Title: Corrosion estimation of Cu and Br based automotive parts exposed to biodiesel environment: case of RSM and ANN

Authors: Olusegun David Samuel; Amin Taheri-Garavand; Marcus A. Amuche; Christopher C. Enweremadu

Addresses: Department of Mechanical Engineering, Federal University of Petroleum Resources, Effurun, Delta State P.M.B 1221, Nigeria; Department of Mechanical Engineering, University of South Africa, Science Campus, Private Bag X6, Florida, 1709, South Africa ' Mechanical Engineering of Biosystems Department, Lorestan University, Khorramabad, 68138-33946, Iran ' Department of Mechanical Engineering, Federal University of Petroleum Resources, Effurun, Delta State P.M.B 1221, Nigeria ' Department of Mechanical Engineering, University of South Africa, Science Campus, Private Bag X6, Florida 1709, South Africa

Abstract: The analysis of the effects of variables when selecting automotive parts operating in biodiesel medium is critical. This novel study employs RSM and 5-fold cross-validation of ANN in the optimisation strategy for minimising corrosion rates (CRs) of automotive parts (APs) in a biodiesel environment. The hardness number (BHN) and tensile strength (TES) as well as the surface morphologies of copper and brass were investigated. The optimum CRs were 0.01656 mpy and 0.008189 mpy at a B 3.91 biodiesel/diesel blend and 240.9-hour exposure. The established ANN model configuration proved superior adaptability and nonlinearity. The ANN model had higher R² and lower values of RMSE, MAE, and AAD when compared to the RSM model, thus validating the superiority of the ANN model. Brass exhibited greater TES while copper had a higher BHN. The database, model, and correlations can assist in mitigating and effectively planning against the corrosiveness of APs when using biodiesel.

Keywords: RSM; response surface methodology; ANN; artificial neural network; corrosion; copper; brass; modelling; biodiesel.

DOI: 10.1504/IJCMSSE.2023.135837

International Journal of Computational Materials Science and Surface Engineering, 2023 Vol.11 No.3/4, pp.187 - 210

Received: 06 Sep 2022
Accepted: 24 Oct 2022
Published online: 08 Jan 2024 *