A novel modified COOT algorithm for parameter extraction of PV modules including measurement noise consideration

The huge integration of photovoltaic (PV) systems into power systems due to fossil fuel scarcity and environmental considerations leads to serious issues in the operation of electrical grids. The reliability and accuracy of PV module models allow the optimization of PV systems. The PV cells are described by a nonlinear current-voltage characteristic, and the parameters not specified by the manufacturer's sheet are generally estimated. Despite the significant effort provided by the PV parameter extraction problem, the latter is still an arduous task. This work suggests a new modified coot optimization algorithm with local search (MCOOT-LS) for effectively extracting the unknown parameters of PV modules. More to the point, the conventional COOT algorithm profits from chaotic dynamic-opposite learning, a new exploration strategy, and a local search to reach a high balance between exploitation and exploration for complex problems. In this work, we employed two PV modules composed of silicon monocrystalline and polycrystalline cells, like the LSM 20 and STE 4/100 photovoltaic panels. Previous works in the domain of PV parameter estimation overlooked the importance of anti-noise ability. However, with a view to revealing the efficacy of the proposed MCOOT-LS, this manuscript examined the solution accuracy, convergence speed, and anti-noise properties of all the algorithms used. To assess the effect of noise on the obtained results, two levels of noise (5% and 10%) have been considered, and the objective function called a Root Mean Square Error (RMSE) is adopted. The results prove that the proposed MCOOT-LS algorithm outperformed all the comparative algorithms regarding convergence feature and solution reliability for both noise-free and noisy results.