A novel stochastic maximum power point tracking control for off-grid standalone photovoltaic systems with unpredictable load demand

Abstract

This paper shows a novel stochastic maximum power point tracking (MPPT) control for photovoltaic (PV) generators. The PV generator is taken disconnected from the grid, i.e., it considers a direct-current (DC) microgrid supplying varying loads. In this setting, we examine how random, varying loads affect the stability and efficiency of the PV generator. The load changes its value according to a Markov chain, the main assumption of this paper. PV generators are complex nonlinear devices, a challenge from the modeling viewpoint. An alternative becomes converting the nonlinear PV generator model into a Takagi-Sugeno (T-S) fuzzy model. The resulting stochastic T-S fuzzy system has its stability characterized by a condition written in linear matrix inequalities (LMIs). The usefulness of our approach is illustrated by simulating the PV generator model fed with real-time weather data collected in Brazil. The corresponding data indicated that the MPPT efficiency was greater than 99.5%, thereby outperforming other methods from the literature. The corresponding data confirm that the DC-DC converter circuit was able to track maximum power from the PV generator against random load variations. Comparisons with other methods indicate the potential of our approach for PV generators.