•  
  •  
 

Keywords

OLTC, neural network, response surface model, fault simulation, fault identification

Abstract

In order to precisely identify the spring energy storage failure in an onload tap changer (OLTC), an identification method is developed for spring energy storage failure of the OLTC based on the neural network response surface model. Firstly, the fault simulation model of the OLTC was established through the finite element method. Then, the training samples of the response surface model were generated from the uniform experiments and simulations, and the neural network response surface model was therefore constructed by training these samples. Finally, the mechanical parameters of the spring energy storage deficiency were identified using the multiobjective identification algorithm constructed by desirability function, and the identification results of spring insufficient energy storage faults of the UCL type OLTC was validated by simulation. The Results show that the fault of spring insufficient energy storage can be identified accuratelyvia the neural network response surface model. The maximum relative error between the identified result and the reference data is 3.93%, which can verifie the effectiveness of this method.

DOI

10.19781/j.issn.1673-9140.2021.03.025

First Page

203

Last Page

210

Share

COinS