变电站密闭空间中设备故障的智能判定方法

变电站设备大多处于密闭空间中,依赖传感器和专家经验来判别故障,为避免当前故障判别方式的弊端并提高变电站智能化水平,提出一种智能判定方法。首先,在对各类巡检需求分析的基础上完成数据的采集和标注,构建数据库;其次,分别介绍门控循环单元和粒子群算法,基于粒子群优化的门控循环单元网络(particle swarm optimization-gated recurrent unit, PSO-GRU)提出智能故障判定方法,给出了具体的网络结构和算法流程;最后,结合实际数据设计仿真实验,提供了具体的实验方式和流程,将本文方法的判别效果进行验证并与其他两种网络进行对比。结果表明:PSO-GRU对故障的判定更加快速和准确。

Research on intelligent judgment method of equipment fault in sub-station confined space

Most of the substation equipment is located in a confined space, and these substations rely on sensors and ex-pert experience to identify faults. In order to avoid the drawbacks of current fault identification methods and improve the level of substation intelligence, an intelligent decision method is proposed. Firstly, data collection and labeling were completed on the basis of analysis of various types of patrol requirements, and a database is completed. Secondly, gated loop unit and particle swarm algorithm were introduced respectively. Based on the gated loop unit network of particle swarm optimization, an intelligent fault diagnosis method is proposed. Be-sides, the specific network structure and algorithm flow are introduced. Finally, a simulation experiment is designed based on the actual data, and the specific experimental methods and processes are provided. The dis-criminant effect of the methods described is verified and compared with the other two networks. The results show that PSO-GRU is more rapid and accurate in fault determination.