基于溯因推理网络的电网故障预测方法研究

Research on Power Grid Fault Forecast Based on Abductive Reasoning Network

电网诊断通常都是故障发生后,根据故障后产生的信息来推断故障,为了能在故障发生前进行预防,提出了模型预测(ModelPrediction,MP)和溯因推理网络(AbductiveReasoning Network,ARN)方法预测电网故障,模型预测利用电力系统中历史数据来预测电网无故障运行时的数据,与电网实际运行时的数据作对比,计算差值。溯因推理网络(ARN)使用带有简单低阶多项式的节点函数的几层网络,能够处理预测数据和相应的候选故障之间的复杂关系。模型预测和溯因推理网络方法相结合,能在保护装置和断路器动作前进行故障定位,具有故障预警功能。仿真结果表明该诊断系统与神经网络方法相比较可以获得快速、准确的诊断结果。

Fault diagnosis is usually judged according to the information after fault happens. In order to prevent the fault before the occurrence of the fault, model prediction(MP) and abductive reasoning network(ARN) are proposed to predict the power grid fault. MP predicts the trouble-free operation of the data of the power grid using the historical data, compares with the actual grid runtime data and calculates the difference. ARN handles complicated relationships between data processing and the corresponding candidate fault section using a hierarchical network with several layers of function nodes of simple low-order polynomials. The combination of model prediction and abductive reasoning network can locate the fault before the protection device and circuit breaker acts, and has the function of fault early warning. The simulation results show that the diagnosis system can obtain rapid and accurate diagnosis results compared with the neural network method.