Multi-group grey wolf optimizer (MG-GWO) for estimating photovoltaic solar cell model

The use of photovoltaic (PV) solar systems as a direct convertor of electricity is increasing on daily basis at industrial and domestic scales. Such systems are still under worldwide investigation due to their low efficiency to have better performance. In this regard, the parameters effecting the PV solar system performance should be defined and investigated using mathematical models, to determine the optimum values of these parameters that result in best possible performance. In this paper, a new and novel method based on grey wolf optimizer (GWO) algorithm is developed for the estimation of the photovoltaic solar cell model. The new proposed method is called multi-group grey wolf optimizer (MG-GWO) where several clans/packs of wolves are searching for the prey. The GWO and MG-GWO are metaheuristic techniques that mimic the leadership hierarchy and hunting behavior of clan or clans of grey wolfs. The clan consists of four levels of leadership ranked from the highest to the lowest as Alpha, Beta, Delta and Omega. On the other hand, the hunting behavior consists of three steps, searching, encircling, and attacking the prey. The algorithm mimics these levels and behavior to find the solution. In the present study, these metaheuristic techniques are used to extract the parameters of a single-diode photovoltaic (PV) solar cell model. The optimization results showed that MG-GWO is better in terms of robustness, and speed of convergence compared to conventional GWO. For more comprehensive comparison, these two methods are compared to the conventional PSO and some recent versions of PSO like time-varying accelerated coefficient PSO (PSOTAC), asymmetric time-varying acceleration coefficient PSO (PSOM), and its improved version (PSOI). The results show that MG-GWO has a superior performance compared to the other algorithms. They also show that MG reduces the values of the RMSE and MAE of GWO up to 77% depending on the number of the packs and population in each pack.