A data-driven photovoltaic string current mismatch fault diagnosis method based on I-V curve

Due to the complex operating environment, current mismatch faults often occur in photovoltaic (PV) strings in the actual PV plants, which seriously reduces the power of the entire string. This paper investigates and collects the data of mismatched PV strings in an actual PV plant, and further the fault characteristics of mismatched PV strings are extracted through the I-V curve. A data-driven online diagnosis method for current mismatch faults based on string I-V data is proposed. The method uses the I-V data to construct a detection line, and the inflection point and step features on the I-V curve caused by current mismatch can be detected simply and quickly through the parallel movement of the detection line and the positional relationship between the line and the I-V data. The experimental verifications from the actual PV plant are demonstrated to show the effectiveness of the proposed method for PV string current mismatch fault.