An improved gaining-sharing knowledge algorithm for parameter extraction of photovoltaic models

The optimal operation of solar cells depends on the accurate determination of parameters in the Photovoltaic (PV) models, such as resistance and currents, which may vary due to unstable weathers conditions and equipment aging. The precise selection of these parameters resembles a multi-variable, nonlinear and multi-modal problem. Despite a few parameter extraction techniques being available to solve such a problem, more-accurate and advanced solutions still present a challenging research question. This paper therefore proposes an improved gaining-sharing knowledge (IGSK) algorithm to accurately and precisely extract the parameters of PV models. The improvement in the classical GSK algorithm is incorporated by introducing an adaptive mechanism to automatically adjust the value of the knowledge rate parameter. This adaptive mechanism ensures the balance between the number of dimensions updated by the junior gaining-sharing phase and the number of dimensions updated by the senior gaining-sharing phase. A bound-constraint handling method is also presented and a linear population size reduction technique is used to boost the speed of convergence and to maintain a tared-off between the exploration and exploitation properties. The efficacy of the proposed IGSK has been demonstrated by considering three different PV modules models, i.e., single diode, double diode, and PV modules and two other commercial ones (Thin Film ST40 and Mono-crystalline SM55). For those modules, the proposed IGSK receptively produces the following outcomes: 0.00098602188, 0.0009827277, 0.0024250749, 0.0017298137, and 0.016600603. The statistical obtained results demonstrate that the IGSK indicates competitive or even better performance on convergence speed, accuracy and reliability compared with other competing techniques. Therefore, the proposed approach is believed to be an effective and efficient alternative for parameter extraction of PV models.