Synthetic fault diagnostic model of oil-immersed transformers utilizing information fusion

For oil-immersed transformers, the dissolved gas-in-oil analysis (DGA) is one of the most effective techniques for the insulation diagnosis. In-depth analysis of gassing mechanism shows that the location information contained in dissolved gases is far from enough. Therefore, fault location based on DGA only is not reliable. In this paper, the fundamental ideas of information fusion are introduced to the diagnostics. Besides the DGA indication, information from routine electrical tests, service condition, histories of maintenance and expertise, etc. are all integrated in the frame of information fusion. Then, based on ANN and the technique of evidence reasoning, a synthetic fault diagnostic model is put forward, which is multi-level, reliable and with an open architecture. With this model, the process of uncertain reasoning could be simulated more directly and realistically