【摘要】 根据大坝原型观测资料识别大坝结构力学参数是典型的非线性反问题,而反问题往往具有Hadamard意义下的不适定性。基于位移反分析方法的基本原理,依据大坝原型观测数据和有限元数值计算成果,建立了优化目标函数后,应用遗传模拟退火算法实现了对大坝及岩基物理力学参数的反演分析。对反演原理、算法及流程进行了研究。将遗传算法和模拟退火算法相结合,既提高了优化问题全局收敛速度,同时克服了传统方法求解病态方程的困难。实例分析表明,该优化反演方法对抵抗噪音有较强的能力,且与正分析形成的闭合环路可以高精度的完成大坝工作性态的分析。更多还原

【Abstract】 It is a typical nonlinear inverse problem to identify the mechanical parameters of dam and its foundation from prototype observations of dam behaviors.According to prototype observations and results of FEM,the basic principle of inverse analysis for displacement is used to build the aim function of above problem.The feedback analysis is implemented with the genetic simulated annealing algorithm.It has the advantages of great efficiency and good fault-tolerance and robustness.Based on the parameters obtained by the proposed feedback analysis method,dam behavior can be analyzed reasonably.更多还原