Photovoltaic Module Temperature Estimation and Sensor Malfunction Detection Algorithm Based on Kalman Filter

This article presents a photovoltaic module temperature estimation and sensor malfunction detection algorithm based on the Kalman filter. At a given instant, an initial estimate of the temperature is determined through a chosen thermal model, which in this research was a dynamic linear regression empirical approach, and compared with the temperature sensor's measurement using the Kalman filter. This comparison results in a gain, which is then used to determine the final and more accurate temperature estimation through a weighted average between the initial estimated value and the temperature sensor's measurement. The method detects a possible malfunction of the temperature sensor and classifies the defect as sensor off/not measuring or with noisy measurements when the differences between the values estimated by the model used and those measured by the instrumentation become sufficiently significant. The proposed methodology was tested against data recorded in a photovoltaic laboratory at the University of São Paulo, Brazil, and the results indicate that this solution can accurately assess the modules' temperature and detect temperature sensor malfunctioning, both on clear and cloudy days, on both temperature sensor shutdowns and noise-corrupted measurements situations.