Short-term power load forecasting for combined heat and power using CNN-LSTM enhanced by attention mechanism

Anping Wan, Qing Chang, Khalil AL-Bukhaiti and Jiabo He

Energy, 2023, vol. 282, issue C

Abstract: This study proposes a new approach for short-term power load forecasting using a combination of convolutional neural networks (CNN), long short-term memory (LSTM), and attention mechanisms to address the issue of information loss due to excessively long input time series data. The objective is to enhance the accuracy of short-term power load prediction, which is crucial for efficient energy management. The study analyzes the relationship between the target load and the collected parameters, identifying the most influential factors using Pearson correlation coefficient analysis. A one-dimensional CNN layer is utilized to extract high-dimensional features from the input data, followed by an LSTM layer that captures temporal correlations within the historical sequences. Finally, an attention mechanism is introduced to optimize the weight of the LSTM output, enhance the influence of key information, and optimize the overall prediction model. The performance of the proposed model is evaluated using two benchmark models based on mean absolute percentage error (MAPE), root mean square error (RMSE), and mean absolute error (MAE) metrics. The results show that the CNN-LSTM-A model outperforms the traditional LSTM model regarding power load prediction accuracy for two thermal power units, with an improvement of 7.3% and 5.7%, respectively, indicating superior performance. Therefore, this study demonstrates the effectiveness of the proposed CNN-LSTM-A model for short-term power load forecasting, which has potential applications in the energy industry. In conclusion, the proposed approach can improve the accuracy of power load forecasting, leading to more efficient energy management and cost savings. Additionally, the study highlights the importance of incorporating attention mechanisms into traditional LSTM models for power load forecasting, as it helps to optimize the weight of the LSTM output and improve the accuracy of the predictions. The proposed CNN-LSTM-A model can be potentially useful for energy companies and policymakers in making informed decisions regarding energy production and consumption. Overall, this study provides a valuable contribution to power load forecasting, and the proposed approach could be extended to other areas of time-series forecasting in the future.

Keywords: Combined heat and power; Attention mechanism; Convolution neural network; Long-short term memory; Power prediction; CNN-LSTM model; Feature extraction (search for similar items in EconPapers)
Date: 2023
References: View references in EconPapers View complete reference list from CitEc
Citations: View citations in EconPapers (29)

Downloads: (external link)
http://www.sciencedirect.com/science/article/pii/S0360544223016687
Full text for ScienceDirect subscribers only

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:eee:energy:v:282:y:2023:i:c:s0360544223016687

DOI: 10.1016/j.energy.2023.128274

Access Statistics for this article

Energy is currently edited by Henrik Lund and Mark J. Kaiser

More articles in Energy from Elsevier
Bibliographic data for series maintained by Catherine Liu ().