A Novel Nondestructive Testing Method for Dielectric Loss Factor of Transformer Oil Based on Multifrequency Ultrasound
Baoliang Li, Qu Zhou, Yupeng Liu, Jianxing Chen
Abstract
Dielectric loss factor is an important index to evaluate the quality of transformer oil, which is very sensitive to the response of polar impurities. As a nondestructive testing technology, ultrasonic can identify polar impurities through attenuation characteristics and then determine dielectric loss. In this article, a novel method for detecting dielectric loss factor of transformer oil based on multifrequency ultrasound and particle swarm optimization-Elman neural network (PSO-ENN) forecasting model was proposed. The analysis of the forecast results and precision shows that the dielectric loss factor can be effectively forecast via this method. The experiment verifies that the ultrasonic effect is nondestructive to transformer oil, which lays a foundation for the development of online monitoring system.