电器与能效管理技术 ›› 2026, Vol. 0 ›› Issue (3): 32-43.doi: 10.16628/j.cnki.2095-8188.2026.03.005

• 研究与分析 • 上一篇    下一篇

基于改进分位数回归的短期负荷预测及在线学习方法研究

储琳琳1, 宗明1, 张宇俊1, 易越1, 郑予容1, 韦宁2, 贾雅君2   

  1. 1 国网上海市电力公司市南供电公司, 上海 200030
    2 上海君世电气科技有限公司, 上海 200240
  • 收稿日期:2025-11-28 出版日期:2026-03-30 发布日期:2026-04-20
  • 作者简介:储琳琳(1978—),女,正高级工程师,主要从事电力系统规划设计研究工作。|宗 明(1969—),男,正高级工程师,主要从事电力系统规划设计研究工作。|张宇俊(1975—),女,高级经济师,主要从事电力系统规划设计研究工作。
  • 基金资助:
    国家电网有限公司科技项目(520933250002)

Research on Short-Term Load Forecasting and Online Learning Method Based on Improved Quantile Regression

CHU Linlin1, ZONG Ming1, ZHANG Yujun1, YI Yue1, ZHENG Yurong1, WEI Ning2, JIA Yajun2   

  1. 1 State Grid Shanghai Municipal Electric Power Company, Shinan Power Supply Company, Shanghai 200030, China
    2 Shanghai Junshi Electric Technology Co., Ltd., Shanghai 200240, China
  • Received:2025-11-28 Online:2026-03-30 Published:2026-04-20

摘要:

针对高比例可再生能源并网与需求侧弹性增强背景下,传统确定性负荷预测难以满足电力系统风险评估与决策优化需求的问题,开展短期负荷不确定性预测研究。构建改进分位数回归神经网络模型,采用倾斜分位数损失函数与区间后处理方法提升预测精准度与可靠性;结合负荷模式动态变化特征,提出基于弹性权重合并(EWC)的在线学习方法实现模型参数动态更新。实验结果表明,所提分位数预测方法在维持高覆盖率的同时,预测区间归一化平均宽度优于传统方法,且在线学习后覆盖率显著提升,验证了方法的有效性与适应性。

关键词: 短期负荷预测, 不确定性预测, 改进分位数回归, 在线学习

Abstract:

Against the backdrop of high penetration of renewable energy integration and enhanced demand-side flexibility,traditional deterministic load forecasting methods struggle to meet the requirements of risk assessment and decision optimization in power systems.Research on short-term load uncertainty forecasting is conducted.An improved quantile regression neural network model is constructed,adopting a tilted quantile loss function and interval post-processing method to enhance prediction accuracy and reliability.Considering the dynamic variation characteristics of load patterns,an online learning method based on elastic weight consolidation (EWC) is proposed to realize dynamic updates of model parameters.Experimental results show that the proposed quantile forecasting method outperforms traditional methods in terms of normalized average width of prediction intervals while maintaining high coverage,and the coverage is significantly improved after online learning,verifying the effectiveness and adaptability of the proposed method.

Key words: short-term load forecasting, uncertainty prediction, improved quantile regression, online learning

中图分类号: