A Rule-Based Fault Detection Method for Photovoltaic Systems Under Persistent Open Circuit Voltage and Low-Current Conditions

Abstract

Real-time Fault Detection and Diagnosis (FDD) is vital to guarantee the efficiency and reliability of photovoltaic (PV) systems despite environmental degradation. Some fault circumstances, however, such as surface occlusion or sensing-loop degradation, exhibit a unique diagnostic challenge in that the array continues to exhibit a completely saturated open-circuit voltage (Voc) despite a total collapse in short-circuit current (Isc) production. This research provide the first step of a low-cost sensor less edge-computing system for the isolation of these specific anomalous situations. The system uses an ESP32 microprocessor in combination with a time-multiplexed relay switching circuit to dynamically separate Voc and Isc in real time, without the need for external solar irradiance sensors. The solar irradiance (G) at the edge layer is directly estimated using a mathematical equation. The proposed methodology was validated on an outdoor experimental setup consisting of a 50W monocrystalline PV module. Empirical data were extracted using MATLAB to identify long-term baseline patterns in normal and abnormal operating windows. Preliminary results show that the sensor less architecture can separate several multi-day fault signals, capturing a full collapse in current and estimated irradiance while Voc stabilizes at its typical daytime potential of 21.2 V. This baseline profiling provides the required empirical basis for the next phase of the diagnostic system, which will integrate advanced time-frequency domain transformations (FFT/DWT).