Research on Real-time Hot-spot Temperature Prediction Method for Transformer Based on Improved Support Vector Machine

LIAO Caibo; RUAN Jiangjun; WEI Chao; LU Yuncai; WU Peng; LI Jiansheng

doi:10.13296/j.1001-1609.hva.2018.12.026

您当前的位置：

首页 >

文章列表页 >

Research on Real-time Hot-spot Temperature Prediction Method for Transformer Based on Improved Support Vector Machine

更新时间：2025-04-26

- Research on Real-time Hot-spot Temperature Prediction Method for Transformer Based on Improved Support Vector Machine
- High VoltageApparatus Vol. 54, Issue 12, Pages: 174-179(2018)
- 作者机构：
  
  1. 国网江苏省电力公司电力科学研究院
  2. 武汉大学电气工程学院
- 作者简介：
- 基金信息：
- DOI：10.13296/j.1001-1609.hva.2018.12.026
  CLC： TM41
- Published：2018
- 稿件说明：
移动端阅览
[1]廖才波,阮江军,蔚超,陆云才,吴鹏,李建生.基于改进支持向量机的变压器实时热点温度预测方法研究[J].高压电器,2018,54(12):174-179.

LIAO Caibo, RUAN Jiangjun, WEI Chao, et al. Research on Real-time Hot-spot Temperature Prediction Method for Transformer Based on Improved Support Vector Machine[J]. High voltageapparatus, 2018, 54(12): 174-179.
[1]廖才波,阮江军,蔚超,陆云才,吴鹏,李建生.基于改进支持向量机的变压器实时热点温度预测方法研究[J].高压电器,2018,54(12):174-179. DOI： 10.13296/j.1001-1609.hva.2018.12.026.

LIAO Caibo, RUAN Jiangjun, WEI Chao, et al. Research on Real-time Hot-spot Temperature Prediction Method for Transformer Based on Improved Support Vector Machine[J]. High voltageapparatus, 2018, 54(12): 174-179. DOI： 10.13296/j.1001-1609.hva.2018.12.026.

摘要

文中提出了一种基于改进支持向量机的变压器实时热点温度预测方法。该方法以变压器负载电流、环境温度和顶层油温为特征值

将变压器热点温度预警与预测相结合

目的在于实现基于近期历史运行数据的变压器实时热点温度预测。文中研究分析了基于改进网格搜索、遗传算法及粒子群算法的支持向量机在变压器短期热点温度预测中的应用效果

结果表明

3种改进方法均可有效提升预测精度

但基于改进网格搜索算法的支持向量机预测效率最高。最后

以某35 k V油浸式变压器实时热点温度数据为样本集

文中所述方法可有效实现变压器短期热点温度预警预测

预测准确度高于90%。

Abstract

A prediction method of hot-spot temperature for transformer based on improved support vector method（SVM）is proposed in the paper.Taking load current

environment temperature and top layer oil temperature as characteristic parameters

the prediction method aims to realize real-time temperature prediction based on the recent running data.The method possesses transformer hot-spot temperature early warning and prediction function.This paper evaluates the application effects of support vector methods based on improved grid search（GS）

genetic algorithm（GA）and particle swarm optimization（PSO）in the short term hot-spot temperature prediction of oil-immersed transformer.The results indicate that these three methods all do benefit to the prediction accuracy

while GS-SVM performs best in prediction efficiency.At last

the validity and feasibility of this method is verified by the test application on a35 k V oil-immersed transformer

and the prediction accuracy of short term hot-spot temperature is higher than 90%.

关键词

Keywords

references

Views

269

下载量

CSCD

Alert me when the article has been cited

提交

Tools

Publicity Resources

Combined Detection Technology of Trace Acetylene Gas Photothermal Interference and Multi-harmonic Signal

Insulation Condition Evaluation of 750 kV Power Transformers Based on Frequency Domain Spectroscopy

Aging Diagnosis and Remaining Life Prediction of Solid Insulation for Oil-immersed Transformers

Study on Prediction of Top Oil Temperature for Power Transformer Based on Kalman Filter Algorithm

Research of Strategies Based on FVM for Large Oil-immersed Power Transformer’s Fan Control

Related Author

WANG Chao

PAN Xingya

QIAO Ruixun

ZHOU Yang

SHAO Jianwei

FENG Yu-Chang2

YU Zheng2

WANG Shi-qiang1

Related Institution

School of Electrical and Automation, Wuhan University

Northwest China Grid Company Limited,Xi′an

State Key Lab.of Electrical Insulation for Power Equipment,Xi′an Jiaotong University,Xi′an

Fujian Electric Power Co.,Ltd.

School of Electrical and Information Engineering,Xihua University

⁰