基于预处理模糊Petri网与改进遗传算法的电网故障诊断方法

为了提升电网复杂故障时的态势感知能力,提出了基于预处理模糊Petri网与改进遗传算法的电网故障诊断方法.首先,基于层次化、动态建模的思想,在建立元件一般故障诊断模型的基础上,引入浮动库所、浮动弧、浮动变迁的概念来合理的体现主保护、断路器以及后备保护之间的逻辑关系,建立了元件复杂故障时的诊断模型;其次,挖掘故障信息源的特性进行预处理,依据故障类型动态建立相应Petri网诊断模型;再次,利用改进的遗传算法对模型中的参数进行了训练、优化;最后,探讨了该Petri网故障诊断模型的容错性、通用性.对算例系统仿真的结果表明:该方法突出了故障诊断过程的层次性、可理解性,提升了诊断模型的透明性、改善了诊断模型的易维护性,在电网复杂故障的情况下诊断结果仍具有较高的准确度,并呈现有较好的容错性.

Method of power grid fault diagnosis based on information-optimized dynamic modeling fuzzy Petri net and improved genetic algorithm

In order to improve situational awareness of complex faults in power grid, a fault diagnosis method based on pretreatment fuzzy Petri net and improved genetic algorithm is proposed. Firstly, based on the idea of hierarchical and dynamic modeling, the concepts of floating storehouse, floating arc and floating transition are introduced to reasonably embody the logical relationship among main protection, circuit breaker and backup protection, and the diagnosis model of complex fault of components is established. Secondly, the characteristics of fault information sources are excavated and preprocessed. According to the fault type, the corresponding Petri net fault diagnosis model is established dynamically. Thirdly, the parameters of the model are trained and optimized by using the improved genetic algorithm. Finally, the fault tolerance and generality of the Petri net fault diagnosis model are discussed. The simulation results of an example system show that the method highlights the hierarchy and comprehensibility of fault diagnosis process, improves the transparency of the diagnosis model, improves the maintainability of the diagnosis model, and still has high accuracy and good fault tolerance in the case of complex faults in power grid.