摘要
滑动电接触是一个具有耦合作用的特殊摩擦副,其中接触电阻和磨耗率是影响接触状态的重要因素。确定接触载荷的数值,使得两个目标最小是本文研究的重点。通过滑动电接触实验,分析了接触电阻和磨耗率存在的相互冲突、矛盾关系,描述了此种多目标问题的Pareto解。并得到了接触电阻和磨耗率为输出,运行速度、接触载流以及载荷为输入的变化数据。进而对其进行了实验数据的Pareto解分布分析,训练了相应的神经网络黑箱模型,依此设计了应用于多目标的人工鱼群优化算法,并进行了测试函数检验,得到了较好效果。最后利用该算法对所建立神经网络模型进行接触载荷的优化,得到了相应工况下最优接触载荷的数值,对结果进行了分析,表明了算法的有效性。
The sliding electric contact can be considered as a current-passing tribosystem with coupling action,and the contact state is influenced by many factors.It is the method to find the best contact load,which makes both of the contact resistant and ratio of wear least.For the conflict of the both,the Pareto meaning is given in the multi-object,and some definitions are given too.Then some experiments about the sliding electrical contact on the designed machine were done.With the data,whose inputs are running velocity,current and contact load,and outputs are contact resistant and ratio of wear,and the black box model using the neural network is trained.So a multi-objective artificial fish arithmetic is proposed to optimize the model to minimize the contact resistant and ratio of wear.At last the contact load is obtained in the giving condition,which shows the validity of the arithmetic.
出处
《电工技术学报》
EI
CSCD
北大核心
2013年第5期196-201,共6页
Transactions of China Electrotechnical Society
基金
国家自然科学基金(50977040)
辽宁省自然科学基金(201102086)资助项目
关键词
电接触
接触电阻
磨耗率
神经网络
优化
Electric contact
contact resistance
ratio of wear
neural network
optimization