A novel group contribution-based method for estimation of flash points of ester compounds

Flammability in esters is one of the most important features for preparation, storage, safe processing, handling and shipping of a substance. In this study, a new accurate group contribution-based method was presented for estimation of the flash points of ester compounds. The predicted flash points for a dataset of 80 esters were in good agreement with the experimental values. The obtained results showed the squared correlation coefficient (R-2) of 0.9902, root mean square error (RMS) of 5371K, and average absolute relative deviation (AARD) of 1.22% for the experimental values. To propose a predictive model, 30 ester compounds in testing set were investigated. The average percent error of 1.96% for the predicted flash point of the investigated compounds was found from the corresponding experimental values. The established model was validated and tested through the use of leave-one-out cross validation method, Y-randomization and applicability domain analysis. The results demonstrated the improved accuracy of the presented method with respect to previously proposed methods in open literatures. Therefore, the model can be well used to predict the flash points of a wide range of ester compounds, which is an accurate method for potential application. (C) 2014 Elsevier B.V. All rights reserved.