Breakthrough in Renewable Energy Forecasting: Deep Learning Model outperforms Conventional Techniques

Strategic Insights -

Research conducted at Delhi Technological University has resulted in a significant advancement in renewable energy forecasting, particularly in solar photovoltaic (SPV) power output. The study developed an innovative deep learning-based model that integrates the advantages of gray wolf optimization (GWO) and whale optimization algorithm (WOA) to enhance performance in real-time data collected from a New Delhi, India location. The proposed model exhibited enhanced performance for forecasting SPV power relative to conventional forecasting techniques, with significant improvements in forecasting accuracy.

Key Takeaways:

  • The deep learning-based model developed in the study integrates the advantages of GWO and WOA to optimize hyperparameters and improve forecasting accuracy.
  • The model was trained on historical meteorological data, encompassing irradiance, cloud cover, wind speed, wind direction, temperature, and SPV power, to enhance performance.
  • The optimized proposed method was compared with conventional deep learning models, and the models optimized with GWO and WOA showed significant improvements in forecasting accuracy, with the GWO-bi-directional-long short-term memory model obtaining the lowest root mean square error of 0.0154 and the highest R2 value of 0.9988 for 5-min interval data.
  • The modified variations decreased predicting errors by up to 45% compared to baseline models such as long short-term memory and bi-directional-long short-term memory models.
  • The research highlighted the potential of the proposed methodology for enhancing smart grid technologies and integrating renewable energy into smart grids.

Statistics:

  • Root mean square error (RMSE) of 0.0154 for the GWO-bi-directional-long short-term memory model.
  • R2 value of 0.9988 for the GWO-bi-directional-long short-term memory model.
  • Up to 45% decrease in predicting errors compared to baseline models.
  • The model was trained on historical meteorological data collected from New Delhi, India.

Sources:

  • A Novel Hybrid Gwo-bi-lstm-based Metaheuristic Framework for Short-term Solar Photovoltaic Power Forecasting. Journal of Renewable and Sustainable Energy, 2025;17(4).
  • NewsRx. Investigators at Delhi Technological University Detail Findings in Renewable Energy (A Novel Hybrid Gwo-bi-lstm-based Metaheuristic Framework for Short-term Solar Photovoltaic Power Forecasting). Ecology, Environment & Conservation. August 8, 2025; p 280.