The paper provides a combined machine learning model to predict the corrosion inhibition behavior of the Mitragyna speciosa extract in 1 M HCl solution on API 5 L X65 steel. The measurements of weight loss under different conditions of inhibitor concentration (0.2–0.8 g/L), temperature (323–343 K), and immersion time (2–10 h) were used to obtain experimental data. 60 experimental datasets were utilized to come up with predictive models using Artificial Neural Networks (ANN) and Adaptive Neuro-Fuzzy Inference Systems (ANFIS), and further refined with Genetic Algorithm (GA) optimization. The results indicate that the effectiveness of an inhibitor is significantly greater with the concentration of the inhibitor, up to around 98 %, whereas the effect of temperature is the opposite. Comparative analysis revealed that the ANN model achieved the highest predictive accuracy, attaining an R² value of 0.863 together with the lowest RMSE and MAE values. Although GA optimization improved the performance of both ANN and ANFIS models, ANN-based approaches consistently outperformed ANFIS-based models.The sensitive analysis indicated that inhibitor concentration is the most sensitive factor, then temperature and immersion time. The developed machine-learning framework effectively captured the nonlinear relationships governing corrosion inhibition behavior and provided robust prediction and optimization capability. The method is an effective and cost-efficient alternative to the traditional experimental methods, and helps in developing sustainable and plant based corrosion inhibitors.
Okuma et al. (Fri,) studied this question.