Noise-Scaled Euclidean Distance: A Metric for Maximum Likelihood Estimation of the PV Model Parameters

This article revisits the objective function (or metric) used in extraction of photovoltaic (PV) model parameters. A theoretical investigation shows that widely current distance (CD) metric does not yield maximum likelihood estimates (MLE) parameters when there is noise both voltage and samples. It demonstrates Euclidean (ED) should be instead, powers are equal. For general case, a new noise-scaled (NSED) proposed as weighted variation ED, which shown to fetch MLE at any conditions. requires ratio (i.e., two variances) an additional input, can estimated by estimation (NE) method introduced this study. One application employ NSED regression follow-up step existing parameter methods toward fine-tuning their outputs. Results on synthetic experimental data show so-called “add-on” improves accuracy five such validate merits metric.