A Fractional-Order Dynamic Photovoltaic Model Parameters Estimation Based on Chaotic Meta-Heuristic Optimization Algorithms

Abstract

Modeling of Photovoltaic (PV) solar modules is an essential target in achieving an efficient emulation for the PV system. Recently, the dynamic PV models were considered to recognize the influence of the switching circuits and load change proprieties. The fractional-order dynamic PV model was the currently proposed one to boost the reliability, accuracy, and efficiency of the classical dynamic PV model. The optimal parameters of this model should be identified; therefore, in this chapter, several Chaotic biologically-inspiring Optimization techniques are proposed to demonstrate the most efficient one for this non-linear optimization problem. The introduced techniques are the chaotic variants of Grasshopper Optimizer, Moth-Flame Optimizer, and Flower Pollination Algorithm in addition to their original versions. To assess the efficiency of the endorsed algorithm, its results are compared to the non-linear least-squares method based on the accuracy, the convergence speed, and the fitting of the experimental dataset. Additionally, another comparison is carried out between the recent fractional-order dynamic PV model and its integer version based on the same algorithm to evaluate the efficiency of using the fractional calculus in the modeling of the PV modules. The overall results show that Chaotic Flower Pollination Algorithm with Chebyshev and Singer chaotic maps in the case of the fractional dynamic PV model offers the best fitting on the load current-time curve. Moreover, the fractional-order dynamic model that can emulate the physical behavior of the real system is efficient than the integer-order dynamic model.