平面弹性位移反演灵敏度分析及随机位移反分析

Study on Parameter Sensitivity Analysi and Stochastic Displacement Back Analysis

提出正算过程采用平面弹性复变方法中的Cauchy积分解法,优化方法选用遗传算法的随机位移正反分析方法。分析了目标函数对待反演参数的灵敏度,研究了位移观测噪声对识别结果的影响规律。结果表明该方法从根本解决了采用优化技术进行位移反分析的求解效率问题;能够反映地应力场和实际岩体的不确定性;能够反映位移观测噪声对反演结果产生的影响,具有较好的容错性能和全局搜索性能。

Aimed at the problems existing in back analysis,stochastic back analysis by use of Cauchy integral method and Genetic Algorithm was presented.Then,parameter sensitivity analysis and the effect on identification results of observation error was studied.The results indicated that the presented method had great fault-tolerance and global searching performance,it improved the solution efficiency basically and could reflect the uncertainty characteristics of ground stress field and engineering rock mass.