Keywords

1 Introduction

The direct coal liquidation is a technology for the clean conversion of coal into light fuel oil and high value-added chemical products. It has become an important initiative for the clean and efficient utilization of coal, which is conducive to the strategic restructuring of China’s coal industry [1]. The direct coal liquefaction residue (DCLR) produced in the process is nearly 1/3 of the original coal, which not only causes waste of resources, but also imposes a burden on the natural environment due to its difficult degradability. The composition and properties of DCLR are similar to those of lake asphalt, with a high asphaltene and gum content and a low saturated and aromatic component. It has the advantages of economy and environmental friendliness as a modifier. And it can change the ratio of asphalt viscoelastic components and improve the high temperature rheological properties of asphalt. However, it also has a negative impact on the low temperature and fatigue properties of asphalt [2].

To enhance the low-temperature performance of DCLR modified asphalt, it was found [3, 4] that the incorporation of SBS can make up for the shortcomings of the low-temperature performance of DCLR modified asphalt to a certain extent. Aromatic oil can promote the reasonable combination of asphaltene, resins, aromatics and saturates in asphalt to form a stable colloidal structure. And it can promote the compatibility of SBS and asphalt to improve the low-temperature performance of asphalt. In this paper, based on the literature studies [5,6,7,8,9,10]. DCLR, SBS and aromatic oils were compounded to modify the matrix asphalt. Nine compounding schemes were designed using orthogonal test method and dynamic frequency scan test was performed using dynamic shear rheometer (DSR). The zero-shear viscosity (ZSV) was then obtained by fitting a simplified Carreau equation model to the complex viscosity. The creep recovery rate R and the irrecoverable creep flexibility Jnr of each modified asphalt were determined by multiple stress creep and recovery (MSCR) test. And the high-temperature rheological properties of the nine composite modified asphalts were evaluated by gray correlation analysis between ZSV and high-temperature rheological parameters. Bending beam rheometer (BBR) tests were conducted on the aged composite modified asphalt, based on the viscoelastic parameters of Burgers model and linear fitting, to analyze its low temperature rheological properties.

2 Experimental Materials and Specimen Preparation

2.1 Test Materials

The selected Tokai brand 90# matrix asphalt, the conventional index test results are shown in Table 1; the selected DCLR performance is shown in Table 2; the selected modifier is a linear YH-791 SBS modifier with a block ratio of 3/7, the basic performance is shown in Table 3; the selected 2#-38 aromatic oil is used as a compatibilizer for the composite modified asphalt.

Table 1 Technical index of 90# matrix asphalt
Full size table
Table 2 Performance of DCLR
Full size table
Table 3 Basic properties of SBS
Full size table

2.2 Method

  1. (1)

    Orthogonal tests were designed to prepare nine DCLR composite modified asphalts, and the results are shown in Table 4.

    Table 4 Composite modified asphalt orthogonal test table
    Full size table
  1. (2)

    High temperature rheological test

Dynamic frequency scan test using DSR. Dynamic shear load loads at low strain levels were applied to asphalt samples at loading frequencies of 0.1–100 rad/s and at 46–82 °C (temperature step of 6 °C). And a simplified Carreau equation model is fitted to the complex viscosity to analyze its high-temperature stability.

MSCR tests were performed on each asphalt after short-term aging using DSR apparatus at 64 and 70 °C with a holding time of 15 min.

  1. (3)

    Low temperature rheological test

Fatigue testing was performed by strain scanning, using strain control mode. The test temperature was 15 °C, the strain was 1%, the number of loading was 50,000, and the frequency was 10 Hz. A low temperature bending beam rheometer was used to test the creep modulus of stiffness S and creep rate m at a load loading time of 60 s. The specimens were asphalt aged by RTFOT short-term and then by PAV long-term at −12, −18 and −24 °C.

3 High Temperature Rheological Test Results and Analysis

3.1 Asphalt ZSV Fitting Results

The ZSV of asphalt can better characterize the storage capacity of asphalt at high temperatures, and the higher its value, the better its high temperature stability. However, the measurement conditions of ZSV are demanding, requiring very low frequency or shear rate, which is difficult to be satisfied by existing test instruments. Therefore, the widely used Carreau model was adopted to fit the measured frequency band to obtain ZSV, whose equation is shown in Eq. (1), and the fitting results are shown in Fig. 1.

$$ \eta = \frac{{\eta_{0} }}{{[1 + (k\omega )^{2} ]^{\frac{m}{2}} }} $$
(1)
Fig. 1
A line graph on Z S V versus temperature. The legend depicts a matrix of asphalt from ranges 1 to 9. The lines are decreasing at different intervals and then converge at approximately 65 degrees Celsius.

ZSV of asphalt at different temperatures

Full size image

where η is the complex viscosity; η0 is zero shear viscosity; ω is the frequency; k is a constant, which is a material parameter with a time scale; and m is a constant, which is a dimensionless material parameter.

From Fig. 1, it can be seen that the ZSV values of the composite modified asphalt are higher than those of the matrix asphalt at all temperatures. This indicates that the composite modifier can improve the high-temperature elastic recovery of asphalt; the ZSV value of No.6 composite modified asphalt is the highest at all temperatures, indicating that its high-temperature performance is more excellent.

3.2 MSCR Test Results

The main evaluation indicators R and Jnr were obtained from the MSCR test and the results are presented in Table 5. The polar differences were calculated using the analysis of range (ANOR) to determine the order of influence of the factors. Where ki denotes the arithmetic mean of the test results obtained at factor level i in any column, and the polar difference R = max{k1, k2, k3}-min{k1, k2, k3}. The results of ANOR of R are presented in Table 6, and the results of ANOR of Jnr are presented in Table 7.

Table 5 MSCR test results
Full size table
Table 6 Results of ANOR of R
Full size table
Table 7 Results of ANOR of Jnr
Full size table

As can be seen from Tables 5 and 6, the R of both the composite modified asphalt 1–9 and the matrix asphalt decreased significantly when the temperature was increased to 70 °C. This indicates that the increase in temperature decreases the elastic properties of the asphalt; R0.1 is significantly higher than R3.2, which shows that it is more difficult for the asphalt to recover from deformation at high stress levels; the greatest effect on the R value is for SBS, and the higher its admixture, the greater the R value.

As can be seen from Table 7, Jnr increased for all specimens when the temperature was increased to 70 °C. This indicates that the higher the temperature, the more viscous the asphalt and the weaker the elastic recovery; Jnr3.2 was higher than Jnr0.1 for all specimens, indicating that the irrecoverable creep deformation was greater at high stress levels; the higher the doping of DCLR with SBS, the smaller the value of Jnr.

In summary, the high temperature resistance to deformation of the composite modified asphalt is better than that of the matrix asphalt, and the modifier with the greatest influence on the modified properties is SBS, and the higher the SBS admixture, the stronger its high temperature resistance to deformation. This is probably because when SBS reaches a certain concentration, its thermoplastic nature enables it to form a stable three-dimensional mesh structure in the asphalt [11, 12]. The addition of aromatic oils enhances this mesh structure [13], thus improving the high temperature deformation resistance of the composite modified asphalt. From the optimum ratio of each index, it can be seen that the high temperature performance of the composite modified asphalt is best when both DCLR and SBS are at the highest dose, but the optimum dose of aromatic oil cannot be determined, and further research and analysis are still needed.

3.3 Grey Correlation Analysis

Due to the number and complexity of the indicators used to evaluate the high temperature performance of asphalt and the large dispersion of the data for each indicator. It is difficult to carry out mathematical and statistical processing to find out the main indicators for evaluating the high temperature performance of asphalt. Grey theory introduces the correlation degree into the system analysis, which can solve the shortcomings of mathematical and statistical methods in the analysis of the large amount of calculation and many samples. The correlations between R and Jnr at different temperatures and stress levels obtained by coupling MSCR tests, combined with the variation pattern of ZSV indicators, were quantitatively compared and analysed, leading to an in-depth analysis of the primary and secondary relationships between them. The results of the ZSV (64 °C) fitted by the Carreau model were chosen as the reference series and the MSCR test results were used as the comparison series to obtain the grey correlations of the factors, as follows.

  1. (1)

    Grey Relational Coefficient

Let the set of grey correlation factors be Zm×n(z1, zi), z1(m) is the reference column and is the comparison column. The number of grey-off links is

$$ \xi = \left| {\frac{{\mathop {\min }\limits_{i = 2,n} \cdot \mathop {\min }\limits_{t = 1,n} {\Delta }_{i} ({\text{t}}) + \rho \mathop {{\text{max}}}\limits_{i = 2,n} \cdot \mathop {{\text{max}}}\limits_{t = 1,n} {\Delta }_{i} ({\text{t}})}}{{{\Delta }_{i} ({\text{t}}) + \rho \mathop {{\text{max}}}\limits_{i = 2,n} \cdot \mathop {{\text{max}}}\limits_{t = 1,n} {\Delta }_{i} ({\text{t}})}}} \right| $$
(2)

where t = 1, 2,…, n; \({\Delta }_{i} (t) = \left| {z_{1} - z_{i} } \right|\); \(\rho\) is a resolution factor generally taken as 0.5.

  1. (2)

    Grey correlation entropy

The distribution density function \(P_{{\text{h}}}\) is noted as:

$$ P_{h} \mathop = \limits^{\Delta } {\xi \mathord{\left/ {\vphantom {\xi {\sum\limits_{t = 1}^{n} \xi }}} \right. \kern-0pt} {\sum\limits_{t = 1}^{n} \xi }} $$
(3)

The grey correlation entropy of zi is:

$$ H(x)\mathop { = - }\limits^{\Delta } \sum\limits_{t = 1}^{n} {P_{{\text{h}}} \ln P_{{\text{h}}} } $$
(4)
  1. (3)

    Grey entropy correlation

The grey entropy correlation of zi is

$$ E(z_{{\text{i}}} )\mathop = \limits^{\Delta } {{H(x)} \mathord{\left/ {\vphantom {{H(x)} {H_{{{\text{max}}}} }}} \right. \kern-0pt} {H_{{{\text{max}}}} }} $$
(5)

where Hmax = lnm is the maximum value of the difference information column consisting of m elements. The larger E(zi) is, the greater the correlation between the comparison sequence zi and the reference sequence z1.

The parameters of the grey correlation analysis are shown in Table 8 and the dimensionless results and grey correlations are shown in Tables 9 and 10.

Table 8 Corresponding parameters of grey correlation analysis
Full size table
Table 9 Nondimensionalize calculation results
Full size table
Table 10 Grey relational calculation results
Full size table

As can be seen from Table 10, the ash correlation between each index and ZSV is high, all above 0.97, with Jnr3.2 of asphalt at 70 °C having the highest ash correlation with ZSV, followed in order by Jnr0.1 at 64 °C, Jnr0.1 at 70 °C, R0.1 at 64 °C, Jnr0.1 at 64 °C, R3.2 at 64 °C, R0.1 at 70 °C and R3.2. This indicates that the 70 °C non-recoverable creep flexibility Jnr3.2 can better respond to the high temperature performance of the asphalt and the best ratio for high temperature performance is 9% DCLR + 4% SBS + 2% aromatic oil.

4 Low Temperature Rheological Test Results and Analysis

4.1 BBR Test Data and Burgers Model Parameters

As a single consideration of the modulus of stiffness S or creep rate m to evaluate the low temperature performance of asphalt is more one-sided, the BBR test data (−18 °C) was subjected to Burgers fitting. The results are shown in Table 11, and the results of ANOR of Burgers model parameters are shown in Table 12. The Burgers model equation is given in Eq. (6).

$$ y = \frac{1}{{E_{1} }} + \frac{{1 - e^{{\frac{{ - tE_{2} }}{{\eta_{2} }}}} }}{{E_{2} }} + \frac{t}{{\eta_{1} }} $$
(6)
Table 11 Fitting results of low temperature viscoelastic parameters for asphalt PAV state
Full size table
Table 12 Results of ANOR of burgers model parameters
Full size table

where E2 is the instantaneous elastic parameter, E2 is the delayed elastic parameter, η1 is the viscous flow parameter and η2 is the delayed viscous parameter. η1 reflects the deformation capacity of the asphalt, the smaller the η1, the better the low temperature performance of the asphalt [14].

The relaxation time λ of asphalt represents the ability of stress dissipation, the shorter the λ, representing the more rapid dissipation of stress within the asphalt, the better the low temperature performance [15]. Calculated as shown in Eq. (7).

$$ \lambda = {{\eta_{1} } \mathord{\left/ {\vphantom {{\eta_{1} } {E_{1} }}} \right. \kern-0pt} {E_{1} }} $$
(7)

The low temperature flexibility parameter JC reflects the viscoelastic properties of the asphalt, the smaller the JC, the higher the proportion of viscous components of the asphalt, the better the low temperature performance [16]. The calculation is given in Eq. (8).

$$ J_{C} = \frac{1}{{J_{V} }}\left( {1 - \frac{{J_{E} + J_{DE} }}{{J_{E} + J_{DE} + J_{V} }}} \right) $$
(8)

where \(J_{V} = \frac{{\text{t}}}{{\eta_{1} }}\), \(J_{E} = \frac{1}{{E_{1} }}\), \(J_{DE} = \frac{1}{{E_{2} }}\left( {1 - {\text{e}}^{{\frac{{ - tE_{2} }}{{\eta_{2} }}}} } \right)\).

From Tables 11 and 12 can be seen, Burgers fitting accuracy R2 are above 0.99, indicating that the model can better reflect the creep process of asphalt. The composite modified asphalt, η1, λ and JC are basically larger than the matrix asphalt, and the higher the DCLR dose, the worse the low temperature performance of the composite modified asphalt. This indicates that the performance of the composite modified asphalt in low temperature performance has some defects. This is mainly due to the high proportion of asphalt in DCLR, which cross-linked with the matrix asphalt and increased the flow resistance of the modified asphalt [17]. The ratio of 5% DCLR + 2% SBS + 2% aromatic oil is better.

4.2 Low Temperature Sensitivity Analysis

The creep stiffness and creep rate versus temperature were regressed for nine asphalt tests and the fitted results are shown in Table 13, and the results of ANOR of Burgers model parameters are shown in Table 14, with the linear regression equation in Eq. (9).

$$ \lg S = SA_{S} \cdot T + C $$
(9)
Table 13 Fitting results of low temperature sensitivity for asphalt
Full size table
Table 14 Results of ANOR of SAS
Full size table

where S represents the creep stiffness; SAS represents the slope of the equation; C represents the temperature; and C is a constant.

As can be seen from Table 13, the linear fit correlation coefficients are all above 0.94, which is a good fit; the values of the composite modified bitumen are equal to or somewhat smaller than the |SAS| values of the matrix asphalt, which indicates that the low temperature sensitivity of the composite modified asphalt has been reduced and the low temperature crack resistance has been slightly enhanced; as can be seen from Table 14, the best low temperature sensitivity is 9% DCLR + 4%/6% SBS + 3% aromatic oil.

5 Conclusions

  1. (1)

    At 46–82 °C, the ZSV values fitted by the Carreau model were higher than those of the matrix asphalt, and the highest ZSV value was obtained with 9% DCLR + 4% SBS + 2% aromatic oil. The higher the amount of SBS, the stronger the high temperature deformation resistance. Based on the grey correlation entropy analysis, the non-recoverable creep flexibility Jnr3.2 at 70 °C can better reflect the high temperature performance of the asphalt, and the DCLR composite modified asphalt with 9% DCLR + 4% SBS + 2% aromatic oil has the best high temperature performance.

  2. (2)

    The accuracy of the Burgers model is above 0.99, which indicates that the model can better reflect the creep process of asphalt. The DCLR composite modified asphalt with a ratio of 5% DCLR + 2% SBS + 2% aromatic oil has the best temperature performance. The low temperature sensitivity of the DCLR composite modified asphalt is reduced, and the low temperature crack resistance is slightly enhanced.