Direct MPP tracking of PV by estimating the Virtual Optimal Resistance

The challenging issue of designing an enhanced, stable and robust method capable to track the maximum power point (MPP) of a PV is presented. In the present work, the notion of the virtual optimal resistance (VOR) is firstly introduced while a merit analysis of its importance on the current and voltage characteristic of a PV is indicated. Secondly, a neural network (NN) based implementation is proposed with its swift and robust response to effectively provide a novel and successful methodology of the MPP tracking. The advantage of the method is that it drives the PV power production to its MPP without creating further ripples or oscillations on the output signal. More specifically, we exploit the fact that the MPP is a unique point on the PV current-voltage (I-V) characteristic and that VOR results from the intersection of the I-V curve with the line connecting the origin with the MPP. Then, in order to utilize this new entity on determining the MPP, a reliable and fast, on-line estimation of the VOR is required. This task is realized by the proposed NN scheme. The NN structure and the training procedure are precisely explained in the paper, whereas the very good performance of the methodology is examined. Different simulation studies are conducted and the results are compared with the most popular perturb and observe (P&O) technique to indicate the improvements achieved.