Relative condition ranking of transformers using evidence theory

In this paper, an evidential reasoning (ER) approach is employed to assess the condition of a power transformer, which also can be extended to rank the condition of a group of transformers for system maintenance purposes. An ER algorithm is briefly introduced, which are utilised to combine pieces of evidence supporting all possible diagnoses and deal with uncertainty and contradictions between the various information resources. Then the transformer condition assessment problem is mapped into a multiattribute decision making (MADM) framework and a specific application is fully discussed to illustrate the functionalities of an ER algorithm. Based on the outcomes of the ER approach, system operators can obtain the overall evaluations for the interested units, which through ER can be ranked in order of severity and risk for system operations.