An online intrusion detection system for photovoltaic generators through physics-based neural networks

• Proposes an online intrusion detection system for photovoltaic (PV) generators. • Integrates physics-informed neural networks (PINN) with LSTM architectures. • Detects cyberattacks by analyzing real-time PV physical and electrical behavior. • Achieves higher accuracy and fewer false alarms than traditional IDS methods. • Provides a scalable approach to enhance the cybersecurity of distributed PV systems. The rapid proliferation of photovoltaic (PV) systems in modern power grids has introduced new cybersecurity vulnerabilities. This paper presents a novel approach to cybersecurity monitoring of PV systems based on real-time analysis of their physical and electrical behavior. We propose an anomaly detection algorithm that combines Long Short-Term Memory (LSTM) networks with Physics-Informed Neural Networks (PINNs) to identify potential cyberattacks. Experimental results demonstrate that the proposed method outperforms conventional techniques in identifying anomalies, offering faster and more reliable detection capabilities.