Abstract
In this research, quantitative structure activity relationship of Azoles as copper corrosion inhibitors was studied by artificial neural networks (ANN). For this purpose, corrosion inhibitor efficiency of Azole compounds was collected from different references. The Azoles structures were optimized by Hyperchem software. Molecular descriptors of Azoles were extracted by Dragon software and selected by principal component analysis (PCA) method. Theses structural descriptors along with environmental descriptors (pH, time of exposure, temperature and concentration of inhibitor) were used as input variables. Corrosion inhibitor efficiency of Azoles was used as output variable. Experimental data were divided randomly into two sets: training set for model building and simulation set for model validation. Linear models were investigated by multiple linear regressions (MLR). Results showed poor correlation between experimental data and model data in linear models. Hence nonlinear method such as artificial neural networks was used for studying nonlinearity of data. After optimization of network by training and validation data, built model was investigated with simulation data. The results showed good agreement between experimental and theoretical data. Therefor ANN can be used as a good tool for predicting Azole’s corrosion inhibitor efficiency for copper in the presence of environmental conditions.
Original language | English |
---|---|
Pages (from-to) | 11293-11297 |
Number of pages | 5 |
Journal | Advanced Science Letters |
Volume | 23 |
Issue number | 11 |
DOIs | |
Publication status | Published - 1 Nov 2017 |
Externally published | Yes |