基于支持向量机的电力系统状态估计多类型数据异常检测

Multi-type data anomaly detection in power system state estimation using support vector machine

为了解决异常数据严重影响电力系统状态估计性能的问题,提出了一种基于支持向量机(SVM)的电力系统预测辅助状态估计(FASE)多类型数据异常检测方法。首先,针对传统FASE的预测准确率欠佳的问题,提出了基于极限学习机的FASE方法,并利用SVM并基于预测数据、量测数据与估计值,实现了对坏数据、负荷突变和单相接地等多种类型的数据异常检测。其次,针对惩罚因子和核函数参数会影响分类精度的问题,提出采用灰狼算法对SVM参数进行优化,在兼顾计算速度的同时提高了数据异常检测的准确率。最后,在IEEE 33和丹麦DTU 7K 47节点主动配电网系统上进行仿真测试,所提方法在正常工况下提升26.08%与26.76%,计算速度提升46.05%,在数据异常情况下准确率综合提升32.04%与29.27%,结果表明,所提方法具备较强的通用性与实时性,可以有效地检测电力系统中各种类型的数据异常,并提高状态估计的性能。

To address performance issues in state estimation caused by anomalous data in power systems,this paper proposes a multi-type data anomaly detection method for power system state estimation based on support vector machine(SVM).Firstly,extreme learning machine is proposed to enhance prediction accuracy in forecasting-aided state estimation(FASE).Then,SVM is utilized to detect various types of data anomalies,including bad data,load changes,and single-phase grounding,by incorporating prediction data,measurement data,and estimated values.To optimize SVM parameters and improve classification accuracy,the grey wolf algorithm is suggested to tackle the issue of penalty factor and kernel function parameters.Finally,simulations are conducted on real distribution systems,specifically the IEEE 33 and DTU 7K 47 systems,using data from different scenarios.The proposed method achieves accuracy improvement of 26.08%and 26.76%and calculating speed improvement of 46.05%under normal operating conditions,and a comprehensive enhancement of 32.04%and 29.27%in accuracy under data abnormality scenarios.The results demonstrate that the proposed method is highly generalizable and real-time,effectively detects various types of data anomalies,and enhances the performance and accuracy of state estimation.