基于时间序列预测模型的并联电容器监测系统研究

并联电容器组是电力系统输配电环节中重要的无功补偿装置。本文针对并联电容器组的故障特点进行分析,研制出了并联电容器监测系统。结合并联电容器组的网络拓扑,对其多个支路进行电流波形采集,采用NAR神经网络建立时间序列预测模型,对获取的电流波形进行实时预测和分析,实现对并联电容器组中故障电容器的快速精确定位。通过试验和仿真,验证了所设计的并联电容器监测系统的正确性和有效性。为并联电容器的故障诊断和快速定位提供参考,大大提高了并联电容器的检修效率和智能化水平。

Study on the Monitoring System of Shunt Capacitor Based on Time Series Prediction Model

Shunt capacitor bank is an important reactive power compensation device in the transmission and distribution of power systems.In this paper,the fault characteristics of the shunt capacitor bank are an?alyzed and the shunt capacitor monitoring system is developed.The current waveform acquisition of multi?ple branches is performed with combination of network topology of shunt capacitor bank,the time series pre?diction model is set up by the use of NAR neutral network so to perform real time prediction and analysis to the acquired current waveform and achieve fast and correct location of the faulty capacitor of shunt capaci?tor.The correctness and effectiveness of the designed shunt capacitor monitoring system are verified by test and simulation,which provides reference for the faulty diagnosis and fast location of shunt capacitor and greatly improves the maintenance efficiency and intelligent level of shunt capacitor.