Investigation of correlation between chemical composition and properties of biodiesel using principal component analysis (PCA) and artificial neural network (ANN)

Biodiesel will provide a significant renewable energy source for transportation in the near future. In the present study, principal component analysis (PCA) has been used to understand the relationship between important properties of biodiesel and its chemical composition. Finally, several artificial intelligence-based models were developed to predict specific biodiesel properties based on their chemical composition. The experimental study was conducted in order to generate training data for the artificial neural network (ANN). Available (experimental) data from the literature was also employed for this modeling strategy. The analytical part of this study found a complex multi-dimensional correlation between chemical composition and biodiesel properties. Average numbers of double bonds in the chemical structure (representing the unsaturated component in biodiesel) and the poly-unsaturated component in biodiesel had a great impact on biodiesel properties. The simulation result in this study demonstrated that ANN is a useful tool for investigating the fuel properties from its chemical composition which eventually can replace the time consuming and costly experimental test.