故障预测下电能计量仪表参数优化控制研究  

Research on Optimal Control of Power Metering Instrument Parameters under Fault Prediction

作　　者：伍莎莎 韦淑敏 谢雄 郭芒 WU Shasha;WEI Shumin;XIE Xiong;GUO Mang(State Grid Hunan Power Supply Service Center(Metrology Center),Changsha 410007,China)

机构地区：[1]国网湖南供电服务中心(计量中心),湖南长沙410007

出　　处：《自动化仪表》2025年第2期80-84,91,共6页Process Automation Instrumentation

摘　　要：电能计量仪表容易受到异常故障的干扰,使得参数控制过程失真。为此,提出了故障预测下电能计量仪表参数优化控制方法。采用层次聚类算法挖掘相关的故障数据,剔除数据挖掘结果中存在的异常数据。利用智能边缘计算技术,将故障数据输入在线贯序极限学习机中,对不同时刻的电能计量仪表故障开展精准预测。以预测结果为基础,引入多聚合方法确定故障预测约束下的相关控制参数,并结合动态平衡方程实现电能计量仪表参数优化控制。试验结果表明,所提方法的最优控制曲线与实际控制曲线几乎一致,电能计量仪表故障预测误差低,控制成功率高达98%。该方法有着广阔的应用前景。Power metering instrument are easily disturbed by abnormal faults,making the parameter control process distorted.For this reason,an optimal control method of power metering instrument parameters under fault prediction is proposed.Hierarchical clustering algorithm is used to mine relevant fault data,and the abnormal data existing in the data mining results are eliminated.Intelligent edge computing technology is utilized to put the fault data into an online sequence extreme learning machine to accurately predict the faults of power measuring instruments at different moments.Based on the prediction results,a multi-aggregation method is introduced to determine the relevant control parameters under the constraints of fault prediction and combined with the dynamic equilibrium equations to realize the optimal control of power metering instrument parameters.The experimental results show that the optimal control curves of the proposed method are almost the same as the actual control curves,and the prediction error of power measuring instrument faults is low,and the control success rate is as high as 98%.This method has a wide range of application prospects.

关 键 词：电能计量仪表 层次聚类算法 故障数据挖掘 异常数据剔除 在线贯序极限学习机 多聚合方法 动态平衡方程 

分 类 号：TH-70[机械工程]