Precise Prediction on the Corrosion Prevention Ability of 1,2,4-Triazole Derivatives: An Artificial Neural Network Approach
Abstract
Corrosion can have a significant impact, including leaking of hazardous substances into the environment and damaging technological devices, which can impair ecosystems and cause pollution. Establishing new methods to prevent corrosion promotes the environmental sustainability. Corrosion inhibitors minimize material waste, prolongs material life, and minimize environmental pollution, with the initiatives for waste optimization of industrial processes. Quantitative structure activity relationship (QSAR) with artificial neural network (ANN) can assist inventory optimization by quickly identifying new corrosion inhibitor compounds, which leads to economically profitable for the end users. Early studies explained the inhibitory effectiveness (IE) and the molecule's quantum characteristics were related, and a process was developed to establish a potential non-linear equation between corrosion inhibitors' experimental inhibition efficiency (IEexp) and specific quantum parameters for 1,2,4-Triazole derivatives. The current work aims to compare the prediction of non-linear model from the previous study and ANN used in this study. According to the research, the ANN predicts inhibition efficiencies more accurately and efficiently than the non-linear model, where correlation coefficient R (between IEexp and calculated inhibition efficiency (IEcalc) increased from 0.888 to 0.967 and mean square error decreased from 4.33×10-4 to 9.81×10-5 respectively. This study summarizes the application of ANN to overcome non-linear problems.