A class of hybrid intelligent system for fault diagnosis in electric power systems

Abstract

In recent years, significant progress has been achieved in the development of a class of artificial intelligent systems. This growing new class of technology has been applied successfully to a wide variety of applications. It is now becoming increasingly evident that in many cases best results can be obtained through the use of combination of such intelligent systems. In this paper, a class of hybrid intelligent systems has been applied in diagnosing a variety of faults in a range of electric power system equipments. The nature of the faults are diverse and no single intelligent algorithm is able to classify the faults. However, it is shown, in this paper, that a combinatorial intelligent system based on neuro-fuzzy, neuro-expert and fuzzy-expert algorithms can be successfully applied in the detection of a number of faults in a range of equipments. Test results using measurements on in-service power transformers in the Australian state of New South Wales and on a distribution feeder belonging to CEMIG, Brazil have been used to show the effectiveness of the proposed class of hybrid intelligent systems in equipment fault diagnosis.