Gradient-based optimization with ranking mechanisms for parameter identification of photovoltaic systems

Abstract

Deriving optimal photovoltaic (PV) models’ optimal parameters have tremendous significance in simulating, evaluating, and controlling the photovoltaic systems. Determining unknown parameters of these PV models is a multimodal, nonlinear, and complex optimization problem. Hence, developing a robust optimization model to achieve optimal parameters of the PV models effectively is essential. This paper proposes an enhanced metaphor-free gradient-based optimizer (EGBO) for extracting PV parameters quickly, precisely, and reliably. In the EGBO, a rank-based mechanism is employed to update its parameters efficiently. Also, the logistic map (LC) is implemented to better use the local escaping operator (LEO) in the original GBO algorithm. The proposed EGBO optimally identifies various parameters in the PV model, such as single diodes, double diodes, and PV modules. The relevant results indicate that compared with most advanced optimization methods, the EGBO algorithm is competitive in reliability, accuracy, and convergence speed. Moreover, the relevant results from the experimental data drawn from the manufacturer’s datasheet demonstrate that the developed approach can offer highly accurate solutions at various irradiances and temperatures. Consequently, the achieved results confirm that the novel approach can be presented as a utility tool for deriving optimal PV models’ optimal parameters, and it can be helpful in modeling PV systems.

Keywords

• Photovoltaic models
• Parameter identification
• Gradient-based optimizer
• Optimization
• Swarm intelligence