PHYSICS-BASED DETECTION OF TELEMETRY ATTACKS IN PV-RICH DISTRIBUTION FEEDERS

Distribution networks that integrate photovoltaic (PV) plants are increasingly exposed to cyber-attacks that can falsify measurements and mislead voltage regulation. This thesis examines how false-data injections targeting PMU/inverter signals can distort voltage stability awareness and trigger unsafe control actions. We propose a simple, physics-guided detection layer that compares local voltage-current measurements with the electrical relationships allowed by the feeder. When data are manipulated, this physical consistency is broken, and the detector issues an alarm without requiring statistical training or prior knowledge of the attack. We also introduce a practical test to determine whether observed changes in PV power result from real weather and environmental variations or malicious manipulation, thereby reducing false alarms during normal fluctuations in irradiance. A system-level model is used to represent how intrusions enter through the communication layer and affect measurement streams. The method is assessed under various operating conditions and measurement noise levels. Results show fast and reliable attack detection while remaining quiet under genuine weather-driven changes. Overall, the work introduces a lightweight, real-time layer that enhances the trustworthiness of voltage regulation and stability monitoring in PV-rich distribution grids.