岩石力学与工程综合集成智能反馈分析方法及应用

首先,给出岩石力学模型和参数综合智能反分析的几种新方法:岩石(体)力学参数的智能反演方法、岩石(体)本构模型的结构和参数耦合智能识别方法、岩石(体)力学参数的时空变异性反分析方法(即集合Kalman滤波和扩展Kalman滤波方法)、力学参数和模型可以不断更新的演化并行有限元反分析方法和基于模型的结构和参数智能识别的岩体工程安全性的综合集成智能分析方法。其中,岩石(体)力学参数的智能反演方法又包括4种,分别为均匀设计(或正交设计)–数值计算方法、粒子群(或遗传算法)–数值方法、均匀设计(或正交设计)–演化神经网络–数值方法–遗传算法(或粒子群)以及均匀设计(或正交设计)–演化支持向量机–并行数值方法–遗传算法(或粒子群)。然后,讨论智能反分析时应注意的待反演参数对监测物理量的敏感性、可以进行反演的参数个数以及模型和参数反演结果的适应性评价等问题。最后,给出这些智能反馈分析方法的工程综合应用概况。

INTEGRATED INTELLIGENT FEEDBACK ANALYSIS OF ROCK MECHANICS AND ENGINEERING PROBLEMS AND ITS APPLICATIONS

Several new methods for intelligent feedback analysis of rockmass models and their parameters are summarized.There include intelligent back analysis methods of mechanical parameters of rockmass,coupling recognition of the structure of the model and parameters,intelligent back analysis of variability of rock mass parameters in time and space,intelligent parallel finite element method with dynamical renew of the model and its parameters and integrated intelligent analysis of rock engineering safety with dynamical recognition of rock model and parameters.Among them,intelligent back analysis methods of mechanical parameters of rockmass include four kinds of methods such as uniform design(or ortho-design)-numerical method,practice swarm(or genetic algorithm) numerical method,uniform design(or ortho-design)-evolutionary neural network-numerical method-practice swarm(or genetic algorithm) and uniform design(or ortho-design)-evolutionary support machine-parallel numerical method-practice swarm(or genetic algorithm) method.Several key problems to obtain global optimum solutions for the back analysis of rock model and parameters are also discussed.Finally case studies have been given to show applicability of the new methods.