چكيده به لاتين
Rutting is one of the three main defects of asphalt pavements and occurs under the passage of cyclic loads on a specific section of the pavement. Studies show that 90% of rutting occurs under the effect of shear weakness of asphalt mixtures. The permanent deformation test under repeated loads is introduced as a simple functional test to investigate the resistance of asphalt mixtures against rutting. The most important output of this test is the Flow Number (the starting point of the third region of the permanent deformation diagram) of asphalt mixtures. The permanent deformation test under repeated loads of the test cycle is very time-consuming. Furthermore, the time-consuming nature of this testing, combined with potential delays in sample arrival, can result in substantial costs and lengthy delays.
Although the parameter G*/Sin(ϭ) utilized in the dynamic shear rheometer test has served as an effective criterion of the rutting tendencies of pure bitumen for numerous years, recent investigations have revealed its inadequacy in accurately predicting the rutting performance of bitumen. As an alternative approach to evaluating the rutting characteristics of modified bitumens, the repeated creep and recovery test has been proposed. This test applies a more realistic loading scenario to the bitumen and takes into account the delayed elastic effect exhibited by modified bitumens. However, due to the low stress level at which this test is conducted, it fails to adequately demonstrate the rutting behavior of asphalt mixtures. Consequently, in order to improve the prediction of bitumen's behavior under high temperature conditions, creep and recovery tests have been suggested at various stress levels.
The correct choice of bitumen plays a crucial role in enhancing the rutting resistance of the asphalt mixture. Therefore, in this study, a novel approach utilizing MODIS and ASTER remote sensing (RS) data, ERA5-Land reanalysis data, NASA/FLDA data, and local meteorological datasets, along with deep learning (DL) techniques, was developed to determine the appropriate bitumen performance grade (PG). Consequently, the bitumen PG selected for the specific study area, located in the southwest region of the country, was determined based on the creep and recovery test temperature under various stress levels.
Then, parameters of permanent creep compliance (Jnr) and recovery percentage (R%) of pure bitumen with SBS, SBR and EVA polymer additives were determined with percentages of 0, 2 and 4. In the following, the asphalt mixture samples of these bitumens were subjected to dynamic creep test at the stress levels of 300 and 450 and at temperatures of 40, 50 and 60 degrees Celsius to determine the Flow Number. By examining the Flow Number of polymer modified asphalt mixtures based on viscoelastoplastic behavior, after measuring the effect of different bitumen rutting on the Flow Number, a model for predicting the rutting of polymer mixtures based on bitumen parameters was presented.
The results of statistical analysis showed that the proposed method for determining bitumen PG based on deep learning can determine the performance temperature of asphalt pavement and estimate bitumen PG with much higher accuracy than conventional methods, and the potential for accurate determination of PG is independent of the distance from the meteorological station. It has a spatial resolution of 1 km. Also, the proposed model for predicting the Flow Number of polymer modified asphalt mixtures based on viscoelastoplastic behavior is able to predict the flow rate with an error of 11% using only the results of the MSCR test.
