A Novel Substation Fault Diagnosis Approach Based on RS and ANN and ES

With the aid of rough set (RS) and artificial neural networks (ANN), expert system (ES) can extend its capability in knowledge representation and acquisition as well as parallel reasoning. However, ANN still can´t completely replace ES due to its inherent flaws such as learning difficulty and interpreting disability, etc. Hence, in the paper we would incorporate ANN with ES to overcome each deficiency and exert each excellence. In addition, rough set is applied to serve for pretreatment unit of ANN so as to simplify networks structure and improve learning quality. Thus, on the one hand, the problems such as inference complexity and time lengthiness of conventional ES are overcome. On the other hand, the flaws such as the incompleteness or error of ANN input data are also resolved well. In the end, a simulation trial in substation fault diagnosis shows the availability of the method