Hydropower Station,which is the largest underground caverns located on the Yellow River.Through getting key rock mechanical parameters by sensitivity analysis and building the united sufficiency function based on the increment displacement and EDZ thickness,the rational values of five key rock mechanical parameters are recognized using the method.The results of qualitative analysis of EDZ and time serial displacement grey correlation analysis indicate that the method is credible.