基于注意力机制的Bi-LSTM光伏发电超短期功率预测方法研究与实施

doi:doi:10.19912/j.0254-0096.tynxb.2024-0564

English Version Reader LoginJournal Login

您当前的位置：

首页 >

文章列表页 >

基于注意力机制的Bi-LSTM光伏发电超短期功率预测方法研究与实施

更新时间：2025-10-22

- 基于注意力机制的Bi-LSTM光伏发电超短期功率预测方法研究与实施
- Vol. 46, Issue 9, Pages: 437-444(2025)
- 作者机构：
- 作者简介：
- 基金信息：
- DOI：doi:10.19912/j.0254-0096.tynxb.2024-0564
  CLC： TM615
- Published Online：14 October 2025，
  
  Published：2025
- 稿件说明：
移动端阅览
[J]. 2025, 46(9): 437-444.
DOI：

[J]. 2025, 46(9): 437-444. DOI： doi:10.19912/j.0254-0096.tynxb.2024-0564.

摘要

针对新能源光伏场站功率预测性能逐年下滑和考核费用居高不下的问题

提出一种融合双向长短期记忆网络（Bi-LSTM）和注意力机制的优化方法。首先

通过应用Bi-LSTM捕捉功率数据的时间依赖性

提升对时间序列变化的理解能力;其次

结合注意力机制使模型能聚焦于对预测结果影响最大的特征

进一步提高预测精度;最后

利用历史数据建模验证准确性。现场模型部署后表明

该模型相比传统单模型和集成学习模型

可更好地捕捉功率快速变化

场站超短期预测准确率可提高10%以上。

Abstract

Aiming at the problems of declining power prediction performance and high appraisal cost of photovoltaic （PV） power plants year by year

this paper proposes an optimization method integrating bidirectional long short-term memory network （Bi-LSTM） and attention mechanism. First

the time dependence of power data is captured by applying Bi-LSTM to improve the understanding of time series changes; second

the combined attention mechanism enables the model to focus on the features that have the greatest impact on the prediction results to further improve the prediction accuracy; and finally

the accuracy is verified by using field historical data modeling. The field model deployment shows that the model can better capture the rapid power changes compared with the traditional single model and integrated learning model

and the ultra-short-term prediction accuracy of the field station can be improved by more than 10%.

关键词

Keywords

references

万灿, 宋永华. 新能源电力系统概率预测理论与方法及其应用[J]. 电力系统自动化, 2021, 45(1): 2-16.

MARKOVICS D, MAYER M J.Comparison of machine learning methods for photovoltaic power forecasting based on numerical weather prediction[J]. Renewable and sustainable energy reviews, 2022, 161: 112364.

陈元峰, 马溪原, 程凯, 等. 基于气象特征量选取与SVM模型参数优化的新能源超短期功率预测[J]. 太阳能学报, 2023, 44(12): 568-576.

王东风, 刘婧, 黄宇, 等. 结合太阳辐射量计算与CNN-LSTM组合的光伏功率预测方法研究[J]. 太阳能学报, 2024, 45(2): 443-450.

林兴宇, 肖迎群, 张苏. 基于极限学习机分位回归的光伏出力区间预测方法[J]. 机械与电子, 2023, 41(6): 3-9, 14.

林帆, 张耀, 东琦, 等. 基于分位数插值和深度自回归网络的光伏出力概率预测[J]. 电力系统自动化, 2023, 47(9): 79-87.

王倩, 张智晟, 王帅, 等. 基于PSO-RNN的光伏发电功率预测研究[J]. 青岛大学学报(工程技术版), 2014, 29(4): 40-43.

蔡源, 吴浩, 唐丹. 光伏发电功率预测方法综述[J]. 四川电力技术, 2024, 47(2): 25-31.

YU M, NIU D X, WANG K K, et al.Short-term photovoltaic power point-interval forecasting based on double-layer decomposition and WOA-BiLSTM-Attention and considering weather classification[J]. Energy, 2023, 275: 127348.

杨春熙, 韩威, 高志球. 基于GRU神经网络的太阳辐照度短期预测[J]. 中国科技论文, 2020, 15(1): 8-14.

姜景芮, 苗田, 彭婧, 等. 基于深度学习算法的光伏发电功率预测方法研究[J]. 自动化应用, 2023(13): 102-104.

ZHANG Y G, PAN Z Y, WANG H, et al.Achieving wind power and photovoltaic power prediction: an intelligent prediction system based on a deep learning approach[J]. Energy, 2023, 283: 129005.

叶林, 裴铭, 路朋, 等. 基于天气分型的短期光伏功率组合预测方法[J]. 电力系统自动化, 2021, 45(1): 44-54.

FENG X, WEI Y J, PAN X L, et al.Academic emotion classification and recognition method for large-scale online learning environment-based on A-CNN and LSTM-ATT deep learning pipeline method[J]. International journal of environmental research and public health, 2020, 17(6): 1941.

PREMALATHA M, VISWANATHAN V, ČEPOVÁ L.Application of semantic analysis and LSTM-GRU in developing a personalized course recommendation system[J]. Applied sciences, 2022, 12(21): 10792.

顾国庆, 李晓辉. 基于箱线图异常检测的指数加权平滑预测模型[J]. 计算机与现代化, 2021(1): 28-33.

DHARMA RAJ T, KUMAR C, SUBRAMANIAM G, et al.A novel ROA optimized Bi-LSTM based MPPT controller for grid connected hybrid solar-wind system[J]. COMPEL-the international journal for computation and mathematics in electrical and electronic engineering, 2023, 42(2): 378-401.

Views

下载量

CSCD

Alert me when the article has been cited

提交

Tools

Publicity Resources

基于相似日理论和IPOA-ELM的短期光伏发电预测

基于VMD-TCN-GRU-AM的超短期风电预测

PROBABILITY PREDICTION OF WIND POWER BASED ON QR-NFGLSTM AND KERNEL DENSITY ESTIMATION

SHORT-TERM PHOTOVOLTAIC POWER FORECASTING BASED ON ATTENTION-GRU MODEL

SHORT-TERM PROBABILISTIC FORECASTING METHOD OF PHOTOVOLTAIC OUTPUT POWER BASED ON IMPROVED TEMPORAL CONVOLUTIONAL NETWORK

Related Author

孔令廉

王海云

黄晓芳

范竞敏

贺广林

王新刚

张奎

李佩怡

Related Institution

School of Electrical Engineering, Shenyang University of Technology

School of Electrical and Information Engineering, Jiangsu University

Jiangsu Zhen An Power Equipment Co., Ltd.

School of Computer Science and Technology, Shanghai University of Electric Power

School of Electrical Engineering, Xinjiang University

AI问答

⁰