An associative binary particle swarm optimization for the diagnosis of transformer failure

Abstract

In this paper an associative binary particle swarms optimization (BPSO) for the diagnosis of transformer failure. In this approach transformer oil gas have been considered for the fault diagnosis so that proper functionality of transformer can be enhanced and the efficiency of transformer can be improved. For this dissolve gas analysis (DGA) and IEC standards have been used for weight assignment of different gas ratios. Rule mining have been applied where these standards fails in the weight assignments. Finally based on the rules associates with different gas ratios have been analyzed separately for each clusters. Finally based on BPSO faults have been diagnosed in several iterations. The results clearly indicate that our approach has better fault diagnosis and individual gas associations.