Data from multimodal functions based on an array of photovoltaic modules and an approximation with artificial neural...

Area of Science:

• Electrical Engineering
• Renewable Energy Systems
• Computational Intelligence

Background:

• Partial shading significantly impacts photovoltaic (PV) module performance, leading to complex power-voltage (P-V) curves with multiple local maxima.
• Accurate modeling of these multimodal P-V curves is crucial for efficient maximum power point tracking (MPPT).
• Previous research explored optimization algorithms for navigating these complex functions.

Purpose of the Study:

• To generate and present a dataset of multimodal functions emulating PV array performance under partial shading.
• To provide data representing both mathematically modeled P-V curves and their neural network approximations.
• To facilitate the evaluation of optimization algorithms and system identification techniques for multimodal functions.

Main Methods:

• Mathematical modeling was employed to derive multimodal functions representing P-V curves of a five-module PV array under partial shading.
• A feedforward neural network was utilized to approximate the derived multimodal functions.
• Simulations were conducted using C code, with data exported to DAT files and organized into Excel tables.

Main Results:

• The study generated datasets containing voltage and power data for five PV modules.
• Data includes both the original multimodal functions and their neural network approximations.
• The dataset is structured to support the analysis of optimization and system identification methods.

Conclusions:

• The presented dataset offers a valuable resource for researchers studying MPPT in PV systems under shading conditions.
• It enables comparative analysis of various optimization algorithms and system identification techniques.
• The data supports the development and validation of advanced control strategies for PV systems.