Creating an Analytical Model to Predict the Phase Angle (Δ) in The Dynamic Shear Rheometer (DSR) Test
Author(s):
Article Type:
Research/Original Article (دارای رتبه معتبر)
Abstract:
Superpave bitumen specifications are designed to improve pavement performance by controlling pavement problems under a wide range of temperatures and aging conditions. Dynamic shear rheometer (DSR) is one of the superpave tests used to determine the rheological properties of bitumen. The aim of this study is to create a prediction model with the ability to predict the phase angle (δ) as a main result of the DSR test method. This, in turn, can reduce the time to obtain laboratory results and, consequently, the cost. For this purpose, a ensemble machine learning method with a random forest approach has been used. Based on this, seven effective variables on bitumen phase angle were collected from the results of 1225 samples from the LTPP website. These factors are: test temperature, type of aging, low performance degree (PG-low), high performance degree (PG-high), penetration, kinematic viscosity and absolute viscosity (dynamic). The proposed method is confirmed through a 10-fold cross-validation method and based on the analysis, it reaches more than 90% accuracy in terms of coefficient of determination. Finally, the effect of some key factors in the random forest approach was also investigated, for example, the effect of the sensitivity of the phase angle input parameters. Also, based on the results of sensitivity analysis, the importance of different input variables was obtained. Based on the research, the test temperature and the type of aging have the greatest effect on the bitumen phase angle. By increasing the number and variety of training data, the model can be used to achieve better results and predict other performance properties of bitumen.
Keywords:
Language:
Persian
Published:
Journal of Transportation Research, Volume:21 Issue: 2, 2024
Pages:
461 to 474
https://www.magiran.com/p2731038
سامانه نویسندگان
مقالات دیگری از این نویسنده (گان)
-
Predicting Marshall Asphalt Stability Using Supervised Machine Learning Algorithms, support vector machine and random forest
*, Alireza Hasani, Sina Arman, Amir Siavash Hejazi
Journal of Transportation Research, -
Detection and classification of asphalt pavement cracks using YOLOv5
*, Ali Ghiami Bajgirani, Mohadeseh Delavarian
Journal of Modeling in Engineering,