Parameter estimation of three diode solar PV cell using chaotic dragonfly algorithm

The efficiency of a photovoltaic system may be increased with the use of effective solar PV cell modelling. Solar cell characteristics with flaws, on the other hand, have an unfavorable impact on PV cell modelling. In most cases, manufacturers do not provide all of the information required for accurate PV cell modelling. As a result, it is vital to accurately anticipate the characteristics of the PV cell. Although the literature describes different optimization techniques, most of them provide unsatisfactory results owing to their convergence toward local minima. This research presents a new stochastic optimization approach for estimating the parameters of solar PV cells. As a result, this work introduces the Chaotic Dragonfly Algorithm, a new chaotic algorithm for evaluating solar cells. The suggested technique has the major benefit of employing chaotic maps to calculate and automatically change the internal parameters of the optimization algorithm. This circumstance is advantageous in difficult situations since the suggested method enhances their ability to seek for the optimal solution throughout the iterative phase. Complex and multimodal objective functions can be optimised using the modified technique. In order to show the potential of the proposed algorithm in the solar cell architecture, it is contrasted with other methods of optimization over two different datasets. The chaotic variants for R.T.C France model (CDA1 1.1543, CDA2 2.1896, CDA3 2.1994, CDA4 2.2011, and CDA5 2.2015) have a faster computation time than the other compared algorithms (particle swarm optimization (PSO) 10.1458, Sine Cosine Algorithm (SCA) 7.9980, Multi Verse optimization (MVO) 4.8450, Grey Wolf optimizer (GWO) 3.9042, and Dragonfly Algorithm (DA) .8056). Similarly, the chaotic variants for Photo-Watt model (CDA1-1.1441, CDA2-1.1864, CTSA3-2.1989, CTSA4-2.2014, and CTSA5-2.2022) had a faster calculation time than the other examined algorithms for the triple-diode model (PSO-10.2045, SCA-7.9999, MVO-4.8544, GWO-3.9057, and DA-3.8110). Nonparametric test, statistical error analysis, and sensitivity temperature variation are all used to prove the suggested algorithm’s superiority.