A nonlinear predictor for the supervision of photovoltaic strings performances

In this paper, a nonlinear predictor of the electrical power produced by a PV string is proposed. The first phase of the approach is the training of the predictor, during which four characteristic parameters are determined. Such coefficients are representative of the string under study and define its electrical signature (identikit). Once trained the model, when new monitoring data are available, the mismatch between the forecasted and measured electrical power can be assumed as a reliable marker of the performances of the string, since the greater the mismatch, the worse the string efficiency. The analysis of the forecasting error, therefore, enables the detection of losses of energy production. In particular, a strength of the proposed approach is the possibility to distinguish the losses due to aging phenomena from the losses due to the dust or dirt accumulation. The method has been tested and validated for a real case study and the obtained results are presented in the paper.