Prediction and Estimation of Maximum Power of a PV String Under Partial Shading Conditions

Under partial shading conditions (PSCs), a photovoltaic (PV) string's power-voltage ($P-V$) curve shows multiple peaks, making global maximum power point (GMPP) estimation challenging. This article proposes a new approximate voltage expression to predict all the maximum power points (MPPs) in the$P-V$curve of a partially shaded PV string. The ratio of short circuit currents is used as a scaling factor to determine the operating voltage of each PV module and approximate the voltage at MPP. Then voltage boundaries around the predicted GMPP are defined, and finally, a new foci of ellipse (FoE) based estimation method is proposed to estimate the GMPP of PV string. The mathematical expressions of the ellipse are utilized to update the foci until the GMPP is achieved. A comparison with existing MPP estimation techniques indicates that the proposed method has the highest accuracy with less computational time. Further, the proposed method is evaluated under 5829 PSCs, showing high effectiveness with a minimal mean bias error in GMPP ∼(10-5). Finally, the accuracy of the proposed method is validated in an experimental setup using PV modules and rheostat. This method is useful for PV plant design and training machine learning algorithms for MPP tracking applications.