Hybrid CNN-LSTM Detection of Cyber-Attacks on MPPT Controllers in PV Systems

Securing Maximum Power Point Tracking (MPPT) controllers in grid-connected photovoltaic (PV) systems is crucial for reliable and efficient energy generation. False Data Injection Attacks (FDIAs) exploit vulnerabilities in sensor data integrity, causing sub-optimal operation of MPPT controllers, leading to power losses and potential system instability. This paper presents a hybrid detection framework combining Long Short-Term Memory (LSTM) and Convolutional Neural Networks (CNNs) architectures are employed to reliably identify FDIAs in real time. The CNN extracts spatial features while the LSTM captures temporal dependencies, enhancing detection accuracy. The framework is validated using a custom-developed cyber-physical testbed that integrates HYPERSIM for detailed electromagnetic transient power system simulation and EXata for high-fidelity communication network emulation. The testbed models a realistic 2 MW grid-connected PV system on an OPAL-RT real-time simulation platform, employing advanced co-simulation techniques to evaluate the impact of FDIAs comprehensively. Experimental results demonstrate a detection accuracy of 99.93%, Showcasing the efficiency and applicability of the proposed approach in strengthening the cybersecurity of PV systems.