摘要
低压断路器的安全可靠是电力系统能否稳定运行的关键一环,因此对断路器进行退化趋势预测和剩余寿命评估具有重要意义。基于鲸鱼优化算法(whale optimization algorithm,WOA)和双向长短期记忆神经网络(bidirectional long short-term memory,BiLSTM)提出了一种断路器操动机构剩余寿命的预测方法,首先采用Pearson相关系数法对获得的原始监测数据进行筛选,选择与断路器开断次数相关度较高的数据作为关键退化特征量,基于主成分分析法进行数据融合获得能够综合表征断路器运行状态的健康指数;随后使用滑动时间窗的方法对健康指数时间序列进行重构,再通过WOA-Bi LSTM寻优获得的最佳模型对健康指数进行时间序列预测,从而获得断路器未来多步的退化趋势;最后再根据设定的失效阈值,确定断路器操动机构的剩余寿命。实例验证表明,该文提出的混合预测模型预测精度最高可达96.43%,相比于其他传统预测模型显著提高,对于断路器的实际运维工作具有一定的指导意义。
The safety and reliability of low-voltage circuit breakers are the key to the stable operation of power systems,so it is of great significance to predict the degradation trend and evaluate the remaining useful life of circuit breakers.In this paper,based on whale optimization algorithm(WOA)and bidirectional long short-term memory(Bi-LSTM)neural network,a method for predicting the remaining useful life of the circuit breaker operating mechanism is proposed.Firstly,the Pearson correlation coefficient method is used to screen the original monitoring data and the data with higher correla-tion with the number of circuit breaker openings were selected as the key degradation feature.The health index which can comprehensively characterize the running state of circuit breaker can be obtained by data fusion based on principal com-ponent analysis(PCA).The time series of health index is reconstructed by the method of sliding time window,and then the best model obtained by WOA-BiLSTM optimization is used to predict the time series of health index,so as to obtain the multi-step degradation trend of circuit breaker in the future.Finally,according to the set failure threshold,the remain-ing useful life of the operating mechanism of circuit breaker is determined.Example validation shows that the hybrid prediction model proposed in this paper has a prediction accuracy of up to 96.43%,which is significantly improved com-pared to other traditional prediction models.So it has certain guiding significance for the actual operation and maintenance of circuit breakers.
作者
李家豪
王青于
范玥霖
史石峰
彭宗仁
曹培
徐鹏
LI Jiahao;WANG Qingyu;FAN Yuelin;SHI Shifeng;PENG Zongren;CAO Pei;XU Peng(State Key Laboratory of Electrical Insulation and Power Equipment,Xi’an Jiaotong University,Xi’an 710049,China;Electric Power Science Research Institute of Shanghai Electric Power Company,Shanghai 200080,China)
出处
《高电压技术》
EI
CAS
CSCD
北大核心
2024年第1期250-262,共13页
High Voltage Engineering
基金
国网上海市电力公司研发项目(520940230003)。
关键词
低压断路器
退化趋势
剩余寿命
双向长短期记忆网络
鲸鱼优化
low voltage circuit breaker
degradation trend
remaining useful life
bidirectional long short-term memory network
whale optimization
作者简介
李家豪,1999-,男,博士生主要从事高电压与绝缘技术、电力设备状态评估与电力设备快速计算方法研究等方面的工作,E-mail:2868259710@qq.com;通信作者:王青于,1991-,女,博士,助理教授,硕导主要从事高端交直流套管关键技术、绝缘结构优化设计及多物理场仿真计算、电力设备状态评估与剩余寿命预测等方面的工作,E-mail:wqy.1991624@xjtu.edu.cn。
