摘要
随着新能源的快速发展以及用电功率的大幅提升,对接触器的性能及可靠性提出了更高的要求,接触器的优化设计也随之成为热点。受接触器性能的低仿真计算效率制约,智能优化算法等全局搜索能力较强的优化算法难以在接触器优化设计中得到很好的应用,限制了接触器优化设计的发展。针对以上问题,有必要对接触器性能的快速计算展开研究。提出了一种接触器近似建模方法,以径向基函数神经网络为基础,通过次胜者受罚竞争学习(RPCL)算法及粒子群(PSO)算法优化径向基函数神经网络中的参数值,从而保证模型的准确度。以直动式大功率接触器的静动态特性为例,对近似模型计算一次动态特性需要40 s,而工程中常用的有限元算法则需要近1 d时间,且近似模型的误差在5%以内。由此可见,该模型极大提升了接触器性能求解效率,为接触器优化设计奠定了基础。
With the rapid development of new energy systems and the significant increase in power consumption,users put forward new requirements for the performance and reliability of contactors.The optimal design of contactors becomes a hot issue.Restricted by the low efficiency for computation of contactor performance,intelligent optimization algorithms with strong global search ability can’t be well applied in contactor optimization design.That limits the development of contactor optimization design so it is necessary to study the fast calculation algorithms of contactor performance.In this paper,an approximate model for contactors was proposed,which is based on radial basis function(RBF)neural network.The parameters of the RBF neural network are optimized by RPCL and PSO to ensure the accuracy of the model.Taking the static and dynamic characteristics of a high-power contactor as an example,it takes 40 seconds for an approximate model to calculate the dynamic characteristics,while nearly one day for the commonly used finite element method.The error of the approximate model is less than 5%.In general,the model greatly improves the efficiency of computation of the performance of contactors and lays the foundation of optimal design of contactors.
作者
肖斌
刘洋
翟国富
XIAO Bin;LIU Yang;ZHAI Guofu(Reliability Institute for Electric Apparatus and Electronics,Harbin Institute of Technology,Harbin 150001,China)
出处
《电器与能效管理技术》
2020年第1期40-45,共6页
Electrical & Energy Management Technology
