On-line power system fault estimation using connectionist models

This paper proposes a new connectionist (or neural network) expert diagnostic system for on-line fault diagnosis of a power substation. The connectionist expert diagnostic system has similar profile of an expert system, but can be constructed much more easily from elemental samples. These sample indicate the association of fault with their protective relays and breakers, as well as the bus voltage and feeder currents. Through an elaborately designed structure, these two types of alarm signals are processed by different connectionist models. The outputs of the connectionist models are then integrated to provide the final conclusion with confidence level. The proposed approach has been practically verified testing on a typical Taiwan Power (Taipower) secondary substation. The test results suggest our system can be implemented by various electric utilities with relatively low customization effort.