A novel hybrid GRU-XGBoost model for day-ahead photovoltaic generation forecasting in microgrids

The global transition toward renewable energy has made photovoltaic (PV) technology a key component in sustainable microgrid operations. However, the inherent variability of PV generation (PVGen), influenced by changing weather conditions, presents significant challenges for efficient energy management and grid integration. Accurate day-ahead PVGen forecasting is crucia8l for optimizing energy storage, load balancing, and minimizing reliance on conventional energy sources. This study introduces a novel hybrid forecasting model, combining Gated Recurrent Units (GRU) and Extreme Gradient Boosting (XGBoost), to address the challenges of PVGen forecasting in microgrids. The proposed model employs a two-stage approach: GRU captures temporal patterns in historical weather and PV data, while XGBoost refines these forecasts by modeling non-linear relationships. Particle Swarm Optimization (PSO) is utilized to fine-tune hyperparameters in both stages, ensuring optimal performance. Evaluations under various weather conditions, using RMSE, MAE, nRMSE, and R² metrics, demonstrate the model’s effectiveness. The proposed model achieved a notably low nRMSE of 4.17 %, outperforming existing models, such as GRU-Informer-SVR (33.9 % improvement) and LSTM-Informer (54.5 % improvement). Under sunny conditions, the model achieved an nRMSE of 2.67 % (38.3 % lower than standalone GRU) and an R² of 0.9956, while on overcast days, it maintained robust performance with an nRMSE of 5.12 % (63.9 % improvement over GRU) and R² of 0.9670. The proposed GRU–XGBoost hybrid model significantly advances day-ahead PV generation forecasting, providing a reliable, adaptable, and efficient solution for microgrid applications. Its strong performance across diverse weather conditions underscores its potential as solution for optimizing PVGen management in real-world energy systems.