基于改进灰狼算法优化SVR的避雷器阻性电流预测
Improved Grey Wolf Algorithm to Optimize SVR for Resistive Current Prediction of Arrester
杨政1
作者信息
- 1. 中国南方电网有限责任公司超高压输电公司南宁局 南宁 530021
- 折叠
摘要
避雷器运行电压下阻性电流大小及其变化反映了避雷器的运行状态.针对避雷器带电测试仅每年开展一次数据较少的问题,论文提出了一种基于改进灰狼算法优化SVR的避雷器阻性电流预测方法.基于环境温度对阻性电流具有较大影响和测量时相间干扰的考虑,该预测方法以避雷器相序、历史阻性电流、环境温度和温度差值作为特征输入.通过Logistic混沌映射初始化种群和采用非线性收敛因子改进了基本的灰狼优化算法.利用某500kV避雷器三相近10年带电测试数据进行仿真分析,验证了预测方法的准确性与可行性.
Abstract
The magnitude and variation of arrester resistive current under the operating voltage reflect the state of arrester.Aiming at the problem that the data samples of on-line measurement are small,the support vector regression(SVR)optimized by improved grey wolf optimizer(IGWO)is proposed to predict resistive current of arrester.In view of the fact that the resistive current is significantly affected by ambient temperature and interphase interference,the prediction method takes the phase,historical resis-tive current,ambient temperature and temperature difference as characteristic inputs.An improved grey wolf optimizer based on Lo-gistic chaotic mapping for population initialization and nonlinear convergence factor is adopted.Three phase on-line measurement datas of a 500kV arrester in recent 10 years are used for modeling and analysis.The results of calculation example show that the pro-posed method is accuracy and feasible.
关键词
阻性电流预测/改进灰狼优化/支撑向量回归/混沌映射/金属氧化物避雷器Key words
prediction of resistive current/improved grey wolf optimizer/support vector regression/chaotic mapping/metal oxide arrester引用本文复制引用
出版年
2024
国家科技期刊平台
NSTL
万方数据