Insights into the corrosion resistance of Thaumatococcus daniellii on carbon steel in simulated sea water: Experimental and machine learning prediction

Abstract

The present work compared the predictive abilities of response surface methodology (RSM) and adaptive neuro fuzzy inference systems (ANFIS) in modeling of carbon steel corrosion inhibition by Thaumatococcus daniellii leaf extract (TDE). Thaumatococcus daniellii leaf extract was examined using Fourier Transform Infrared Spectroscopy (FTIR) methods which revealed that the phytochemical components present in TDE has high-value flavonoids, tannins, and dominating functional groups needed to support long-term corrosion inhibition. Error indices revealed the superiority of ANFIS (R2 = 0.99443) and RSM (R2 = 0.98932) in predicting the inhibition efficiency of carbon steel corrosion, while statistical metrics confirmed the application of RSM and ANFIS techniques in modeling the corrosion inhibition of carbon steel. Weight loss and electrochemical, methods were used to validate the predictive abilities of RSM and ANFIS. In addition, the carbon steel surface was examined post immersion using fourier transform infrared spectroscopy (FTIR), UV–Visible spectroscopy, scanning electron microscopy (SEM), energy dispersive x-ray spectroscopy (EDX), and atomic force microscopy (AFM). According to the results, the optimal values of the percentage inhibition efficiencies (IE%) of TDE from weight loss, electrochemical impedance spectroscopy (EIS), and potentiodynamic polarization (PDP) were found to be in close relationship as 88 %, 86 %, and 81 %, respectively, at a concentration of 2.0 g/L. TDE was confirmed to function as a mixed-type corrosion inhibitor according to potentiodynamic polarization results. The results of the machine learning are in line with the experimental findings.

Keywords

• Sea water corrosion
• Carbon steel
• Electrochemical
• FTIR
• UV–Visible
• Surface analyses machine learning