Quantitative Study of Non-Linear Convection Diffusion Equations for a Rotating-Disc Electrode.

Area of Science:

• Electrochemistry
• Computational Fluid Dynamics
• Applied Mathematics

Background:

• Rotating-disc electrodes (RDEs) are crucial for analyzing electrochemical processes in machinery like engines and generators.
• Modeling these processes often involves complex nonlinear entropy convection-diffusion equations.
• Accurate characterization of ion concentrations is vital for performance optimization.

Purpose of the Study:

• To investigate surrogate solutions for non-dimensional OH- and H+ ion concentrations at RDEs.
• To develop and validate an intelligent hybrid technique for this modeling.
• To assess the accuracy and reliability of the proposed computational method.

Main Methods:

• Utilized nonlinear entropy convection-diffusion equations with semi-boundaries.
• Employed an intelligent hybrid technique combining neural networks (NN) and the Levenberg-Marquardt algorithm (LMA).
• Calculated reference solutions using the RK-4 numerical method for training, validation, and testing.

Main Results:

• The NN-BLMA approximations accurately represented ion concentrations at the RDE.
• Error histograms, absolute error, curve fitting, and regression graphs confirmed the method's resilience and accuracy.
• Comparison graphs demonstrated the NN-BLMA procedure's reliability against reference solutions.

Conclusions:

• The intelligent hybrid technique (NN-BLMA) provides a reliable and accurate paradigm for modeling ion concentrations in RDEs.
• This approach enhances the analysis of electrochemical processes in revolving machines.
• The study validates the efficacy of AI-driven methods in complex scientific modeling.