Sistema di monitoraggio per la produzione di energia rinnovabile e sistema acquisizione dati

This thesis presents a comprehensive framework for the monitoring, performance analysis, and forecasting of renewable energy systems, integrating Artificial Intelligence (AI), Machine Learning (ML), Deep Learning (DL), digital technologies, and multi-source meteorological data. The research is structured around eight milestones, encompassing the development of a modular data acquisition and monitoring platform, the assessment of climatic anomalies on energy production, sensitivity analysis of ML hyperparameters, and the deployment of advanced forecasting methodologies for short- and medium-term Photovoltaic (PV) energy generation and consumption prediction. The study demonstrates the feasibility of combining real-time data from inverters, numerical weather models (NWM), satellite-derived irradiance estimates, and historical PV production to enhance system efficiency, fault detection, and predictive energy management. ML models were rigorously evaluated and optimized, showing significant improvements in forecasting accuracy through hyperparameter tuning and feature selection. Hybrid AI approaches, such as Random Forest - Long Short-Term Memory (RF-LSTM) and Sequential Support Vector Regression - Multi-Layer Perceptron (Seq-SVR-MLP), further improved performance for hourly and multi-scale energy forecasts, supporting smart household and building-level energy management. Findings highlight the impact of environmental factors, topographical variations, and climatic anomalies on PV performance, providing actionable insights for system design, operation, and long-term planning. The thesis also identifies avenues for future research, including the integration of high-resolution satellite and Internet of Things (IoT) data, the development of physics-informed hybrid models, the implementation of digital twin ecosystems, and the evaluation of AI-based PV management frameworks in diverse geographic and climatic contexts. Overall, this work contributes to the advancement of data-driven, resilient, and scalable solutions for renewable energy management, supporting the transition toward sustainable, climate-resilient energy systems.