特高拱坝多参数反演的智能优化径向基函数网络模型  

Intelligent Optimization Radial Basis Function Network Model for Multi-parameter Inversion of Super-high Arch Dam

作　　者：寇文状 康飞[1] 梅智 KOU Wen-zhuang;KANG Fei;MEI Zhi(School of Hydraulic Engineering,Dalian University of Technology,Dalian 116023,China;PowerChina Zhongnan Engineering Corporation Limited,Changsha 410014,China)

机构地区：[1]大连理工大学水利工程学院,辽宁大连116023 [2]中国电建集团中南勘测设计研究院有限公司,湖南长沙410014

出　　处：《水电能源科学》2023年第11期69-72,共4页Water Resources and Power

基　　金：国家重点研发计划(2022YFB4703400);国家自然科学基金项目(52079022,51979027);中央高校基本科研业务费项目(DUT21TD106)。

摘　　要：利用大坝的监测数据对大坝力学参数进行精确反演,对于确保大坝的安全稳定运行至关重要。对此,提出基于径向基函数(RBF)网络和人工大猩猩部队优化算法(GTO)的拱坝参数反演模型。首先,采用RBF代理模型代替有限元模型,探讨材料参数与监测点位移响应之间的关系,RBF代理模型的采样数据由高效的拉丁超立方采样技术生成;其次,采用GTO智能优化算法,使材料参数识别的目标函数最小。工程实例分析结果表明,RBF-GTO模型能够在降低计算成本的同时实现高精度的混凝土特高拱坝参数反演分析。Accurate inversion of dam mechanical parameters based on dam monitoring data is crucial to ensure the safe and stable operation of the dam.This paper presented an arch dam parameter inversion model based on radial basis function(RBF) network and Artificial Gorilla Troops Optimizer(GTO).Firstly,the RBF surrogate model was used to replace the finite element model to discuss the relationship between the material parameters and the displacement response of the monitoring point.The sampling data of the RBF surrogate model was generated by the efficient Latin hypercube sampling technology.Secondly,the GTO intelligent optimization algorithm was adopted to minimize the objective function of material parameter identification.The analysis results of engineering examples show that the RBF-GTO model can achieve high-precision parameter inversion analysis of concrete super-high arch dams while reducing the calculation cost.

关 键 词：径向基函数网络 代理模型 位移反分析 智能优化 特高拱坝 

分 类 号：TV642.4[水利工程—水利水电工程]