Dynamic fault recognition for power transformers

At present, routine diagnostic methods just can reflect the state of the faults which is represented by a power transformer at the certain time. Obviously, it is not enough, as the fault omens caused by different reasons are so similar at any time that it is difficult to distinguish between them. However, it is still possible for these faults to be recognized, so long as the increasing tendency of the characteristic gases is observed. Because there are so many differences between each developing course of all kinds of faults, owing to the dissimilar mechanisms of the faults. In this paper, by analyzing a large number of fault examples statistically, a new dynamic fault recognition method is presented according to the operating characteristics of transformers. It is proved that this method overcomes the shortage of static analysis, and improves the accuracy of the recognition distinctly, and also help to forecast the developing trend of the fault preferably