发电技术 ›› 2022, Vol. 43 ›› Issue (2): 313-319.DOI: 10.12096/j.2096-4528.pgt.21006

• 智能电网 • 上一篇    下一篇

基于Logistic快速最小误差熵算法的配电变压器停电预测

许中1, 栾乐1, 莫文雄1, 罗思敏1, 叶宗林2, 陈超2, 赖轩达2, 解明辉2   

  1. 1.广东电网有限责任公司广州供电局, 广东省 广州市 510000
    2.西安交通大学电气工程学院, 陕西省 西安市 710049
  • 收稿日期:2021-06-30 出版日期:2022-04-30 发布日期:2022-05-13
  • 作者简介:许中(1986),男,硕士,高级工程师,主要研究方向为电能质量,348867958@qq.com
    栾乐(1982),女,硕士,高级工程师,主要研究方向为设备状态评价,149529958@qq.com
    莫文雄(1971),男,硕士,教授级高级工程师,主要研究方向为电气工程,gzmwx@139.com
    罗思敏(1988),男,硕士,高级工程师,主要研究方向为智能配电网,本文通信作者,rushlsm@163.com
    叶宗林(1990),男,博士,主要研究方向为数据挖掘、过程控制和工业自动化,yezonglin1990@gmail.com
    陈超(1995),男,硕士研究生,主要研究方向为配电网大数据分析与态势感知研究,610825983@qq.com;
    赖轩达(1999),男,硕士研究生,主要研究方向为电力大数据,lxd593@stu.xjtu.edu.cn
    解明辉(1998),男,硕士研究生,主要研究方向为电力大数据,619425976@qq.com
  • 基金资助:
    中国南方电网责任有限公司科技项目(GZHKJXM20180068)

Distribution Transformer Outage Prediction Based on Logistic Fast Minimum Error Entropy Algorithm

Zhong XU1, Le LUAN1, Wenxiong MO1, Simin LUO1, Zonglin YE2, Chao CHEN2, Xuanda LAI2, Minghui XIE2   

  1. 1.Guangzhou Power Supply, Guangdong Power Grid Co. , Ltd. , Guangzhou 510000, Guangdong Province, China
    2.School of Electrical Engineering, Xi’an Jiaotong University, Xi’an 710049, Shaanxi Province, China
  • Received:2021-06-30 Published:2022-04-30 Online:2022-05-13
  • Supported by:
    the Science and Technology Project of China Southern Power Grid Co., Ltd(GZHKJXM20180068)

摘要:

为了提高配电变压器停电预测的速度和准确性,提出了一种基于Logistic快速最小误差熵算法的配变停电预测方法。在最小熵回归算法的基础上,提出了快速最小误差熵算法,基本保持了最小熵回归的回归效果,并且显著地减少了算法的运行时间;针对配变停电预测适用Logistic回归的情况,提出了基于Logistic的快速最小误差熵回归算法,选取配电变压器重过载时长、最大有功负载率、平均有功负载率、平均三相不平衡度以及重三相不平衡度作为配变停电预测的特征变量数据,建立了配电变压器停电预测模型,实验预测结果优于Logistic回归。

关键词: Logistic快速最小误差熵, 配电变压器, 停电预测

Abstract:

In order to improve the speed and accuracy of distribution transformer outage prediction, a distribution transformer outage prediction method based on Logistic fast minimum error entropy algorithm was proposed. Aiming at the problem that the basic minimum entropy regression algorithm runs too slowly, a fast minimum error entropy algorithm was proposed, which can keep the same regression effect as the minimum entropy regression, and greatly reduce the running time of the algorithm. In view of the application of Logistic regression in outage prediction, a fast minimum error entropy regression algorithm based on logistic was proposed, and the weight of distribution transformer was selected. The overload duration, maximum active load rate, average active load rate, average three-phase unbalance degree and heavy three-phase unbalance degree were used as the characteristic variable data of distribution transformer outage prediction. A distribution transformer outage prediction model was established, and the effect was found to be better than Logistic regression in the comparative experiment.

Key words: Logistic fast minimum error entropy, distribution transformer, outage prediction

中图分类号: