基于模型的配电网故障诊断关键问题研究

Key issues with model-based diagnosis in distribution network

现有配电网故障诊断方法主要是根据故障后保护装置产生的报警信息,断路器的状态变化信息来推断可能故障位置和类型,属于后验故障诊断方法。基于模型诊断方法(Model-based diagnosis,MBD)则直接利用电压电流等量测量来判断故障元件,可以在保护装置和断路器动作前进行故障定位,具有一定的故障预警功能。给出一个完整的基于模型的配电网诊断方案,采用优化的离散二进制粒子群算法(BPSO,Binary Particle Swarm Optimization)求取冲突集的最小碰集,诊断识别过程中以贝叶斯后验概率形式量化了候选诊断的衡量标准,从而降低诊断过程中的不确定度。最后以某10kV配电网为诊断实例,通过实际建模、编程和实验,取得了良好的效果。

Most of the diagnosis systems used in the electric distribution network are based on the action information of circuit breakers and protective relays, which are also called the posteriori diagnostic method. The model-based diagnosis method could directly use the measurements of voltage and current to determine faulty components and locate the fault before the operation of protection device and breaker, so it has a certain early fault warning function. In this paper, a diagnosis plan based on MBD is given. The minimal hitting set （that is, candidate diagnosis） could be calculated from the obtained minimal conflict set based on improved BPSO. The posteriori fault probabilities of each diagnosis candidate could be calculated by the Bayesian probability method. So, we could quantify the measure standard of the diagnosis identification in the form of the posteriori fault probabilities, which lowers the uncertainty of diagnosis. Finally, an example of actual electric distribution network is presented to verify the feasibility and validity of this method through modeling, programming and testing.