Performance evaluation and viscoelastic model establishment of recycled hot mix asphalt

. In this paper, the pavement performance of recycled hot mix asphalt with 30% recycled asphalt pavement (RAP) content is studied by mix design, splitting test, freeze-thaw splitting test and dynamic modulus test. The results show that the recycled hot mix asphalt with 30% RAP content has better high-temperature deformation resistance than hot mix asphalt, but low temperature performance is slightly worse, and the water stability is not much different. It is proved that the road performance of recycled hot mix asphalt with 30% RAP content is the same as that of hot mix asphalt. Recycled hot mix asphalt with 30% RAP content can be widely used. The flexural creep test data at different temperatures were fitted to the exponential decay function, which proved that the recycled hot mix asphalt conformed to the generalized Maxwell model, and provided parameters for the establishment and analysis of the viscoelastic mechanical model of the recycled hot mix asphalt.


Introduction
In recent years, in order to save resources and protect the environment, the recycling of asphalt mixtures has become a hot topic. Many literatures have pointed out that the performance of recycled hot mix asphalt (RHMA) is optimal when the recycled asphalt pavement (RAP) content ranges from 20%~30% [1] [2] . In this paper, the performance of recycled hot mix asphalt with 30% RAP content is evaluated by splitting test, freeze-thaw splitting test and dynamic modulus test.
As a viscoelastic material, the mechanical properties of asphalt mixture are significantly affected by the loading temperature and time. Therefore, it is very important to describe the viscoelastic model correctly and quantitatively. For this reason, many scholars have carried out a lot of experimental research and established a variety of viscoelastic models of asphalt mixture, in order to grasp the viscoelastic characteristics of asphalt mixture more accurately [3] ~ [6] . However, these studies are based on hot mix asphalt (HMA), and there is no relevant research on viscoelastic model of recycled hot mix asphalt. In this paper, the viscoelasticity of recycled hot mix asphalt is studied based on fitting the bending creep test data at different temperatures with an exponential decay function.

Mix design
In this paper, the gradation type of asphalt mixture is AC-16 dense. In order to verify the performance of RHMA, HMA with the same raw materials and conditions was used as a comparison. All technical indexes of raw materials meet the requirements of "Technical Specifications For Construction Of Highway Asphalt Pavement" (JTG F40-2004). According to the aging of the asphalt in the old material, it is determined that the mixing ratio of the rejuvenating agent is 8% of the old asphalt. According to the Marshall Design method in "Technical Specifications for Construction of Highway Asphalt Pavement" (JTG F40-2004), the optimum asphalt content was determined to be 4.7%. The gradation curve of HMA AC-16 and RHMA AC-16 is shown in Fig. 1 The test temperatures were -20℃, -10 ℃, 0℃, 10℃, and 20℃. When testing at -20℃, -10℃ and 0℃, the loading rate is 1mm / min; when testing at 10℃ and 20℃, the loading rate is 50mm / min [7] .  (1) The splitting tensile strength of RHMA and HMA increases with the decrease of temperature at first, and then decreases after the peaking. At high temperature, the splitting tensile strength of RHMA AC-16 is 6.1% higher than that of HMA AC-16, and 1.2% lower than that of low temperature. This shows that the high temperature deformation resistance of RHMA AC-16 is strong, and the low temperature performance is slightly poor; (2) According to the relationship between the failure strain and temperature, the failure strain of RHMA and HMA gradually decreases with the decrease of temperature, and there is a cross phenomenon, and the failure strain of the two is not much different;

freeze-thaw splitting test
The test is to concentrate and strengthen the impact of water on the actual road surface, so that the impact of the road surface can be simulated in a shorter period of time. Table 2 for freeze-thaw splitting strength test results. According to the data in Table 2, the splitting tensile strength of RHMA AC-16 without freeze-thaw is greater than that of HMA AC-16. This is because the splitting strength is mainly composed of the internal friction resistance of the mineral and the cohesive force of the asphalt. Due to RHMA AC-16 contain 30% RAP content, the old asphalt in the old material has a higher viscosity after being regenerated by the rejuvenating agent, which makes the splitting strength relatively high. However, it decreased rapidly after freeze-thaw cycles. Compared with the HMA AC-16, the freeze-thaw splitting strength of the RHMA AC-16 is reduced by 3.2%, and the change is not great, which proves that the water stability of RHMA AC-16 is the same as that of HMA AC-16 basically.

Dynamic modulus test
The simple performance tester (SPT) is used for the dynamic modulus test. At the specified temperature, the sinusoidal compressive stress test of 0.01 ~ 25Hz is applied to the test specimens [7] . Table 3 is the average value of three groups of dynamic modulus test results under the same temperature (15℃), the same load and different frequency. It can be known from Table 3 that the dynamic modulus of the two asphalt mixtures increases with increasing load frequency. The dynamic modulus of RHMA AC-16 is greater than that of HMA AC-16, which indicates that the deformation resistance of RHMA AC-16 is stronger than that of HMA AC-16. Table 4 shows three groups of dynamic modulus test results under the same frequency (10Hz), same load and different temperature. It can be known from Table 4 that the dynamic modulus of the two asphalt mixtures decreases with the increase of temperature. At the same temperature, the dynamic modulus of RHAM AC-16 is higher than that of HAM AC-16, especially at higher temperature. It shows that the deformation resistance of RHAM AC-16 is better than that of HAM AC-16 at high temperature.

Viscoelastic model establish
The establishment of viscoelastic model could reflect the mechanical properties of materials better, and the model should be as simple and intuitive as possible. The mechanical parameters in the model are determined by experimental simulation. The test simulation uses the theoretical solution of the mechanical model to fit the test results. By adjusting the values of the mechanical parameters, the two results can achieve the best fit effect. On the one hand, the rationality of the selected mechanical model can be verified. On the other hand, it can determine the mechanical parameters of the material. The research on the viscoelastic mechanics model of RHMA AC-16 is based on the above methods. And the creep test is carried out at -20℃, 0℃ and 25℃. The experimental data are sorted out and simulated by the generalized Maxwell model. Therefore, it is studied whether the viscoelasticity of RHMA AC-16 conforms to the generalized Maxwell model.
The second-order exponential decay function (1) is used to fit the test data of three groups (9 specimens) at -20℃, 0℃ and 25℃ each temperature. Before fitting the data, the unreasonable negative strain and the falling section of the post creep section are deleted.
The average value is shown in Table 5 and the fitting curve of fitting parameters is shown in Fig. 3~ Fig. 5. It can be known from Table 5 and Fig. 3~Fig.5 that the fitting result of -20℃ is poor, and the correlation coefficient is only 0.7755. However, the fitting effect of 0℃ and 25℃ is good, and the correlation coefficient reaches above 0.9985. Because the creep effect is not obvious at -20℃, this article only analyzes the two cases of 0℃ and 25℃. The stiffness modulus can be obtained by substituting the vertical force concentration applied in the span of the test specimens at 0℃ and 25℃ into equation S(t)=σ 0 ∕ε(t), as shown in Fig.6 The third-order exponential decay function is used to fit the two curves in Fig.6 The general equation of the fitting curve is as follows: The coefficients in equation (2) at 0 ℃ and 25 ℃ are shown in Table 6.
By fitting the correlation coefficients in Table 6, it can be known that the generalized Maxwell model better. Therefore, according to specific test conditions, multiple Maxwell unit models in parallel can well characterize the viscoelastic state of RHMA AC-16, that is, increasing the number of parallel unit models to improve the accuracy of curve fitting [4] [5] .

Conclusions
(1) The high temperature performance of RHMA with 30% RAP content is better than that of HMA. Low temperature performance is slightly worse than HMA.
(2) The water stability of RHMA with 30% RAP content is similar to that of HMA.
(3) RHMA with 30% RAP content has the same road performance as HMA, which can be widely used.
(4) The fitting results of the creep data of the test specimens at -20℃, 0℃, and 25℃ with theoretical calculations show that the viscoelastic characteristics of RHAM AC-16 conform to the generalized Maxwell model. (5) According to the fitting results of the test data and the generalized Maxwell model, the parameters are provided for the establishment of the mechanical model of the RHMA.