基于数据驱动的500 kV高压并联电抗器过流误报警在线判别方法研究

为提升500 kV并联电抗器在线监测的准确性,提出了基于逻辑回归和主成分分析的并联电抗器过流误报警在线判别方法。首先,分析了并联电抗器所在母线电压互感器以及电抗支路电流互感器测量误差对过流报警条件概率的影响。其次,定义逻辑回归拟合函数的中心点及不确定域为故障检测的二维关键指标,并利用主成分分析法生成一维综合指标。最后,通过将检测日综合指标与设备正常运行工况下综合指标进行对比,对检测日过流报警的正确性进行判别。该方法只需使用高压并联电抗器所在母线的电压测量有效值和监控系统所报出的过流报警信号,无需额外增加测量设备。以某实际500 kV高压并联电抗器作为算例进行的仿真实验,验证了该方法的有效性。

Online identification of error overcurrent alerts of 500 kV high voltage shunt reactor based on data-driven

In order to increase the accuracy of the online monitoring of shunt reactor, an online identification method for false overcurrent alarm of 500 kV shunt reactor based on the logical regression and Principal Component Analysis (PCA) method is proposed. Firstly, the influences of measurement error of the transformers including the PT in the bus connected to shunt reactor and the CT in the reactor branch on the conditional probability of overcurrent alarms are analyzed. Then the central point and the uncertain domain of the logistic regression function are defined as the two-dimension key indicators for the detection of false alarms, and PCA method is used to generate the one-dimension comprehensive index. Finally, the validity of overcurrent alarm is judged by comparing the comprehensive index of the detection day with that under normal conditions. The proposed method only uses the RMS of voltage in the bus connected to shunt reactor and the overcurrent alarm information obtained from monitoring system, so no additional measurement equipment is needed. The effectiveness of the method is verified by case study of a practical 500 kV high voltage shunt reactor. This work is supported by Key Research and Development Program of Sichuan Department of Science and Technology (No. 2017GZ0163) and Science and Technology Program of State Grid Corporation of China (Technology Research of Supercomputer-based Power System Operation Mode Calculation Platform Support).