• MHFTNet introduces breakthrough renewable energy forecasting architecture combining solar and wind power prediction. • MHFTNet utilizes advanced temporal feature extraction mechanisms. • Proposed model achieves superior performance with RMSE of 0. 5627 and R 2 of 0. 9652 in day-ahead forecasting. • MHFTNet outperforms competing methods by 15% average accuracy with MAPE ranging from 0. 9% to 4. 7% • Generates 250–450 k annual revenue through optimized trading strategies and market timing. Energy security, sustainability, and management are extremely crucial for the efficient operation of a large-scale renewable energy mix in national grids. For the first time, the economic value of accurate forecasting has become increasingly critical for market participants. The stochastic nature of natural phenomena impacts wind and solar energy generation severely. To effectively operate hybrid grids, it is extremely important to forecast available wind and solar energy with high precision. MHFTNet represents a breakthrough advancement in renewable energy forecasting, introducing an innovative architecture designed explicitly for integrated solar and wind power prediction. This enhanced temporal feature extraction framework leverages advanced deep learning techniques to optimize short and medium-term power forecasts, addressing critical challenges in renewable energy integration. Key features include advanced temporal feature extraction mechanisms for improved forecast accuracy, an integrated processing of solar and wind energy patterns for enhanced multi-feature extraction and multi-horizon forecasting capabilities. Enhanced prediction models are optimized for renewable energy integration. The architecture demonstrates superior performance in handling complex temporal dependencies and multi-modal patterns characteristic of solar and wind energy generation. Numerically, MHFTNet demonstrates clear superiority over competing techniques. In day-ahead forecasting, it achieves an RMSE of 0. 5627 and an R 2 of 0. 9652. The proposed MHFTNet model demonstrates superior forecasting accuracy across multiple time horizons (t + 1 to t + 1344 h) with MAPE values ranging from 0. 9% to 4. 7%, consistently outperforming baseline methods, including AE-CNN, GRNN, LSTM, and KNN by 15% on average in accuracy metrics. Integration with real energy market data reveals that MHFTNet’s enhanced forecasting accuracy translates directly to substantial economic benefits, generating 250–450 k in additional annual revenue through optimized trading strategies, including opportunistic spot trading, portfolio optimization, and predictive market timing. The economic analysis establishes clear performance thresholds where forecast accuracy below 15% MAPE enables value-maximizing strategies with ROI ranging from 17. 8% to 19. 5%, demonstrating that accurate wind power forecasting serves as a fundamental economic enabling in modern electricity markets rather than merely an operational tool. Furthermore, market integration ensures these strategies align with real-time market dynamics, enabling proactive adjustments to trading positions.
Mansoor et al. (Sun,) studied this question.