Prediction of upper flammability limits of binary fuel mixtures

The main affecting factors of the upper flammability limits(UFL)of fuel mixtures were determined in this paper, such as fuel stoichiometric concentration, critical pressure, heat of combustion. Based on the above affecting factors, the physico-chemical parameters which characterize fuel mixtures were constructed. And then, the quantitative relationship between those parameters and the upper flammability limits of binary fuel mixtures was studied by using multiple linear regression and multiple nonlinear regression method, respectively, and the upper flammability limit prediction model of binary fuel mixtures was established. For the training set, the average absolute deviations between the experimental and predicted values of upper flammability limit were 2.39% by linear regression analysis method and 1.272% by nonlinear regression method, while for the testing set, they were 2.185% and 1.888%, respectively. The results showed that the predicted upper flammability limits were in good agreement with the experimental data by both linear regression analysis method and nonlinear regression method. The paper provides a new method for predicting UFL of binary fuel mixtures in engineering.