Using simulation and rough set learning to detect fault location in distribution network

Fault occurs owning to a variety of reasons in distribution network, such as equipment failure, overloading, tree, vehicle etc. It is very important for utility to detect the fault location as quickly as possible for helping to reduce the outage time. This paper proposed a method for distribution network fault location diagnosis which employs simulation and rough set learning. Based on the topology structure of distribution network and the probability model of equipment failure, the simulation model is firstly built for training the sample data. The rough set theory is applied to establish the rules of the relationship between outage zone and the equipment failure. And the enhanced learning process is used to improve the completeness of rules library. The numerical testing results are also presented to illustrate the method.