输电线路的雷电过电压的识别方法

Identification method of lightning over voltage of transmission line

雷电是一种落雷时间、地点、能量均无法预测的随机产生的自然现象,这对雷电过电压识别带来了难度,针对传统识别方法使用单一特征信息作为判断依据导致的识别率低等问题,本文针对时域波形、雷击波头特征以及HHT时频谱这三个方面进行特征提取,建立的了普通短路过电压、感应雷击过电压、直击雷过电压、反击雷过电压以及绕击过电压这5种过电压类型与与特征对应关系,并提出一种结合协同量子粒子群优化算法和最小二乘支持向量机的雷击过电压识别模型,通过仿真研究,与基于QPSO-LSSVM算法进行对比,本文研究的识别模型的准确率提高了6.46%。

Thunder is a natural phenomenon in random locations,time,lightning energy are unpredictable,so it is difficult to get the lightning over-voltage identification. Due to low recognition for the traditional identification methods using single feature information as the basis of judgment result rate,this article extracts these three features like the time domain waveform,lightning wave head features and HHT spectrum to establish the common short pass by induction lightning overvoltage,voltage,lightning over-voltage,counter lightning overvoltage and shielding failure overvoltage of the 5 types of overvoltage and their feature correspondence and to proposed a combination of lightning quantum particle swarm optimization algorithm and least squares support vector machines over voltage the recognition model.Through the simulation and compared with the algorithm based on QPSO-LSSVM,the accuracy rate of recognition model of this study is increased by 6. 46%.