基于实时监测数据的海上风电分钟级超短期多步功率预测模型

黄秋璇; 唐靖泊; 邢志鹏; 刘明波

引用本文:	黄秋璇,唐靖泊,邢志鹏,等.基于实时监测数据的海上风电分钟级超短期多步功率预测模型[J].电力系统保护与控制,2026,54(09):151-162.
	HUANG Qiuxuan,TANG Jingbo,XING Zhipeng,et al.Minute-scale ultra-short-term multi-step power forecasting model for offshore wind farm based on real-time monitoring data[J].Power System Protection and Control,2026,54(09):151-162

【打印本页】【下载PDF全文】【查看/发表评论】【EndNote】【RefMan】【BibTex】

←前一篇|后一篇→

过刊浏览高级检索

本文已被：浏览 201次下载 17次	码上扫一扫！
分享到：微信更多字体:加大+\|默认\|缩小-
基于实时监测数据的海上风电分钟级超短期多步功率预测模型
黄秋璇^1,2，唐靖泊^1,2，邢志鹏^1,2，刘明波^1,2
1. 华南理工大学电力学院，广东广州 510640；2. 华南理工大学广东省绿色能源技术重点实验室，广东广州 510640

摘要:

为给系统调度中心修正日内发电计划偏差提供高精度风电预测数据，提出一种融合双向门控循环单元 (bidirectional gated recurrent unit, BiGRU)和多头注意力机制的分钟级超短期编码器-解码器预测模型。首先，采用改进的动态时间弯曲算法量化各风机之间的相似度，从而实现对多台风机的聚类。然后，采用改进的自适应噪声完全集成经验模态分解算法将各个机群的历史功率序列分解为多个候选子序列，再结合皮尔逊相关系数筛选出与原始功率序列相关性高的子序列。最后，搭建编码器-解码器预测网络，对各个机群的功率子序列进行独立预测，并将这些功率子序列预测结果叠加获得风电场总功率的预测值。利用某实际海上风电场数据进行验证，结果表明所提方法较以BiGRU-CNN为代表的4种典型模型，在未来1—4 h超短期预测场景下表现更佳，能有效抑制风电不确定性的干扰，提升了超短期预测精度。

关键词: 海上风电场分钟级超短期风电功率预测编码器 - 解码器模型改进动态时间弯曲算法自适应噪声完全集合经验模态分解

DOI：10.19783/j.cnki.pspc.251264

分类号:

基金项目:广东省省级重点基金项目资助(2024B1515250007)

Minute-scale ultra-short-term multi-step power forecasting model for offshore wind farm based on real-time monitoring data

HUANG Qiuxuan^1,2, TANG Jingbo^1,2, XING Zhipeng^1,2, LIU Mingbo^1,2

1. School of Electric Power Engineering, South China University of Technology, Guangzhou 510640, China; 2. Guangdong Key Laboratory of Clean Energy Technology, South China University of Technology, Guangzhou 510640, China

Abstract:

To provide highly accurate wind power forecasting data for system dispatch centers to correct intraday generation schedule deviations, a minute-scale ultra-short-term encoder-decoder forecasting model integrating bidirectional gated recurrent units (BiGRU) and multi-head attention mechanism is proposed. First, improved dynamic time warping algorithm is used to quantify the similarity and cluster wind turbines. Next, improved complete ensemble empirical mode decomposition with adaptive noise is adopted to decompose historical power series. Pearson correlation coefficients are used to select highly correlated subsequences. Finally, an encoder-decoder network is built to predict each subsequence and aggregate to obtain total wind farm power. Validation shows the proposed method performs better in 1-4 hours ahead forecasting.

Key words: offshore wind farm minute-scale ultra-short-term wind power forecasting encoder-decoder model improved dynamic time warping complete ensemble empirical mode decomposition with adaptive noise

X关闭