A novel multi-task learning model based on Transformer-LSTM for wind power forecasting

The integration of multi-task learning into multi-step deterministic and probabilistic prediction frameworks plays a pivotal role in augmenting the accuracy of wind power forecasts and mitigating associated operational uncertainties. Current wind power forecasting research usually focuses on single-...

Full description

Saved in:
Bibliographic Details
Main Authors: Rongquan Zhang, Siqi Bu, Yuxia Zheng, Gangqiang Li, Xiupeng Wan, Qiangqiang Zeng, Min Zhou
Format: Article
Language:English
Published: Elsevier 2025-08-01
Series:International Journal of Electrical Power & Energy Systems
Subjects:
Online Access:http://www.sciencedirect.com/science/article/pii/S0142061525002832
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849689252018782208
author Rongquan Zhang
Siqi Bu
Yuxia Zheng
Gangqiang Li
Xiupeng Wan
Qiangqiang Zeng
Min Zhou
author_facet Rongquan Zhang
Siqi Bu
Yuxia Zheng
Gangqiang Li
Xiupeng Wan
Qiangqiang Zeng
Min Zhou
author_sort Rongquan Zhang
collection DOAJ
description The integration of multi-task learning into multi-step deterministic and probabilistic prediction frameworks plays a pivotal role in augmenting the accuracy of wind power forecasts and mitigating associated operational uncertainties. Current wind power forecasting research usually focuses on single-task deterministic forecasting, but there is less research on multi-task learning for multi-step deterministic and probabilistic forecasting. To achieve this, a novel multi-task learning model based on Transformer-Long short-term memory (LSTM) for multi-step deterministic and probabilistic wind power forecasting. First, a novel deep learning hybrid approach (DLH) based on the dilated causal convolutional network, Transformer, LSTM, and L2 regularization is proposed to extract multi-dimensional and complex nonlinear features of wind power time series. Then, the DLH is introduced into the task-sharing layer of the multi-task learning model for multi-step deterministic prediction. In addition, a probabilistic forecasting model that integrates the proposed multi-task learning model and quantile regression is formulated to describe the forecast uncertainty of wind power. Finally, to improve the prediction accurateness, a new heuristic algorithm, namely hybrid Cauchy mutation-based Elk herd and sand cat swarm optimization, is proposed to optimize the model hyperparameters. The proposed deterministic and probabilistic forecasting models have been applied to operational datasets from a northwest China wind farm. Compared to 23 state-of-the-art deterministic models, the proposed model reduces the mean absolute error by a minimum of 0.3174 and a maximum of 9.190, with an average reduction of 2.278. Experimental outcomes also substantiate that the proposed probabilistic model exhibits superior interval sharpness when juxtaposed against six advanced benchmarks.
format Article
id doaj-art-ce8df1cc4a2c4c3b83e7b32b9fcd39e5
institution DOAJ
issn 0142-0615
language English
publishDate 2025-08-01
publisher Elsevier
record_format Article
series International Journal of Electrical Power & Energy Systems
spelling doaj-art-ce8df1cc4a2c4c3b83e7b32b9fcd39e52025-08-20T03:21:42ZengElsevierInternational Journal of Electrical Power & Energy Systems0142-06152025-08-0116911073210.1016/j.ijepes.2025.110732A novel multi-task learning model based on Transformer-LSTM for wind power forecastingRongquan Zhang0Siqi Bu1Yuxia Zheng2Gangqiang Li3Xiupeng Wan4Qiangqiang Zeng5Min Zhou6College of Transportation, Nanchang JiaoTong Institute, Nanchang, 330044, China; Department of Electrical and Electronic Engineering, Hong Kong Polytechnic University, 999077, Hong KongDepartment of Electrical and Electronic Engineering, Hong Kong Polytechnic University, 999077, Hong KongCollege of Transportation, Nanchang JiaoTong Institute, Nanchang, 330044, ChinaDepartment of Electrical and Electronic Engineering, Hong Kong Polytechnic University, 999077, Hong KongSino-German Colledge of Intelligent Manufacturing, Shenzhen Technology University, Shenzhen, 518118, ChinaDepartment of Electrical and Electronic Engineering, Hong Kong Polytechnic University, 999077, Hong KongSino-German Colledge of Intelligent Manufacturing, Shenzhen Technology University, Shenzhen, 518118, China; Corresponding author.The integration of multi-task learning into multi-step deterministic and probabilistic prediction frameworks plays a pivotal role in augmenting the accuracy of wind power forecasts and mitigating associated operational uncertainties. Current wind power forecasting research usually focuses on single-task deterministic forecasting, but there is less research on multi-task learning for multi-step deterministic and probabilistic forecasting. To achieve this, a novel multi-task learning model based on Transformer-Long short-term memory (LSTM) for multi-step deterministic and probabilistic wind power forecasting. First, a novel deep learning hybrid approach (DLH) based on the dilated causal convolutional network, Transformer, LSTM, and L2 regularization is proposed to extract multi-dimensional and complex nonlinear features of wind power time series. Then, the DLH is introduced into the task-sharing layer of the multi-task learning model for multi-step deterministic prediction. In addition, a probabilistic forecasting model that integrates the proposed multi-task learning model and quantile regression is formulated to describe the forecast uncertainty of wind power. Finally, to improve the prediction accurateness, a new heuristic algorithm, namely hybrid Cauchy mutation-based Elk herd and sand cat swarm optimization, is proposed to optimize the model hyperparameters. The proposed deterministic and probabilistic forecasting models have been applied to operational datasets from a northwest China wind farm. Compared to 23 state-of-the-art deterministic models, the proposed model reduces the mean absolute error by a minimum of 0.3174 and a maximum of 9.190, with an average reduction of 2.278. Experimental outcomes also substantiate that the proposed probabilistic model exhibits superior interval sharpness when juxtaposed against six advanced benchmarks.http://www.sciencedirect.com/science/article/pii/S0142061525002832Wind powerProbabilistic predictionMulti-task learningDilated causal convolutionalTransformer-LSTMElk herd optimizer
spellingShingle Rongquan Zhang
Siqi Bu
Yuxia Zheng
Gangqiang Li
Xiupeng Wan
Qiangqiang Zeng
Min Zhou
A novel multi-task learning model based on Transformer-LSTM for wind power forecasting
International Journal of Electrical Power & Energy Systems
Wind power
Probabilistic prediction
Multi-task learning
Dilated causal convolutional
Transformer-LSTM
Elk herd optimizer
title A novel multi-task learning model based on Transformer-LSTM for wind power forecasting
title_full A novel multi-task learning model based on Transformer-LSTM for wind power forecasting
title_fullStr A novel multi-task learning model based on Transformer-LSTM for wind power forecasting
title_full_unstemmed A novel multi-task learning model based on Transformer-LSTM for wind power forecasting
title_short A novel multi-task learning model based on Transformer-LSTM for wind power forecasting
title_sort novel multi task learning model based on transformer lstm for wind power forecasting
topic Wind power
Probabilistic prediction
Multi-task learning
Dilated causal convolutional
Transformer-LSTM
Elk herd optimizer
url http://www.sciencedirect.com/science/article/pii/S0142061525002832
work_keys_str_mv AT rongquanzhang anovelmultitasklearningmodelbasedontransformerlstmforwindpowerforecasting
AT siqibu anovelmultitasklearningmodelbasedontransformerlstmforwindpowerforecasting
AT yuxiazheng anovelmultitasklearningmodelbasedontransformerlstmforwindpowerforecasting
AT gangqiangli anovelmultitasklearningmodelbasedontransformerlstmforwindpowerforecasting
AT xiupengwan anovelmultitasklearningmodelbasedontransformerlstmforwindpowerforecasting
AT qiangqiangzeng anovelmultitasklearningmodelbasedontransformerlstmforwindpowerforecasting
AT minzhou anovelmultitasklearningmodelbasedontransformerlstmforwindpowerforecasting
AT rongquanzhang novelmultitasklearningmodelbasedontransformerlstmforwindpowerforecasting
AT siqibu novelmultitasklearningmodelbasedontransformerlstmforwindpowerforecasting
AT yuxiazheng novelmultitasklearningmodelbasedontransformerlstmforwindpowerforecasting
AT gangqiangli novelmultitasklearningmodelbasedontransformerlstmforwindpowerforecasting
AT xiupengwan novelmultitasklearningmodelbasedontransformerlstmforwindpowerforecasting
AT qiangqiangzeng novelmultitasklearningmodelbasedontransformerlstmforwindpowerforecasting
AT minzhou novelmultitasklearningmodelbasedontransformerlstmforwindpowerforecasting