Research On Fault Diagnosis Method Of Dry - Type Transformer Based On Soft - Sensing Technology

Dry-type transformer is one of the key equipment of power system, once the fault can directly affect the operation of the power system, its service life mainly depends on the insulation performance, when dry, will cause different degrees of temperature changes, such as not timely detection, will lead to insulation breakdown, power outages, explosion, and so on.This paper through theoretical analysis, simulation and experimental validation of dry-type transformers temperature under different fault conditions were studied systematically. First, the establishment of a three-dimensional finite element model of the electromagnetic hot dry-type transformers, combined with theoretical knowledge on the distribution of dry variable temperature field were described in detail, to find the most hot dry change of position, and through the simulation with the measured temperature value temperature value than the right, to verify the correctness of the simulation. Secondly, on this basis, the use of direct measurement and soft measurement method of combining dry-type transformers temperature information to be collected through the soft measuring technique point of failure information to simulate dry-type transformers, and actual lists of five kinds of dry-type transformers operating conditions, these five kinds of operating conditions were simulated, and set the 45 monitoring points for temperature data collection, these data as training data dry-type transformer fault diagnosis model.finally, this paper proposes a fault diagnosis algorithm based on genetic algorithm and BP neural network. The algorithm has the ability of global optimization, and improves the convergence speed and precision of BP neural network, and verifies the feasibility of genetic algorithm to optimize BP network. The accuracy of the proposed algorithm is 92%, which is verified by experiments.