Conditional Probability Approach for Fault Detection in Photovoltaic Energy Farms

Detection of electric faults in photovoltaic (PV) farms enhances a sustainable service continuity of farm energy generation. In this paper, a probabilisticfunction is introduced to detect the faults in the PV farms. The conditional probability functions are adopted to detect different fault conditions such as internalstring faults, string-to-string faults, and string-to-negative terminal faults. As thediodes are important to make the PV farms in-service safely during the faults,the distribution currents of these faults are evaluated with different concepts ofdiode consideration as well as without considering any diode installation. Thispart of the study enhances the diode utilization in the PV farms concerning theprotection point of view. The PV string currents are used as inputs to the conditional probability detection algorithms. However, the setting of the fault detectiontechnique is not portable for the other PV systems due to broad ranges of PV system ratings. To accordingly generalize the proposed fault detection algorithm, thePV string currents are first normalized to the total array current for universallyapplying the detection function at different PV string ratings. The limiting faultresistances are evaluated to show the sensitivity of the proposed fault detector.The results ensure the application of the proposed probabilistic detection functionfor PV farms.