基于优化遗传神经网络的水轮发电机组辨识研究.pdf

of the撕ChineseControleonreronce Proceedings Guani留．hou,P．R．China Julyl量工8：2005 基于优化遗传神经网络的水轮发电机组辨识 赵宏 (中国计量学院机电工程分院，浙江杭州310027) E-mail：pzhzy@163，com 摘要：针对水轮发电机组具有非线性、时变、非最小相位且结构参数变化范围较大的特点，提出了将优化遗传算 法和神经网络相结合实现对水轮发电机组辨识的新方法，阐述了其原理并给出了相应的算法和计算公式，该方 法结合优化遗传算法与神经网络各自的优点，克服了传统BP神经网络收敛速度较慢以及易于收敛到局部极小点 等缺点，仿真试验表明，采用该方法能较迅速、准确地逼近实际系统，表明了其有效性． 关键词：水轮发电机组，辨识，遗传算法，神经网络 The ofthe Identificationof Generator Study Optimum Hydraulic UnitBasedOnGeneticNeural Networks Zhao hong (College Institute 310027China) ofMechanicalElectricalEngineering．ChinaofMetrology，Hangzhou Abstraet：Tothe and nonlinear，time．variablenon-minimumcharatterthe and varied ofthe phase large region new unit’sstructureand methodis forththat identification hydraulicgenerator parameters，A put optimum ofhydraulic unit the ofthe andneuralnetwork．The characteristics generatorby optimumalgorithms combining genetic principle is the andformulasare cancombinethe of and expounded，andalgorithm presented，It advantagesgenetic algorithms neuralnetworks，andovercomesthe ofBPnetworks，suchasthe rate，andto shortcomings slowlylearning easy tothe