摘要
分析了液压缓冲器的结构及其动态工作过程 ,介绍了基于结构的神经网络建模方法 .该建模方法根据系统结构组成特点将复杂系统分解为相互关联的简单子系统 ,用函数链神经元分别建立子系统模型 ,然后根据子系统间固有的连接关系将子系统神经元模型连接成一个网络 ,所得网络模型即为原系统模型 .应用该方法建立了 5 2 SFZ- 1 40 - 2 0 7B液压缓冲器的动态模型 .结果表明 ,基于结构的神经网络建模方法对复杂非线性系统建模是有效的 .
The structure and dynamic working process of hydraulic bumper were discussed. The modeling method using architecture-based neural network was introduced. Using this method, the complex nonlinear system is divided into several simple sub-systems according to its structure, each sub-system is learned by a functional link neuron respectively, then the neurons are connected into a network according to the coherent relations among sub-systems, the network is the system model. The dynamic model of 52SFZ-140-207B type of hydraulic bumper was established using this modeling method. The result shows that the modeling method using architecture-based neural networks is suitable to the modeling of complex nonlinear system.
出处
《上海交通大学学报》
EI
CAS
CSCD
北大核心
2003年第5期741-744,共4页
Journal of Shanghai Jiaotong University
关键词
液压缓冲器
结构
神经网络
建模
Dynamics
Hydraulics
Neural networks
Nonlinear systems
