%0 Journal Article
%T Predicting Flow Number of Asphalt Mixtures Based on the Marshall Mix design Parameters Using Multivariate Adaptive Regression Spline (MARS)
%J International Journal of Transportation Engineering
%I Tarrahan Parseh Transportation Research Institute
%Z 2322-259X
%A Ghanizadeh, Ali Reza
%A Safi Jahanshahi, Farzad
%A Khalifeh, Vahid
%A Jalali, Farhang
%D 2020
%\ 04/01/2020
%V 7
%N 4
%P 433-448
%! Predicting Flow Number of Asphalt Mixtures Based on the Marshall Mix design Parameters Using Multivariate Adaptive Regression Spline (MARS)
%K Flow Number
%K Rutting
%K Marshall mixing design parameters
%K Multivariate Adaptive Regression Spline (MARS)
%R 10.22119/ijte.2020.184115.1476
%X Rutting is one of the major distresses in the flexible pavements, which is heavily influenced by the asphalt mixtures properties at high temperatures. There are several methods for the characterization of the rutting resistance of asphalt mixtures. Flow number is one of the most important parameters that can be used for the evaluation of rutting. The flow number is measured by the dynamic creep test, which requires advanced equipment, notable cost, and time. This paper aims to develop a mathematical model for predicting the flow number of asphalt mixtures based on the Marshall mix design parameters using the Multivariate Adaptive Regression Spline (MARS). The required parameters for developing the model are as follows: percentage of fine and coarse aggregates, bitumen content, air voids content, voids in mineral aggregates, Marshall Stability, and flow. The coefficient of determination (R2) of the model for training and testing set is 0.96 and 0.97, respectively, which confirms the high accuracy of the model. Moreover, the comparison of the developed model with the existing models shows the superior performance of the developed model. It should be noted that the developed model is valid only in the range of dataset used for the modeling.
%U http://www.ijte.ir/article_102587_a8f07fe2e66ae1372afad21b947226c5.pdf