Cascade fuzzy neural network based voltage contingency screening and ranking

摘要

A method based on cascade fuzzy neural network (CFNN) comprising of a filter module and ranking module is proposed for online voltage contingency screening and ranking under known but uncertain loads. A new fuzzy performance index, which combines voltage violations and voltage stability margin following a contingency, is proposed for effective voltage security ranking. All the selected contingency cases are first applied to a filter module, which filters out the non-critical contingencies and passes on the critical ones to the ranking module for on-line ranking. The uncertainty associated with loads is modeled by representing them as fuzzy quantities using non-linear membership functions. The performance index is also translated into fuzzy set notations to ensure a flexible and more realistic ranking. Due to the fuzzy nature of the performance index, the proposed method is particularly useful for ranking contingencies, which lie on the boundary between two severity classes. The potential of the CFNN to provide insight into the ranking process, without having to go through the complicated task of rule framing has been demonstrated on IEEE 30-bus test system and a practical 75-bus Indian system.