Identification of photovoltaic module ageing from symmetrically and asymmetrically cut current–voltage curves

Condition monitoring of photovoltaic (PV) power systems is necessary for their optimal power production: ageing of PV modules must be detected in time. Measured current–voltage curves subjected to single-diode model parameter identification and series resistance -based ageing analysis could provide a low-cost solution for online condition monitoring. However, in practice only small portions of the current–voltage curves around the maximum power point (MPP) can be measured for the parameter identification. This reduces the accuracy of series resistance identification, which complicates the ageing diagnosis. The exact effect of measurement range restrictions on the ageing detection is yet unclear constituting a clear research gap. Indeed, studies where different measurement regions are asymmetrically set on either side of the MPP are lacking, although they would have direct practical value in minimizing the power losses stemming from the measurement. This article addresses this issue for the first time by systematically investigating current–voltage curves limited asymmetrically around the MPP. Both power and voltage-based principles in limiting the measurement region are utilized. It is proven that the ageing-induced increase of series resistance of PV modules can be accurately identified from current–voltage curves measured close to the MPP. Also, the accuracy of the irradiance identification very near the MPP can be clearly improved by reducing the number of identified parameters. The results provide clear guidance for selection of the suitable measurement region for diagnostics. Accordingly, they serve practical condition monitoring of PV power systems. • Ageing of photovoltaic modules was quantified from measured current–voltage curves. • Series resistance values obtained by fitting the single-diode model were examined. • Partial current–voltage curves measured close to the maximum power point were used. • Ageing was excellently detected from curves measured near the maximum power point. • The used identification procedure can be used for online condition monitoring in practice.