An enhanced neural network algorithm and its applications for numerical optimization and parameter extraction of...

Area of Science:

• Renewable Energy Systems
• Artificial Intelligence in Engineering
• Computational Optimization

Background:

• Photovoltaic (PV) systems are crucial for solar energy conversion.
• Optimizing PV systems requires accurate parameter extraction from PV models.
• Existing methods face challenges in efficiently and accurately estimating these parameters.

Purpose of the Study:

• To propose an Enhanced Neural Network Algorithm (ENNA) for accurate PV model parameter extraction.
• To improve the efficiency and performance of solar energy conversion.
• To provide a novel computational approach for complex optimization problems in renewable energy.

Main Methods:

• Development of ENNA featuring a novel transfer operator with three learning strategies.
• Integration of perturbation and elite operators to leverage population information.
• Application of ENNA to benchmark functions and three PV models: Single Diode Model (SDM), Double Diode Model (DDM), and PV Module Model (PVM).

Main Results:

• ENNA achieved optimal root mean square errors of 0.00098602 (SDM), 0.000982485 (DDM), and 0.00242507 (PVM).
• ENNA outperformed 10 other metaheuristic algorithms in numerical comparison, ranking, and convergence.
• Demonstrated excellent performance in PV model parameter estimation.

Conclusions:

• ENNA is a highly effective algorithm for estimating unknown parameters in PV models.
• The proposed method significantly enhances the optimization and simulation of PV systems.
• ENNA offers a robust and efficient solution for solar energy conversion efficiency improvement.