基于径向基函数神经网络的汽轮机转子等效应力计算模型  

Computation Model for Equivalent Stresses in Steam Turbine Rotors Based on RBF Neural Network

作　　者：方宁[1] 纪冬梅[1] 姚秀平[1] 杨宇[2] 史进渊[2] 

机构地区：[1]上海电力学院能源与环境工程学院,上海200090 [2]上海发电设备成套设计研究院,上海200240

出　　处：《动力工程学报》2011年第4期268-272,共5页Journal of Chinese Society of Power Engineering

基　　金：国家863计划资助项目(2007AA04Z429);上海教委重点学科(第5期)(G51304);上海市教委第3期本科教育高地建设项目

摘　　要：为解决汽轮机转子应力的在线监测问题,建立了基于径向基函数(RBF)神经网络的汽轮机转子等效应力计算模型,对比了用有限元和RBF神经网络模型两种方法计算得出的冷态、温态、热态和极热态4种启动工况下转子调节级叶轮根部圆角处的等效应力.计算结果表明:RBF神经网络模型计算结果与有限元法解的结果非常相近,且计算简便、耗时少,可以应用于汽轮机转子等效应力的在线计算,为汽轮机转子寿命在线管理提供依据.A model for on-line calculation of equivalent stresses in steam turbine rotors has been built up based on radial basis function（RBF）neural network.Equivalent stresses at root area of rotor impeller respectively obtained with finite element and RBF model are compared under cold,warm,hot and extreme hot start-up conditions.Results show that similar calculation results of equivalent stress may be obtained with either finite element or RBF model.However,with the latter one,the calculation becomes easier and fast,which is applicable for on-line calculation of equivalent stresses in turbine rotors,and may serve as a reference for on-line life management of turbine rotors.

关 键 词：汽轮机 转子 等效应力 径向基函数 神经网络 在线计算 

分 类 号：TK263.6[动力工程及工程热物理—动力机械及工程]