article · Mathematics
Accurate solar energy production forecasting is critical for feeding renewable energy into power grids, particularly to support developing cities with high energy consumption. To address this challenge, a deep learning approach combines a Convolutional Neural Network, a Long Short-Term Memory network, and a Transformer architecture. The methodology incorporates clustering techniques to analyse correlations in the input data, alongside a self-organising map used to identify and select the most relevant historical features. Evaluation was conducted using the Fingrid open dataset. The resulting hybrid CNN-LSTM-Transformer model demonstrated superior forecasting accuracy when benchmarked against existing models and alternative configurations such as LSTM-CNN. By delivering reliable and accurate predictions of solar power output, this integrated forecasting technique offers a dependable mechanism to facilitate smoother, more efficient integration of solar power into wider electrical grids.
Integrating variable renewable energy into electrical networks is difficult without dependable forecasts. Improved predictions of solar energy production assist grid operators in balancing electricity supply and demand effectively. By combining advanced deep learning architectures with feature-selection techniques, this approach helps reduce uncertainty in power generation, thereby supporting the broader adoption of green energy in expanding cities and aiding environmental preservation.
This forecasting tool is targeted at grid operators and energy utilities that need to integrate solar power generation into electrical networks. The work represents applied research that has been tested on the Fingrid open dataset. Commercial deployment would require embedding the model into operational control software and validating it against live grid feeds to advance it from an evaluated prototype into a production-ready utility application.
AI-generated from the published abstract. Always read the original work before citing.
Green energy is very important for developing new cities with high energy consumption, in addition to helping environment preservation. Integrating solar energy into a grid is very challenging and requires precise forecasting of energy production. Recent advances in Artificial Intelligence have been very promising. Particularly, Deep Learning technologies have achieved great results in short-term time-series forecasting. Thus, it is very suitable to use these techniques for solar energy production forecasting. In this work, a combination of a Convolutional Neural Network (CNN), a Long Short-Term Memory (LSTM) network, and a Transformer was used for solar energy production forecasting. Besides, a clustering technique was applied for the correlation analysis of the input data. Relevant features in the historical data were selected using a self-organizing map. The hybrid CNN-LSTM-Transformer model was used for forecasting. The Fingrid open dataset was used for training and evaluating the proposed model. The experimental results demonstrated the efficiency of the proposed model in solar energy production forecasting. Compared to existing models and other combinations, such as LSTM-CNN, the proposed CNN-LSTM-Transformer model achieved the highest accuracy. The achieved results show that the proposed model can be used as a trusted forecasting technique that facilitates the integration of solar energy into grids.
This page summarises published work. The authoritative version sits with the publisher.
DOI: 10.3390/math11030676
Is something wrong with this record? Report it or request removal.
Discussion
Have you built on this work, tried to replicate it, or seen it applied in practice? Share what you know. Verified researchers and MARATTO™ domain experts can open a discussion, and any member can reply. Contributions are reviewed before they appear.
No discussion yet. Open the first thread.
New to MARATTO™? Create a free account.
