Dynamic Lévy-Brownian marine predator algorithm for photovoltaic model parameters optimization.

Area of Science:

• Renewable Energy Systems
• Computational Intelligence
• Electrical Engineering

Background:

• Solar photovoltaic (PV) systems present complex characteristics due to their dynamic and multimodal nature, complicating accurate modeling.
• Traditional optimization methods, like the marine predator algorithm (MPA), face challenges with unpredictable transitions between search mechanisms (Lévy flight and Brownian walk).

Purpose of the Study:

• To introduce a novel dynamic shift function for the marine predator algorithm (MPA) to enhance its performance in solar photovoltaic (PV) modeling.
• To develop a more robust and efficient optimization technique for accurate PV parameter estimation.

Main Methods:

• Development of a dynamic shift function to modulate the interplay between Brownian walk (BW) and Lévy flight (LF) in the MPA.
• Integration of a constraint handling technique to address parameterization limitations in PV modeling, creating the dynamic Lévy-Brownian MPA (DLBMPA).
• Comparative performance analysis of DLBMPA against ten established optimization algorithms using various PV models (SDM, DDM, TDM).

Main Results:

• DLBMPA achieved a statistically significant average RMSE of 9.7 × 10^-4 in PV parameter estimation, outperforming ten other algorithms (p < 0.05).
• The optimized DLBMPA demonstrated high accuracy across different irradiance and temperature levels with an average computation time of 13 ms.
• DLBMPA exhibited superior speed of convergence and accuracy compared to existing techniques for PV parameter estimation.

Conclusions:

• The proposed dynamic shift function significantly enhances MPA performance for PV modeling.
• DLBMPA offers a highly efficient, accurate, and fast solution for estimating solar photovoltaic parameters, overcoming limitations of traditional methods.
• DLBMPA represents a significant advancement in computational intelligence for reliable solar energy system analysis.