No abstract

No abstract

No abstract

The determination of photovoltaic (PV) parameters is of great importance for the reliability of solar system operation, continuity of the load power consumption, and control management of the energy source. Therefore, this study proposes an advanced backtracking search optimization algorithm (BSA) equipped with teaching and learning-based optimization (TLBO), named TLBOBSA, to accurately simulate the PV model. During the evaluation

Accurate determination of photovoltaic (PV) parameters holds immense significance for ensuring the reliability of solar system operations, uninterrupted power supply for load consumption, and efficient control and management of energy sources. I-V curves transform parameter extraction into a nonlinear optimization problem supported by the I-V data points in the PV model to characterize the PV model macroscopically. Therefore, this paper proposes a novel parameter extraction model using the Q -learning-based multistrategy improved shuffled frog leading algorithm (CRNSFLA). During the evaluation process of the proposed algorithm, the colony predation algorithm (CPA) is utilized to expand the search range of the worst individual, which is no longer confined to the line segment range between the current and best values. In the later stage of evaluation, the optimal individual serves as the starting point and is applied to the Nelder-Mead simplex (NMS) for forming a simplex to mine higher-quality solutions. Besides, the simplest reinforcement Q -learning allows for a reasonable switch between these two mechanisms. A reasonable balance between exploration and exploitation trends is ensured while making full use of the advantages of both according to the reward and punishment mechanisms. The comprehensive test results under various optimization functions, different PV models, and environmental conditions demonstrate that the proposed algorithm is more advantageous than existing algorithms for parameter extraction problems. Specifically, CRNSFLA had RMSEs of 9.8602E-04, 9.8248E-04, and 9.8248E-04 in the single-diode model (SDM), double-diode model (DDM), and three-diode model (TDM), respectively. Moreover, compared with the original shuffled frog leading algorithm, the CRNSFLA showed significant improvements in 62% of the optimization functions. Therefore, CRNSFLA can be considered an effective tool for solar cell parameter extractions.