Keywords

1 Introduction

In recent years, the recycling of asphalt pavement has gained considerable momentum due to its numerous advantages, including cost reduction during initial implementation, resource and waste area preservation, and positive environmental implications [1, 2]. The disposal of reclaimed asphalt pavement (RAP) not only causes air and water pollution but also results in substantial resource wastage [3, 4]. The production of new asphalt mixture involves extensive mining of sand and gravel, posing a threat to soil and vegetation preservation. To align with the country’s environmental protection efforts and global sustainable development principles, the research on recycling old asphalt pavement has gained significant traction in China [5].

The regeneration process of asphalt pavement encompasses in-place recycling and central plant recycling, with in-place recycling further classified into cold in-place recycling (CIR) and hot in-place recycling (HIR) [6,7,8]. Similarly, central plant recycling is categorized into central plant cold recycling and central plant hot recycling [6, 9]. Currently, considerable investigation has been undertaken on the selection and action principles of regenerants, the aging mechanism of asphalt, performance characteristics of RAP, and the mixing ratio of new asphalt mixture in the HIR process of asphalt [10,11,12]. HIR technology is becoming increasingly prevalent in asphalt pavement maintenance engineering. The process involves three primary stages: pavement preheating, milling and loosening of the original surface, and paving and rolling. During the process, the original road surface is fully heated using a heating machine, milled and dispersed, and then mixed with recycled asphalt mixture in a continuous mixer [8, 13]. Appropriate amounts of regenerant, new asphalt, and new aggregate are added to achieve uniform mixing. The mixture is subsequently transported to the paver for paving and vibrating before being compacted using roller equipment. The geothermal regeneration technology offers six advantages: rapid construction with short cycle duration, minimal noise pollution, suitability for repairing pavement damage and cracks, the potential to improve aggregate gradation, reduce asphalt content, and enhance pavement performance [14].

HIR of asphalt pavement involves several key technologies that play crucial roles in the construction process [15]. Research in China has yielded promising results in this area. However, previous studies have primarily focused on the ratio and performance of regenerants and recycled asphalt mixtures, mainly at the basic theory and method research stage [10, 16, 17]. Few studies have explored the technology’s construction applicability or its integration with field construction conditions. The influence of HIR on the propagation of reflection cracks has also received limited attention [18].

The pretreatment of the original pavement in the HIR process is of utmost importance [19]. Neglecting this stage can lead to the rapid occurrence of early diseases in the recycled pavement, leading to unsatisfactory results. In China, semi-rigid base asphalt pavement is a prevalent structural form, often accompanied by reflection cracks extending from the base layer to the asphalt surface layer [20,21,22]. Since HIR targets only a 4 cm surface layer, neglecting the pretreatment of reflective cracks can result in their penetration through the 4 cm HIR layer. Addressing reflection cracks before HIR can be labor-intensive and slow down construction progress.

Hence, it is imperative to investigate the propagation mechanism of reflection cracks in HIR of asphalt pavement [23, 24]. This analysis aims to determine the necessity of treating cracks in the original pavement during construction, considering economic and practical feasibility. The study assesses whether this treatment affects other pavement indicators. Employing a combination of physical engineering and theoretical analysis, the paper uses a fatigue fracture joint simulation program to numerically simulate the fatigue fracture process in pavement structures with HIR materials under wheel load and temperature conditions. Exploring the influence of material properties, layers, and initial crack positions of HIR materials on fatigue cracking, the paper seeks to provide insights into managing reflection cracks effectively and optimizing the performance of asphalt pavement structures.

2 Establishment of Numerical Model

2.1 Determination of Pavement Structure

Taking into consideration the common semi-rigid base asphalt pavement structure in China, the pavement structure combination comprises the following layers: a 4 cm hot in-place recycling AC-13 surface layer, a 6 cm AC-20 middle surface layer, an 8 cm AC-25 lower surface layer, a 30 cm cement stabilized macadam base, and a soil base (as illustrated in Fig. 1).

Fig. 1.
figure 1

Pavement structure.

Full size image

A 2D pavement structure model is established in finite element software with a width and depth of 4 m and 2 m, using linear plane strain elements. Applying a standard axle load of 100 kN, a pressure of 0.7 MPa, an equivalent circle diameter of 21.3 cm, and a center distance of 31.95 cm between wheels, a vertical pressure of 117371 Pa is exerted on the upper surface. This pressure is converted based on static equivalence principles, maintaining the same spacing and range of action as the standard load. Symmetrical boundary conditions are set on the left and right sides, while fixed boundary conditions are applied at the bottom. A meshing scheme (Fig. 2) enhances computational efficiency by increasing density from top to bottom and middle to sides, resulting in sparser meshes at the sides and lower part where deformation and damage typically occur within the specified wheel load range for each structural layer.

Fig. 2.
figure 2

Pavement structure mesh.

Full size image

2.2 Calibration of Material Parameters

The mechanical response and fatigue accumulation of an asphalt pavement structure under load are significantly influenced by the material parameters of each structural layer. In this paper, particular attention is given to the impact of HIR asphalt mixture, which introduces notable changes to the material parameters of this layer. The detailed material parameters for each structural layer in the model are presented in Table 1. It is important to note that the analysis in this paper focuses solely on the influence of HIR asphalt mixture, and does not account for fatigue damage caused by the soil base.

Table 1. Material parameters
Full size table

3 Fatigue Analysis of Pavement Structure

3.1 Evolution Process of Pavement Structure Damage Field

If the fatigue damage of the pavement structure is disregarded, the initial application of an approximate load on the pavement generates the maximum tensile stress beneath the centerline of the two wheels at the bottom of the base, leading to damage initiation from this point. Figure 3 illustrates the damage distribution near this location for various loading cycles. The warmer colors in the figure represent greater damage. The initial damage predominantly occurs at the center of the two wheels on the bottom of the base, where the maximum tensile stress is concentrated. This damage gradually diminishes towards the surrounding area, with the surface structure experiencing primarily compressive stresses, resulting in lower damage.

Fig. 3.
figure 3

Damage field evolution of pavement structure.

Full size image

As loading cycles accumulate, the fatigue damage field in the base layer extends laterally and upwards, exhibiting a horn-like distribution. The damage pattern at the bottom of the base layer becomes evident, with the highest damage concentration near the center of the two wheels and decreasing progressively towards the periphery. The damage distribution in the surface layer differs from that in the base layer. Notably, the prominent damage under the load is relatively small, and the maximum damage does not occur near the centerline of the two-wheel load. Instead, it is situated closer to the edge of the wheel load, displaying an approximate V-shaped distribution from top to bottom.

Figure 4 presents the damage distribution of the upper, middle, and lower layers after 3.25 million loading cycles, considering the pavement and load’s symmetrical structures (only one side of the change is shown). For the bottom of the upper layer (Fig. 4a), some damage occurs near the center of the double wheel due to the proximity to the wheel load acting on the surface, aligning with slight uplift and cracking phenomena observed in the actual road surface’s wheel track belt’s center. However, as one moves further from the wheel load, the damage rapidly decreases to almost zero due to the pressure at the bottom of the layer directly below the wheel load, which is less prone to damage. Once the area directly beneath the wheel load is surpassed, the damage increases significantly, peaking at a distance of 1–3 cm from the wheel’s edge and then declining. For the bottom of the middle layer (Fig. 4b), there is minimal damage near the center of the two wheels. As one approaches the edge of the wheel load, damage gradually initiates and grows rapidly, reaching a peak at approximately 1–3 cm outside the wheel’s edge. The bottom layer (Fig. 4c) exhibits a similar pattern to the middle layer, with minimal damage near the center of the two wheels but starting at a slightly smaller range and showing some damage earlier. Beyond the wheel’s edge, the range of higher damage is more extensive, and the reduction is less apparent after reaching the peak value at 1–3 cm from the outer edge of the wheel load.

Fig. 4.
figure 4

Horizontal distribution of surface layer bottom damage.

Full size image

The center line of the two wheels at the bottom of the base layer and the outer edge of the wheel load at the bottom of each layer are chosen, with approximately 1.5 cm outward from these points serving as the observation point. The resulting variation laws of the damage degree for each layer are presented in Fig. 5.

Fig. 5.
figure 5

The damage of each layer of pavement structure.

Full size image

As the number of loading cycles increases, the maximum damage degree of each layer shows a rising trend. The early base layer and lower layer exhibit significantly greater damage compared to the middle and upper layers, attributed to force distribution within the pavement structure. In the later stage, damage in the middle and upper layers progresses at a faster rate. This can be partly attributed to the materials of the base and lower layers having accumulated considerable damage, making it difficult for them to sustain the same load-carrying capacity. Consequently, the main body of load-bearing shifts towards regions with less damage or unaffected material in the middle and upper layers, allowing for good bearing capacity diffusion and transfer. Notably, the upper layer, consisting of hot in-place recycled material, exhibits faster damage accumulation compared to fresh asphalt concrete, as observed in related research. The later stage shows a slightly higher damage degree in the upper layer than in the middle layer.

3.2 Sensitivity Analysis

Material parameters of HIR asphalt mixture. The variation in RAP content and the use of different recycling agents in various HIR projects lead to differences in the nature and dosage of RAP. Consequently, the changes in pavement damage may also vary. This section takes into account the relationship between the material parameters obtained in the third chapter and the RAP content. Material parameters for HIR mixtures with two different RAP contents are set as presented in Table 2.

Table 2. Material parameters of HIR mixture with different RAP content
Full size table

The variation laws of the damage degree for each layer are analyzed in both cases using the same method as in the previous section, as shown in Fig. 6.

Fig. 6.
figure 6

Pavement damage with different material parameters.

Full size image

Comparing Fig. 6(a) and (b), it becomes evident that the material parameters of the HIR mixture significantly influence the damage in the upper layer, and the fatigue damage of the hot recycled mixture with higher RAP content accumulates faster. When the RAP content is low, the damage degree of the upper layer remains at a minimum level. However, with an increase in RAP content, the damage degree of the upper layer exceeds that of the middle and lower layers after 2.6 million loading cycles, becoming second only to the damage degree of the base layer.

HIR asphalt mixture layer. In practical HIR projects, a thin surface layer is sometimes applied over the thermal regeneration layer, which is positioned similarly to the middle surface layer. To investigate the damage changes in the pavement structure under this scenario, the aforementioned pavement structure is modified. Specifically, the upper layer is composed of SMA-13, the middle layer is the hot recycled mixture, and the remaining layers remain unchanged. The elastic modulus of SMA-13 is 1800 MPa, while other parameters are the same as those of AC-13. Figure 7 illustrates the pavement damage changes of the HIR mixture in different layers.

Fig. 7.
figure 7

Pavement damage of different layers of HIR mixture.

Full size image

Observing Fig. 7, it becomes apparent that when the HIR layer is used as the middle layer and a mixture with better fatigue performance, such as SMA-13, is employed as the upper layer, the damage in the upper layer is reduced to a certain extent. However, the damage of the HIR structure layer itself does not show significant reductions.

Middle surface layer and base layer modulus. It is essential not only to use asphalt concrete with strong anti-rutting capability on the surface layer but also to enhance the strength of the middle layer. In this section, the modulus of the middle layer is increased from 1500 MPa to 1800 MPa, and the change in pavement structure damage is reanalyzed, as depicted in Fig. 8.

Fig. 8.
figure 8

Pavement damage of different layers of HIR mixture.

Full size image

Upon comparing Fig. 8 with Fig. 5, it is evident that the increase in the middle layer’s modulus results in the upper layer’s damage degree becoming 0.44 after 3.25 million loading cycles, which is slightly lower than the value of 0.45 before the increase. However, the change is negligible. The enhancement of the middle layer’s stiffness exhibits little effect on reducing damage in the upper layer and has minimal influence on the overall damage of the pavement structure.

The base layer constitutes the thickest part of the pavement structure and serves as the primary load-bearing component. Consequently, any changes in the material parameters of the base layer often lead to significant alterations in the overall pavement structure’s damage. As accurate research results regarding the influencing factors and variation rules of base material fatigue parameters are lacking, this section focuses on changing the base modulus alone. The overall pavement structure’s damage under different base modulus values is obtained and depicted in Fig. 9.

Fig. 9.
figure 9

The influence of different base modulus on pavement damage.

Full size image

Observing Fig. 9, it becomes evident that changes in the base modulus do not substantially affect the overall damage development pattern. However, the damage degree in each layer does exhibit considerable variation. Table 3 presents the damage results of each layer after 3.25 million loading cycles.

Table 3. Different base modulus of each layer damage
Full size table

Remarkably, when solely increasing the modulus of the base layer, not only does the damage in the base layer decrease, but the damage in other layers also reduces. This phenomenon may be attributed to the fact that under actual wheel loads, the pavement experiences deformation. An increase in the base layer’s modulus, owing to its larger thickness, significantly enhances the overall pavement structure’s stiffness, thereby reducing the deformation generated under the same load. This mechanism is akin to a loading method that controls stress, leading to a reduction in the cumulative damage rate.

Overload. Currently, road transportation in China still faces a serious problem of overloading. The repetitive impact of overloading significantly degrades the pavement surface quality. This section primarily investigates whether overloading affects the fatigue behavior of the in-situ thermally regenerated pavement structure. The overload is set to 1.3 times the standard axle load, and the applied pressure in the model is changed from 117,371 Pa to 152,582 Pa. The resulting pavement structure damage development is presented in Fig. 10.

Fig. 10.
figure 10

Pavement structure damage under overload.

Full size image

Comparing Fig. 10 with Fig. 5, noticeable changes are observed in the growth trend of the middle layer and the upper layer. The curves exhibit steeper inclines as a whole, and the gap between the upper and middle layers and the lower layer reduces during the later stage. Table 4 summarizes the damage degrees of each layer after the standard axle load and overload are respectively applied for 3.25 million loading cycles.

Table 4. Different base modulus of each layer damage
Full size table

The base layer shows the smallest growth rate in damage degree, while the middle layer experiences the largest growth rate, reaching 11.63%. The upper layer also exhibits a significant growth rate. Consequently, it is evident that increasing the axle load will invariably escalate the damage degree of each layer in the pavement under the same number of loading cycles, with a more pronounced effect in the middle and upper layers. Hence, the adverse impact of overloading on pavement structure damage should be carefully considered during actual usage.

4 Cracking Analysis of Pavement Structure

4.1 Cracking Analysis of Pavement Without Initial Cracks

Assuming the pavement surface has no initial crack, the distribution of the damage field in the pavement structure just before crack initiation (at 3.3 million loading cycles) is depicted in Fig. 11. The damage primarily occurs at the bottom of the semi-rigid layer and on both sides of the tire’s grounding position. Concurrently, the main stress field of the pavement before crack initiation is presented in Fig. 12. The large modulus of the semi-rigid base results in the maximum principal stress occurring at the bottom of the semi-rigid layer. The main strain distribution on the road surface before crack initiation is illustrated in Fig. 13. The main strain concentrates primarily on both sides of the tire’s grounding position and the semi-circular arc area beneath the tire.

Fig. 11.
figure 11

Damage distribution.

Full size image
Fig. 12.
figure 12

Master stress field distribution.

Full size image
Fig. 13.
figure 13

Master strain distribution.

Full size image

Considering damage, principal stress, and principal strain, there are two potential locations for crack propagation. From the perspective of damage and principal stress, the bottom of the semi-rigid layer is the primary candidate for crack propagation. However, regarding principal strain and damage, the two sides of the tire on the asphalt surface and the semi-circular arc area below the tire are also possible locations for crack expansion. The initiation point of crack expansion predominantly depends on the cracking criterion of the pavement material.

Based on the maximum principal stress criterion. If the maximum principal stress criterion of the semi-rigid base reaches the critical value, crack initiation occurs, and the crack starts propagating upward from the bottom of the semi-rigid base. Numerical simulation results reveal the initiation of two symmetrical cracks propagating from the location of maximum principal stress at the bottom of the semi-rigid layer. As the crack propagates, the principal stress concentration at the crack tip gradually diminishes. However, when the crack extends to approximately 70% of the thickness of the semi-rigid base (corresponding to 4.11 million loading cycles), the crack propagation comes to a halt, despite the increase in load cycles. Based on the horizontal stress distribution map in the final state of crack propagation, it is observed that when the crack propagates to about 70% of the thickness of the semi-rigid base, the crack tip nears the compression zone above the semi-rigid layer. As per the maximum principal stress criterion, crack propagation can only occur under tensile stress conditions. At this length of crack extension, the crack reaches the compressive zone, rendering further extension unfeasible in Fig. 14.

Fig. 14.
figure 14

Fracture analysis based on maximum principal stress criterion.

Full size image

Based on the maximum principal strain criterion. If the crack resistance strength of the semi-rigid base is sufficiently high, it is possible for cracks to initiate due to the maximum principal strain criterion of the asphalt layer reaching the critical value. Figure 15(a) illustrates the distribution of the maximum principal strain in the asphalt layer. As evident from Fig. 15(b), when the number of loading cycles reaches 3.67 million, cracks form at the location of maximum principal strain concentration within the asphalt layer. With the progression of loading cycles, some cracks in the middle area of the tire extend to the surface of the asphalt layer, while cracks in other regions either remain unchanged or expand within a limited local area. Upon reaching 4.28 million loading cycles, the crack state stabilizes, and further increases in load cycles do not result in additional crack propagation.

Fig. 15.
figure 15

Fracture analysis based on maximum principal strain criterion.

Full size image

In the aforementioned scenarios, the cracking of the semi-rigid base occurs due to the maximum principal stress criterion, while the cracking of the asphalt layer bottom is caused by the maximum principal strain criterion. Given the distinct material properties of the semi-rigid base and asphalt surface, their respective cracking criteria also differ. The semi-rigid layer, characterized by significant stiffness and brittleness, is more susceptible to the maximum principal stress fracture criterion. Conversely, the asphalt layer, featuring lower stiffness and greater ductility, is more inclined towards the fracture criterion controlled by the maximum principal strain. Considering the material differences and potential for distinct cracking criteria, it is important to acknowledge that both cases may occur simultaneously in real pavement structures, leading to the initiation of cracks in both the asphalt layer and the semi-rigid base layer. This understanding contributes to a comprehensive analysis of crack formation mechanisms and aids in the design and maintenance of asphalt pavement structures to enhance their longevity and performance.

4.2 Cracking Analysis of Pavement Without Initial Cracks

When the initial crack is located at the bottom of the semi-rigid base, directly below the center of the two wheel loads, the crack propagates vertically upward along the load centerline. However, it ceases to expand when it approaches the compression zone. The crack propagation is thus confined within the semi-rigid layer, and after reaching a certain length (at 310,000 load cycles), further expansion is hindered due to the presence of compressive stress at the crack tip, in accordance with the maximum principal stress fracture criterion (Fig. 16).

Fig. 16.
figure 16

Crack propagation status and principal stress distribution (310,000 times).

Full size image

When the crack deviates 20 cm from the semi-rigid base below the wheel load center, similar to the previous case, the crack extends to a specific position and stops expanding. The crack’s propagation path is upward and tilts toward the center of the wheel load (Fig. 17).

Fig. 17.
figure 17

Crack propagation status and principal stress distribution (430,000 times).

Full size image

When the crack deviates 40 cm from the semi-rigid base below the wheel load center, the crack still extends to a certain position and halts expansion, following a path inclined towards the center of the wheel load. Compared with the 20 cm deviation case, the crack propagation path shows a greater inclination towards the wheel load center (Fig. 18).

Fig. 18.
figure 18

Crack propagation status and principal stress distribution (450,000 times).

Full size image

When the crack deviates 20 cm from the asphalt layer below the center of the wheel load, the crack expands upwards by two unit grid lengths but then stops as the load continues. Similar to the previous case, the crack cannot continue to expand when the crack tip approaches the compression zone (Fig. 19).

Fig. 19.
figure 19

Crack propagation status and principal stress distribution (220,000 times).

Full size image

When the crack deviates 40 cm from the asphalt layer below the center of the wheel load, the crack propagates upwards, inclined towards the center of the wheel load. Similar to the 20 cm deviation case, the crack reaches a certain extent and ceases further expansion, but the propagation distance is longer (Fig. 20).

Fig. 20.
figure 20

Crack propagation status and principal stress distribution (270,000 times).

Full size image

The following figure presents a statistical diagram of the initial crack position and the number of load cycles. It shows that pavement structures without initial cracks have a life 10–20 times longer than those with initial cracks. In the absence of initial cracks, the fatigue stage accounts for about 80% of the total life. With initial cracks, the farther the crack is from the centerline, the longer the fatigue life of the structure. The pavement structure’s life is greater when the initial crack is located at the bottom of the semi-rigid layer compared to being at the bottom of the asphalt layer (Fig. 21).

Fig. 21.
figure 21

Statistics of crack propagation.

Full size image

5 Conclusion

This paper simulates the fatigue fracture process of a pavement structure with hot in-place recycling material under wheel load and temperature. Key findings include:

  1. 1.

    Initial damage occurs at the center of the two wheels at the base, extending as the number of load actions increases, with the surface layer showing a V-shaped distribution pattern.

  2. 2.

    Maximum damage generally increases from top to bottom, influenced by hot in-place recycling layer properties. Increased RAP content raises cumulative damage in later stages, with minimal impact on other layers.

  3. 3.

    Base modulus significantly affects overall damage; increasing it reduces damage in each layer. Lowering the modulus of the intermediate layer increases deformation and damage. Overloading exacerbates damage, particularly in upper and middle layers.

  4. 4.

    Cracking patterns vary based on the presence of initial cracks. Without initial cracks, pavement structures have a lifespan 10–20 times longer. Crack propagation direction is influenced by the initial crack’s location and distance from the wheel load center.

  5. 5.

    Pavement structures without initial cracks have longer lifespans, with fatigue stages contributing around 80% of the total life. The location of the initial crack affects the structure’s life.

  6. 6.

    Regardless of the presence of an initial crack, fatigue load under wheel load prevents crack extension to the pavement surface due to force distribution.