Maximum power point tracking algorithm of PV system based on irradiance estimation and multi-Kernel extreme learning machine

This paper proposes a maximum power point tracking (MPPT) algorithm based on irradiance estimation and multi-kernel extreme learning machine (MKELM) to reduce investment costs and improve PV system efficiency. First, because irradiance sensors are relatively expensive, an irradiance estimation method based on the grey wolf optimization (GWO) algorithm was used to replace the sensors to estimate irradiance value. Next, a prediction model based on MKELM was used to model the PV system. By inputting temperature and irradiance, the prediction model can output a reference voltage of the maximum power point (MPP), allowing the system to operate at the MPP. Experimental results showed that the irradiance estimation method based on GWO can accurately estimate irradiance value in real time, and the MKELM-based prediction model is highly accurate. Through simulation experiments on the PV system, validity and advantages of the proposed method over the traditional MPPT algorithm are verified under different operating environments.