基于Transformer Encoder和LSTM的计及多变量的风电功率短期预测
Transformer Encoder-LSTM Based Short-Term Wind Power Forecasting Considering Multiple Variables
黄佳骏1
作者信息
- 1. 国网上海市电力公司市南供电公司,上海 200000
- 折叠
摘要
短期风力发电的准确可靠预测对电网的稳定运行具有重要意义.针对风力发电的不稳定性,本文提出了一种结合Transformer的 Encoder和LSTM的混合构架,考虑多种影响因素对风力发电这种复杂的时间序列数据进行建模和预测.实验结果表面,本文所提出的Transformer Encoder-LSTM模型在风力发电预测任务上取得了明显的性能提升,在平均绝对误差(MAE)、均方根误差(RMSE)和平均绝对百分比误差(MAPE)等评估指标上都优于单一的LSTM和GRU模型.
Abstract
Accurate and reliable short-term wind power prediction is crucial for the stable operation of the power grid.To address the instability of wind power generation,this paper proposes a hybrid architecture combining Transformer Encoder and LSTM,considering various influencing factors to model and predict this complex time series data.Experimental results indicate that the proposed Transformer Encoder-LSTM model achieves significant performance improvements in the wind power prediction task,outperforming single LSTM and GRU models in terms of Mean Absolute Error(MAE),Root Mean Square Error(RMSE),and Mean Absolute Percentage Error(MAPE).
关键词
Transformer/LSTM/风力发电预测/电力系统Key words
transformer/LSTM/wind power prediction/power system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