Characteristic curve diagnosis based on fuzzy classification for a reliable photovoltaic fault monitoring

Implementing automatic fault diagnostic tools will contribute to reducing the operations and maintenance costs and duration associated with PV systems. This paper presents a fuzzy diagnostic algorithm relying on the electrical parameters classification which values are extracted from experimental I-V curve measurements of crystalline modules. It is dedicated to the identification of common faults in a power plant, such as uniform dust, partial shade and potential induced degradation using a curve tracer at the system level. The fuzzy processing of the electrical parameters permits recognizing the shape of the characteristic by discerning different signatures of each fault. Also, a novel algorithm is presented to detect the deviation occurred in the measured I-V curve caused by the by-pass activation due to partial shade. It depends on the quadratic and cubic polynomial regression which concavities are highly sensitive to noisy data.

Automated summary

Implementing automatic fault diagnostic tools in PV systems can reduce operation and maintenance costs. This paper presents a fuzzy diagnostic algorithm based on electrical parameters classification, aimed at identifying common faults and detecting deviations in measured I-V curves caused by partial shade.