Dynamic Model Updating for On-load Tap Changer Based on Neural Network Response Surface Model

LIU Zhiyuan; CHEN Haijun; MIAO Hui; ZHAO Xinyang; CHEN Dapeng

doi:10.13296/j.1001-1609.hva.2021.06.022

您当前的位置：

首页 >

文章列表页 >

Dynamic Model Updating for On-load Tap Changer Based on Neural Network Response Surface Model

更新时间：2025-04-26

- Dynamic Model Updating for On-load Tap Changer Based on Neural Network Response Surface Model
- High VoltageApparatus Vol. 57, Issue 6, Pages: 171-178(2021)
- 作者机构：
  
  1. 国网宁夏电力有限公司
  2. 国网宁夏电力有限公司检修公司
  3. 南京航空航天大学能源与动力学院
  4. 南京优能特电力科技发展有限公司
- 作者简介：
- 基金信息：
- DOI：10.13296/j.1001-1609.hva.2021.06.022
  CLC： TP183;TM403.4
- Published：2021
- 稿件说明：
移动端阅览
LIU Zhiyuan, CHEN Haijun, MIAO Hui, et al. Dynamic Model Updating for On-load Tap Changer Based on Neural Network Response Surface Model[J]. High voltageapparatus, 2021, 57(6): 171-178.
DOI：

LIU Zhiyuan, CHEN Haijun, MIAO Hui, et al. Dynamic Model Updating for On-load Tap Changer Based on Neural Network Response Surface Model[J]. High voltageapparatus, 2021, 57(6): 171-178. DOI： 10.13296/j.1001-1609.hva.2021.06.022.

摘要

为提高模型修正的精度和效率

发展了一种基于神经网络响应面的动力学模型修正方法

将该方法用于有载分接开关底座有限元简化模型的修正。首先

根据底座简化模型的误差来源确定待修正参数及其设计空间

采用均匀设计生成神经网络响应面模型的训练样本;然后

通过对样本数据的训练构造修正参数与动力学响应关系的神经网络响应面模型;最后

基于神经网络响应面模型

采用线性组合的多目标优化方法对底座简化模型进行修正

并将修正结果与超模型仿真数据进行对比。研究表明

基于神经网络响应面模型修正的底座简化模型的动力学特性与超模型结果基本一致

验证了该方法的有效性。

Abstract

In order to improve the accuracy and efficiency of model updating

a method is developed for dynamic model updating based on neural network response surface model.By this method

the simplified finite element model of the on-load tap changer was updated.First

the parameters to be updated and their design space were determined by the error sources of the simplified base model

and the training samples of the neural network response surface model were generated by uniform design.Then

the neural network response surface model for the relation between updating parameters and dynamic responses was established by training the sample data.Finally

the simplified base model was updated by the multi-objective optimization with linear combination based on the neural network response surface model

and the updated results were compared with the simulation results of the super model.The results show that the dynamic characteristics of the simplified base model updated by the neural network response surface model is basically consistent with that of the super model

which validates the effectiveness of this method.

关键词

Keywords

references

Views

下载量

CSCD

Alert me when the article has been cited

提交

Tools

Publicity Resources

Research on Capacity Configuration Technology of Distributed Energy Storage Based on Multi-objective Constraints

Opening and Closing State Monitoring for GIS Disconnector Based on Voiceprint Characteristic Analysis

Optimization Design on Spring Contact Finger Structure Based on Finite Element Method and Neural Network Method

Simulation Design and Optimization of Permanent-magnet-motor Operating Mechanism for Vacuum Circuit Breaker with Long Clearance

Study on Envelope Extraction Method of Vibration Signal of OLTC Based on Savitzky-Golay Filter

Related Author

GAO Xiang

AN Yulun

ZHANG Yitao

CHENG Peng

ZHOU Wei

CHEN Fuguo

JIANG Zhuangqiang

LI Zhongqi

Related Institution

State Grid Jibei Electric Power Co, Ltd, Langfang Power Supply Company

Pinggao Group Co., Ltd.

School of Electrical Engineering, Xi'an Jiaotong University

College of Electrical Engineering, Shanghai University of Electric Power

State Key Laboratory of Electrical Insulation and Power Equipment, Xi'an Jiaotong University

⁰