Computational approach using machine learning modelling for optimization of transesterification process for linseed biodiesel production

Abstract

In this research work, various machine learning models such as linear regression (LR), KNN and MLP were created to predict the optimized synthesis of biodiesel from pre-treated and non-treated Linseed oil in base transesterification reaction mode. Three input parameters were included for modelling, reaction time, catalyst concentrated ion, and methanol/oil-molar ratio. In biodiesel transesterification reaction 180 samples run with non-Pre-treated Linseed Methyl Ester (NPLME), Water Pre-treated Linseed Methyl Ester (WPLME) and Enzymatic Pre-treated Linseed Methyl Ester (EPLME) oil as feed stocks and optimized parameters are find out for maximum biodiesel yield to be 8:1 molar ratio, 0.4% weight catalyst, 60 °C reaction temperature.To test the technique, R2 and MAPE parameters were used. The average R2 values for linear regression, KNN, and MLP are 0.7030, 0.8554 and 0.7864 respectively. Moreover, the average MAPE values for these models are 11.1886, 6.0873 and 8.0669 respectively. Hence, it is observed that the KNN model outperforms other models with higher accuracy and low MAPE score.