Laboratory-based modelling of viscoelastic properties of asphalt pavement
Abstract
The Norwegian Public Roads Administration (NPRA) is running a project “VegDim” with the aim of developing and implementing a Mechanistic-Empirical (ME) pavement design system for Norwegian conditions. The traditional empirical method uses resilient modulus to characterise the asphalt material by considering pure elastic properties. However, the asphalt material demonstrates a viscoelastic behaviour in servicing conditions of the flexible pavement. Therefore, the dynamic modulus closer to the real situation, which is time and temperature dependent, is adopted in the ME pavement design. The objective of this study is to investigate the dynamic modulus of various Norwegian asphalt mixtures and model the viscoelastic properties of asphalt mixtures and pavements.
Asphalt mixtures mostly used in Norway are divided into 20 types, based on the kinds of mixture and bitumen. In this study, the Marshall mix design was first conducted to determine the optimum binder contents of 20 types of asphalt mixtures containing Norwegian local aggregate materials. Then the asphalt slabs and gyratory samples based on the required aggregate gradings, and the optimum binder content were created by the roller compactor and the gyratory compactor. The roller compression method was determined as the asphalt sample preparation approach as it is believed to be closer to the paving situation in the field.
Afterwards, the Cyclic Indirect Tensile Test (CITT) and Cyclic Compression Test (CCT) under the same temperature and loading frequency conditions were compared regarding the measured dynamic modulus of asphalt mixtures in the laboratory. The results showed that the two test modes led to similar dynamic modulus values at intermediate temperatures (frequencies) and some differences at extreme conditions. Thus, the CCT was considered to measure the dynamic modulus of asphalt mixtures due to the sample size advantage for both laboratory and field samples in this study.
The master curves were established through the test results to predict the dynamic modulus of asphalt mixtures at any temperature and loading frequencies. Different models, Standard Logistic Sigmoidal (SLS) model, the Generalised Logistic Sigmoidal (GLS) model and the Christensen-Anderson-Marasteanu (CAM) model, for master curve construction were estimated to find an optimum model for Norwegian asphalt mixtures as well as the shifting techniques of log-linear, quadratic polynomial function, Arrhenius, William-Landel-Ferry (WLF) and Kaelble methods. The result indicated that the SLS model had the most suitable fit for the specimens tested. The WLF equation was recommended for master curve construction due to better fit and convenience of temperature and shift factor transition. Therefore, the SLS model and WLF equation were used to construct the dynamic modulus master curves.
As the modelling method was determined, the dynamic modulus master curves of asphalt mixtures were established as a database for Norwegian materials, which further was the input for the ERAPave program for asphalt pavement structure design.
This research also investigated the relationship between material factors and dynamic modulus. The rheological properties of bitumen were characterised using a dynamic shear rheometer, in which the bitumen viscosity was fitted log-linearly by the temperature and the complex shear modulus master curve of bitumen was constructed by the modified Huet-Sayegh model. The grey relational analysis revealed the correlation existing between material factors and dynamic modulus. A prediction model was established using multiple linear regression based on the material factors to reduce dynamic modulus tests. The results indicated that the dynamic modulus of asphalt mixtures mainly depended on the bitumen type. The correlation between dynamic modulus and the material factors was found to show a good fit of the model with Coefficients of Determination, R2 ≥ 0.973.
Finally, the Finite Element (FE) method was implemented to predict the stress-strain response of the asphalt pavement. The viscoelastic behaviour of the materials was described by the generalised Maxwell model. The FE model was established through the relaxation modulus converted from the dynamic modulus master curve. The FE modelling was verified through the stress-strain response of the CITT. The results of the FE model matched the measured values. After validation, the numerical model incorporating the dynamic modulus master curve developed in this study was a very useful tool to predict the stress and strain responses of flexible road pavements.
This research provided a Norwegian material database and proposed the dynamic modulus modelling method and implement the modelling for the stress-strain response of asphalt pavements. This study can reduce huge test processes and provide a more accurate method to predict the viscoelastic behaviour of asphalt materials in cold regions.
Has parts
Paper 1: Chen, Hao; Barbieri, Diego Maria; Hoff, Inge; Mork, Helge; Wathne, Pål; Liu, Gang. Construction of asphalt mixture master curves for a Norwegian mechanistic-empirical pavement design system. Eleventh International Conference on the Bearing Capacity of Roads, Railways and Airfields, Volume 2. Proceedings of the eleventh international conference CRC Press 2021 https://doi.org/10.1201/9781003222897 CC BY-NC-ND 4.0 licensePaper 2: Chen, Hao; Barbieri, Diego Maria; Zhang, Xuemei; Hoff, Inge. Reliability of Calculation of Dynamic Modulus for Asphalt Mixtures Using Different Master Curve Models and Shift Factor Equations. Materials 2022 ;Volum 15.(12) https://doi.org/10.3390/ma15124325 This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license
Paper 3: Chen, Hao; Garba,Saba, Rabbira; Gang, Liu; Barbieri, Diego Maria; Zhang, Xuemei; Hoff,Inge. Influence of material factors on the determination of dynamic moduli and associated prediction models for different types of asphalt mixtures Construction and Building Materials 365 (2023) 130134 https://doi.org/10.1016/j.conbuildmat.2022.130134 This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).
Paper 4: Chen, Hao; Mulugeta,Alamnie, Mequanent; Barbieri, Diego Maria; Zhang, Xuemei; Gang, Liu; Hoff,Inge. Comparative study of indirect tensile test and uniaxial compression test on asphalt mixtures: Dynamic modulus and stress-strain state. Construction and Building Materials 366 (2023) 130187 https://doi.org/10.1016/j.conbuildmat.2022.130187 This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).
Paper 5: Chen, Hao; Hoff,Inge; Gang, Liu; Zhang, Xuemei; Barbieri, Diego Maria; Fusong,Wang; Jianan, Liu. Development of finite element model based on indirect tensile test for various asphalt mixtures