Iterated unscented Kalman filter-based maximum power point tracking for photovoltaic applications

Abstract

One of main challenges in harvesting power from a PV source is maximum power point tracking (MPPT). This is due to the nonlinear behaviour and characteristics of PV arrays. Conventional MPPT techniques usually utilize a hill climbing process which requires partial/full scan of the array power-voltage (P-V) curve resulting in high power fluctuation during peak searching. Dynamic estimation techniques, such as the Kalman filter, benefit from their ability to estimate non-measurable signals with rapid convergence. In this paper, a MPPT technique based on the Iterated Unscented Kalman Filter (IUKF) is presented. The proposed technique achieves: (i) satisfactory MPPT for PV arrays working under varying environmental conditions, (ii) PV array modelling with full parameter estimation including temperature and insolation level, and (iii) full estimation of the working P-V curve for the PV array. Only six measurement points are required for MPPT, modelling, and curve estimation; hence no full scan for the P-V curve is needed. This paper presents the system mathematical model and simulations. Moreover, an experimental setup is implemented illustrating practical results at various insolation levels and temperatures to validate the proposed technique.