A novel sensorless support vector regression based multi-stage algorithm to track the maximum power point for photovoltaic systems

This paper proposes a new approach for maximum power-point tracking (MPPT) process in photovoltaic (PV) systems. Based on the theory of support vector regression (SVR), a multi-stage algorithm (MSA) is proposed for MPPT to estimate the temperature and solar irradiation without a need to measure them. The only needed measurements for the proposed MSA are the output voltage and current of the PV panel. The MSA consists of three stages: The first stage estimates the initial values of temperature and irradiation; the second stage instantaneously estimates the irradiation assuming that the temperature is constant within a one-hour time span; and the third stage updates the estimated temperature once every one hour. The proposed method is robust, not only to changes in solar irradiation and load, but also to variations in temperature. Moreover, using fewer sensors improves the reliability of the system. The effectiveness of the proposed method is demonstrated through simulation studies conducted in the PSCAD/EMTDC and Matlab software environment.