基于贝叶斯估计的故障电流过零点预测

Prediction of zero crossing point of fault current based on Bayesian estimation

目前断路器开断能力难以满足新型电力系统故障电流持续攀升的需求,采用过零点预测的相控开断技术是一种重要的解决手段。研究了基于贝叶斯估计的故障电流过零点预测方法,通过对故障电流衰减时间常数的不同处理策略建立了故障电流模型,提出了扩展卡尔曼滤波、标准卡尔曼滤波和粒子滤波3种算法及其实现,并通过算例仿真对比分析了算法的预测结果。仿真分析表明,通过事先计算衰减时间常数进而采用标准卡尔曼滤波进行状态估计的方法在预测时间、准确度和计算量等方面整体性能较优,适用于控制保护装置的工程实践。

Breaking capacity of circuit breakers can no longer meet the continuous rising demand of fault current in present new power systems.The phrase-controlled fault interruption technology based on zero crossing point prediction is an important solution.Zero crossing prediction method of fault current based on Bayesian estimation is studied in this paper.The fault current model through different processing strategies for the attenuation time constant of fault current is established,and three algorithms such as extended Kalman filter,standard Kalman filter and particle filter as well as their implementation are proposed.The prediction results of the algorithms through numerical simulation are compared and analyzed.The simulation analysis shows that the method of calculating attenuation time constant in advance and then using the standard Kalman filter for state estimation is superior in prediction time,accuracy and amount of calculation,and is suitable for the engineering practice of control and protection device.