Modelling and optimization of corrosion rate of Ni-Co-P coating using box-behnken design, gradient descent with RMSprop & metaheuristic algorithm

• Published in: International journal on interactive design and manufacturing Vol. 19; no. 8; pp. 5641 - 5656
• Main Authors: Hassan, Tarik, Mallick, Ankita, Lamichaney, Sameer, Mandal, Sukhen Das, Sarkar, Subhasish, Mondal, Nitesh, Majumdar, Gautam
• Format: Journal Article
• Language: English
• Published: Paris Springer Paris 01.08.2025
• Subjects: CAE) and Design; Computer-Aided Engineering (CAD; Electronics and Microelectronics; Engineering; Engineering Design; Industrial Design; Instrumentation; Mechanical Engineering; Original Article; Machine learning (ML); Electroless coating; Corrosion rate (CR); Gradient descent with RMSprop; Optimization
• ISSN: 1955-2513; 1955-2505
• DOI: 10.1007/s12008-024-02156-8

The present study is conducted to assess the corrosion rate of electroless Ni-Co–P alloy coating thereby optimizing and characterizing the coating over the copper substrate. The response parameter and corrosion rate have been optimized by the variation of factors like cobalt sulfate, sodium hypophosphite, and bath temperature using gradient descent with RMSprop, Artificial Bee Colony, Firefly, and Teaching-Learning-Based-Optimization algorithms. Fifteen coated samples with variable compositions are synthesized as per the three-factor three-level Box-Behnken design and their corrosion rates are evaluated from a Potentiodynamic polarization test in aerated 3.5wt% NaCl solution. A second-order regression model is obtained from the Box-Behnken design result in Minitab which is used as the input function for machine learning algorithm in Python and Matlab. The Analysis of Variance depicts that the cobalt sulfate concentration and the interaction between temperature and cobalt sulfate are the most significant parameters in determining the corrosion rate as per the regression model. The optimum corrosion rate of coated surface suggested by the algorithms is 0.435 μm/Y whereas it is 2.12 μm/Y in the case of the pure copper substrate as explored and investigated experimentally. The surface morphology of the optimized coating, its phases, and compositions are analyzed using Scanning Electron Microscopy, X-ray Diffraction, and Energy Dispersion X-ray analysis. It showed spherical deposits in a regular fashion and disc plates of Ni-Co-Phosphates with negligible cracks and crevices conforming to the optimum corrosion rate. Different Machine Learning algorithms have been successfully utilized in this study to determine the optimum corrosion rate and the corresponding concentration of independent variables and these are found under experimental results.