An Improved Method to Predict the Position of Maximum Power Point During Partial Shading for PV Arrays

Abstract

For a photovoltaic (PV) system, it is widely assumed that the peaks of a partially shaded P-V curve are located at the multiples of 0.8V <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">oc</sub> , where V <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">oc</sub> is the open-circuit voltage of the PV module. However, this assumption - known as the of 0.8V <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">oc</sub> model for partial shading - is not necessarily true. If the same model is used to design the maximum power point tracker (MPPT), it is very likely that the algorithm will scan the wrong region of the P-V curve and, hence, an incorrect global peak detection. This paper attempts to prove the inadequacy of the 0.8V <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">oc</sub> model when the number of modules in the string increases. Further, the work suggests a simple relationship to predict the correct position of the peaks. It is found that for a string of 20 modules (V <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">oc</sub> of 400 V), the maximum deviation between the actual and computed peak is less than 3 V, while for the 0.8V <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">oc</sub> model, the deviation reaches up to 50 V. In addition, using the proposed method, the MPPT efficiency can be improved by approximately 2%.