Accurate day-ahead forecasting of photovoltaic (PV) generation is essential for the reliable operation of smart grids and energy management systems (EMS). This paper proposes a lightweight, two-stage neural network model for 24-hour ahead PV prediction at 15-minute resolution. The framework integrates a daylight classifier, implemented using a bagged ensemble of decision trees, with a Bayesian-regularized feedforward neural network (FFNN) trained exclusively on daylight samples. The model is trained and validated using real-world PV and environ- mental data from the LAMBDA Microgrid Laboratory in Rome, Italy. Experimental results demonstrate an RMSE of 0.0755 and a daylight MAPE of 19.94%, along with 100% accuracy in detect- ing zero-generation periods. The architecture ensures physically consistent forecasts and supports real-time deployment in EMS. Compared to standard baselines such as SVR and single-stage FFNN, the proposed model achieves superior daylight accuracy and robustness. It offers a practical foundation for real-time EMS deployment and future extensions involving probabilistic forecasting and hybrid deep learning architectures.
A Two-Stage Neural Network for Day-Ahead PV Forecasting Using Environmental Data from the LAMBDA Microgrid in Rome / Nat, Aslam; Of Computer, Dept.; Control, ; Management Engineering, And; University Of Rome, Sapienza; Of Astronautical, Italy Email. Aslam. Nat@uniroma1. It Riccardo Loggia Dept.; Electrical, ; Energy Engineering (DIAEE), And; University Of Rome, Sapienza; Of Astronautical, Italy Emai :riccardo. Loggia@uniroma1. It Cristina Moscatiello Dept.; Electrical, ; Energy Engineering (DIAEE), And; University Of Rome, Sapienza; Italy, ; Of Astronautical, Cristina. Moscatiello@uniroma1. It Luigi Martirano Dept.; Electrical, ; Energy Engineering (DIAEE), And; University Of Rome, Sapienza; It, Italy Luigi. Martirano@uniroma1.. - (2025). (Intervento presentato al convegno IEEE PES GTD Asia 2025 – Power and Energy Conference and Exposition tenutosi a Bangkok,Thailand).
A Two-Stage Neural Network for Day-Ahead PV Forecasting Using Environmental Data from the LAMBDA Microgrid in Rome
Aslam Nat
;
2025
Abstract
Accurate day-ahead forecasting of photovoltaic (PV) generation is essential for the reliable operation of smart grids and energy management systems (EMS). This paper proposes a lightweight, two-stage neural network model for 24-hour ahead PV prediction at 15-minute resolution. The framework integrates a daylight classifier, implemented using a bagged ensemble of decision trees, with a Bayesian-regularized feedforward neural network (FFNN) trained exclusively on daylight samples. The model is trained and validated using real-world PV and environ- mental data from the LAMBDA Microgrid Laboratory in Rome, Italy. Experimental results demonstrate an RMSE of 0.0755 and a daylight MAPE of 19.94%, along with 100% accuracy in detect- ing zero-generation periods. The architecture ensures physically consistent forecasts and supports real-time deployment in EMS. Compared to standard baselines such as SVR and single-stage FFNN, the proposed model achieves superior daylight accuracy and robustness. It offers a practical foundation for real-time EMS deployment and future extensions involving probabilistic forecasting and hybrid deep learning architectures.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
