The feature analysis for transformer reliability assessment based on the improved artificial fish optimization algorithm and BP neural network

The power transformer is the most important equipment in the power system, so the condition-based maintenance of the transformer has an important significance for the safe and stable operation of the power grid, and condition assessment provides an important basis for the condition-based maintenance. In this paper, the calculation principle of the reliability assessment method is analyzed, and the contribution of each feature value is calculated. Then fit the reliability of a single transformer based on support vector machine, optimizing the weights and threshold of the BP neural network through improved artificial fish optimization, which does not only improve the computational rate, but also the reliability is more accurate. It has an important guiding significance for the Power Supply Bureau to provide the condition assessment data, the data with a contribution of large degree should be collected forcibly, so that the data for transformer reliability assessment is the most simplified, improving the assessment rate.