Study on the Application and Deformation Characteristics of Construction Waste Recycled Materials in Highway Subgrade Engineering

Abstract

It is difficult to meet environmental requirements via the coarse treatment methods of landfilling and open-air storage of construction waste. At the same time, the consumption of building materials in highway engineering is enormous. Using construction waste as a filling material for proposed roads has become a research hotspot in recent years. This paper starts with basic performance tests of recycled construction waste materials, and then moves on to laboratory experiments conducted to obtain the road performance of the recycled materials, the testing of key indicators of post-construction filling quality of the roadbed, and analyses of the deformation pattern of roadbed filled with construction waste. Additionally, the ABAQUS finite element software was used to establish a numerical model for roadbed deformation and analyze the roadbed deformation under different compaction levels and vehicle load conditions. The experimental results show that the recycled material has a moisture content of 8.5%, water absorption of 11.73%, and an apparent density of 2.61 g/cm³, while the liquid limit of fine aggregates is 20% and the plasticity index is 5.4. Although the physical properties are slightly inferior to natural aggregates, its bearing ratio (25–55%) and low expansion characteristics meet the requirements for high-grade highway roadbed filling materials. The roadbed layer with a loose compaction of 250 mm, after eight passes of rolling, showed a settlement difference of less than 5 mm, with the loose compaction coefficient stabilizing between 1.15 and 1.20. Finite element simulations indicated that the total settlement of the roadbed stabilizes at 20–30 mm, and increasing the compaction level to 96% can reduce the settlement by 2–4%. Vehicle overload causes a positive correlation between the vertical displacement and shear stress in the base layer, suggesting the need to strengthen vehicle load control. The findings provide theoretical and technical support for the large-scale application of recycled construction waste materials in roadbed engineering.

1. Introduction

In the process of urbanization, the construction industry is developing rapidly, but the problem of construction waste is becoming increasingly serious. The demolition of old buildings and the construction of new buildings generate a large amount of construction waste with a wide range of types and complex compositions [1,2]. Traditional disposal methods, such as landfills and open piles, not only take up land resources but also may pollute groundwater and soil, posing a threat to human health. Meanwhile, a large amount of construction materials are required in the construction of highway networks, and over-exploitation of natural sand and gravel has led to resource scarcity and ecological damage [3,4]. Therefore, the application of construction waste as a roadbed filler in road projects can not only solve the problem of shortage of construction materials but also reduce environmental damage and realize the resource utilization of construction waste, which has significant social and economic benefits [5,6,7].

In recent years, many researchers have explored the feasibility of using various industrial waste and solid waste materials (such as fly ash, lime–fly ash construction waste mixtures, fiber-reinforced soil, river sediment, lignin, construction waste fine aggregates, red mud, windblown sand, nickel–iron slag, unburned coal gangue, recycled concrete aggregates, etc.) as road subgrade filling materials. Chunyu Cui et al. [8,9] found that fly ash and lime–fly ash construction waste mixtures exhibit good embankment filling performance. Yucheng Huang et al. [10] pointed out that polypropylene fibers and cement reinforcement can effectively improve the soil’s mechanical properties. Qingzhou Wang et al. [11] improved river sediment to prevent heavy metal contamination. Qiang Gao et al. [12] studied the effect of lignin and construction waste fine aggregates on the modified sediment’s properties. Ruifeng Chen et al. [13,14] found that an appropriate amount of red mud waste can improve the performance of loess subgrade. Xiangdong Zhang et al. [15] improved windblown sand subgrade performance using alkali-activated slag. Pingbao Yin et al. [16] studied the performance of nickel–iron slag as a subgrade material. Yan Feng et al. [17] evaluated the road performance of unburned coal gangue. Fan He et al. [16] and Yu Zhao et al. [18], respectively, assessed the performance of stabilized soil and microbiologically improved rock subgrades. Mohd Hafizan Md. Isa et al. [19] and Mukhtar Abukhettala Mamadou Fall [20] enhanced the mechanical properties of soil by adding waste materials. Javad Shamsi Sosahab et al. [21] and Cesar Hidalgo [22] studied the improvement of subgrade performance through the use of recycled concrete aggregates and chemical stabilizers, as well as demolition waste. Yassine Abriak [23] explored the feasibility of using dredged sediment and recycled concrete aggregates as subgrade materials. These studies indicate that with appropriate mixing ratios and modification treatments, waste materials can meet or exceed the minimum standards for road subgrade materials under specific conditions, providing a scientific basis for their resource utilization.

In recent years, researchers have deeply explored the potential applications of various waste materials, particularly construction waste, in road subgrade. Wuping Ran et al. [24] conducted experimental studies on the dynamic modulus of asphalt mixtures made from lime composite-modified oil sludge pyrolysis residues. Xiao Zhi et al. [25] found that cement-based additives can enhance the mechanical properties of cement-stabilized permeable recycled aggregates. Li Li-Hua et al. [26] verified the feasibility of using waste tires and construction waste for road subgrade reinforcement through physical model tests. Zhe Li et al. [27] demonstrated that the gradation and bearing ratio of recycled construction waste fillers meet the requirements for highway subgrade. Junhui Zhang et al. [28] introduced the construction procedure for demolition waste subgrades and analyzed their impact on subgrade performance and the variation in dynamic elastic modulus. Chenglin Shi et al. [29] studied the physical performance indicators of floor waste as subgrade fill. Haiying Wang et al. [30] developed a model to predict the settlement of demolition waste subgrades. Le Ding et al. [31] evaluated the saturation and leaching characteristics of recycled aggregates in road subgrades. Although progress has been made in these studies, the overall utilization rate of construction waste remains low, especially in high-grade highway projects. Therefore, further exploration of efficient ways to utilize construction waste in road subgrades is of significant importance.

Existing research focuses on testing the basic performance of recycled materials from construction waste, providing a preliminary theory for highway reuse. However, there is still a lack of systematic demonstration of the deformation mechanism of recycled roadbed under complex stress, the adaptability of layered filling, and long-term performance. Under complex stress, the deformation mechanism of recycled roadbed is complex. Due to its irregular shape, wide particle size distribution, and rough surface, its deformation behavior may be different from that of traditional materials. In-depth research is needed to reveal the deformation laws and factors [32]. The adaptability of layered filling is also critical and affects the performance of recycled roadbed. Process selection and parameter setting directly affect the compaction effect and overall performance. It is necessary to study the adaptability of different recycled materials, optimize parameters, and improve compaction and stability.

This paper takes a highway project as the research carrier and adopts the method of combining indoor tests, field tests, and numerical simulation to study the engineering performance of construction waste recycled materials and their application effect as roadbed fillers. At the same time, cement is added to the construction waste mixture to expand the scope of construction waste utilization. Finally, combined with numerical simulation and on-site monitoring, the deformation law of roadbeds filled with construction waste fillers is analyzed. The technical route diagram is shown in Figure 1.

2. Basic Properties of Recycled Materials from Construction Waste

2.1. Basic Properties of Recycled Materials

2.1.1. Water Absorption and Apparent Density

The water absorption test of construction waste is based on the “Highway Engineering Aggregate Test Code” (JTG E42-2024) [33]. First, take samples, take 5 kg of coarse aggregate and 1 kg of fine aggregate, and put them into water. After 24 h, take out the test, wipe the aggregate until no shiny water marks can be seen on the surface, which is the saturated surface dry state, and weigh its saturated surface dry mass (m₁), then put it in an oven at a temperature of 105 °C to dry, and weigh its dry mass (m₂) after cooling. Calculate the water absorption of coarse and fine aggregates according to the Formula (1).

$$\omega ={\displaystyle \frac{m_1-m_2}{m_1}}\times 100\%$$

(1)

In the formula, $\omega$ is the water absorption rate (%); $m_1$ is the saturated surface dry mass; and $m_2$ is the dry mass after drying.

Apparent density refers to the dry mass of material particles per unit volume (including the solid mineral components and closed pore volume of the material). This test is carried out in accordance with the “Highway Engineering Aggregate Test Code” (JTG E42-2005) [33]. Construction waste with a particle size of less than 4.75 mm is calibrated as fine aggregate, and construction waste with a particle size of more than 4.75 mm is calibrated as coarse aggregate.

Construction waste recycled materials refer to materials that are recycled after processing. This study uses construction waste raw materials generated by the renovation and demolition of brick–concrete structure houses. Before use, they need to be crushed and screened accordingly. The recycled materials are mainly divided into three groups of particle sizes: <4.75 mm, 4.75 mm~37.5 mm, and >37.5 mm. Water absorption and density are two fundamental properties of building materials. As a key component of recycled materials, bricks undergo changes in their engineering properties after absorbing water, which in turn reduces the overall strength of the subgrade. Therefore, it is important to understand the water absorption characteristics of recycled coarse aggregates and compare them with natural aggregates. The measured water absorption and density values for recycled coarse aggregates and natural aggregates are presented in

Table 1. Water absorption and density of recycled and natural aggregates.

Materials	Apparent Density (g/cm3)	Packing Density (g/cm3)	Water Absorption (%)	Natural Moisture Content (%)
Natural aggregate	2.77	1.62	5.90	1.90
Recycled aggregate	2.61	1.42	11.73	8.5

.

Due to the residual cement mortar on the surface of recycled coarse aggregates and the presence of numerous flaky brick fragments, the apparent volume of recycled aggregates is increased, resulting in a lower apparent density compared to natural aggregates. Specifically, the bulk density of recycled aggregates is 12.3% lower than that of natural aggregates. As for water absorption, recycled coarse aggregates exhibit a 98% higher absorption rate (5.90%) than natural aggregates. This is primarily due to the lightweight, porous nature of the brick fragments in the recycled material, as well as the low density and high porosity of the cement mortar adhered to the surface, which leads to an increased water absorption rate.

2.1.2. Crushing Value and Content of Flaky Particles

Construction waste, when exposed to the natural environment over an extended period, tends to develop microcracks and experience aggregate loosening, leading to a reduction in the strength of recycled materials compared to natural aggregates. The strength of recycled materials can be quantified by the crushing value, where a lower crushing value indicates higher aggregate quality.

The experiment was conducted in accordance with the “Aggregate Testing Method for Highway Engineering (JTG E42-2005)” [33]. The particle size was controlled between 9.5 mm and 13.2 mm. A predetermined mass (m₀) of recycled material, in its air-dried state, was placed in a cylindrical container and positioned on a compression machine. A load of 400 kN was applied and maintained for 5 s before unloading. Subsequently, finer material crushed during the test was separated using a sieve with a pore size of 2.36 mm, and the remaining sample weight (m₁) was measured. Two parallel tests were performed, and the crushing value (CV) of the material was calculated using the following Formula (2).

$$\delta ={\displaystyle \frac{m_0-m_1}{m_0}}\times 100\%$$

(2)

The experiment shows that the crushing value of recycled construction waste materials is 15.3%, slightly higher than the 10.7% of natural aggregates, but still suitable for the subgrade backfill. Flaky particles, due to their shape, are more easily crushed, which increases the void ratio and affects the subgrade’s water stability and fatigue resistance. Therefore, it is necessary to test the content of flaky particles in both recycled and natural aggregates. Following the JTG E42-2005 standard [33], the content of flaky particles in recycled materials was measured using a caliper method, identified with a calibration device, and calculated using Formula (3). The results are shown in

Table 2. Recycled aggregate and natural crushed stone needle-like particle content.

Aggregate Type	Recycled Material (mm)		Natural Aggregate (mm)
	9.5~19.0	19.0~26.5	9.5~19.0	19.0~26.5
Recycled aggregate	16.1	16.0	10.6	9.8

.

$$\mathrm{Content} \mathrm{of} \mathrm{needle} \mathrm{and} \mathrm{flake} \mathrm{particles}={\displaystyle \frac{m_1}{m_0}}\times 100\%$$

(3)

The content of flaky particles in the recycled material is 16%, which is 6% higher than the 10% in natural aggregates. However, it still meets the requirement in the “Inorganic Mixture of Recycled Construction Waste for Road Use (JC/T 2281-2014)” [34], which stipulates that the content of flaky particles in recycled coarse aggregates should be less than 20%.

2.1.3. Recycled Fine Aggregate Properties

In the construction waste mixture, the fines can fill the gaps between the coarse materials and play a connecting role and stabilizing role. To understand the characteristics of <4.75 mm recycled fine aggregate in construction waste, the fines obtained from sieving were subjected to a bounded moisture content test, and the specific test steps are shown in Figure 2.

The liquid limit of the fine aggregate was determined to be 25.4% based on the relationship between water content and cone depth, and the plastic limit was determined to be 20% and the plasticity index to be 5.4 using polynomial curves, which meets the provisions of the “Highway Roadbed Design Specification JTG D30-2015” [35] that the liquid limit is less than 50% and the plasticity index is small 26. This indicates that the recycled fine aggregate is mainly powder and sand particles, with less content of clay particles, and belongs to sandy soil, which is characterized by high permeability and low cohesion. When the content of fine aggregate increases, the structure of the mixture will change from a void structure to a dense structure, so the content of recycled fine aggregate with a particle size <4.75 mm has a key role in the construction waste mixture system.

2.2. Roadworthiness of Recycled Materials

2.2.1. Grain Gradation

Two kinds of on-site randomly selected raw materials were used for the backfill according to JTG E40-2007 [36] in “T0115-1993 sieving method” for the sieving test (test results shown in Figure 3).

The maximum particle size of the first type of raw material is 60 mm, primarily used for the subgrade backfill, while the maximum particle size of the second type is 150 mm, intended for embankment backfill. The particle gradation is assessed using the uniformity coefficient (C_u) and the curvature coefficient (C_c). The uniformity coefficient indicates the uniformity of the particle distribution, and the curvature coefficient is used to describe the shape of the cumulative curve. The calculation formulas are provided in Equations (4) and (5). For materials to be used as subgrade fill, the following criteria must be met: for the subgrade C_u ≥ 10, for the embankment C_u ≥ 5, and the curvature coefficient C_c should be within the range of 1 to 3.

$$C_u={\displaystyle \frac{d_{60}}{d_{10}}}$$

(4)

$$C_c={\displaystyle \frac{d_{30}^2}{d_{10}\times d_{60}}}$$

(5)

In the formula, d₁₀, d₃₀, and d₆₀ are the grain sizes corresponding to 10%, 30%, and 60% of the longitudinal coordinates on the grading curve, which are obtained by interpolation. The unevenness coefficients of the embankment and roadbed are 40.6 and 35, respectively, and the curvature coefficients are 1.61 and 1.06, respectively, which belong to the well-graded gravels, and a high degree of densification can be achieved during on-site rolling.

2.2.2. Maximum Dry Density and Optimum Moisture Content

(1)

Experimental design

The indoor compaction test is limited to a maximum particle size of 40 mm, while the construction waste filler contains particles exceeding this size, which affects the maximum dry density. Therefore, a compaction test for the designed ratio of recycled coarse and fine aggregates was conducted to establish the relationship curve between particle size and maximum dry density. During on-site testing, the content of particles between 4.75 mm and 40 mm was measured, and the results were corrected to resolve the compaction detection issue. Additionally, the compaction characteristics of the construction waste mixture under different ratios were obtained.

With reference to the relevant contents in “Highway Geotechnical Test Specification JTG E40-2007” [36] and “Technical Specification for Roadbed Construction with Recycled Construction Waste Materials DB61/T 1149-2018” [37], seven groups of proportioning schemes for recycled coarse aggregate and recycled fine aggregate are prepared, as shown in

Table 3. Compaction test mix design program.

Component Content (%)	Specimen Number
	1	2	3	4	5	6	7
Recycled Aggregate	20	30	40	50	60	70	80
Recycled fine aggregates	80	70	60	50	40	30	20

.

(2)

Experimental process

A multi-function electric compactor was selected for heavy-duty compaction, with the following main parameters: diameter of 15.2 cm, height of 17 cm, hammer weight of 4.5 kg, hammer bottom diameter of 5 cm, hammer drop distance of 45 cm.

The water content of the 5 kg mixture was configured at 10%, 12%, 14%, 16%, and 18%. For the coarse aggregate with strong water absorption, water was first sprayed on the fine aggregate mixture, followed by simmering the coarse aggregate and static closure. The mixture was then placed into a compacting cylinder in three layers, each layer being compacted 98 times with 1.7 kg of force. The layers were brushed, and the test procedure is shown in Figure 4. After compaction, the sample was weighed, and the center of the soil sample was dried for more than 8 h to measure its moisture content. The dry density was calculated, and a dry density–moisture content curve was plotted. The dry density and water content at the peak of the curve represent the maximum dry density and optimum water content of the mixture.

(3)

Test results and analysis

During the compaction process, the change in water content will affect the compaction of the specimen. Take 20% coarse aggregate content of compaction as an example, as shown in Figure 5. When the water content is 10%, the specimen has a fracture in the middle after demolding, and there are large pieces of granules falling around, indicating poor molding; when the water content is 12%, the water content is still low, and the specimen drops small pieces of soil at the corners after demolding, which indicates that the particles are not fully compacted, and there are gaps between the particles; when the water content is 14%, the specimen after demolding is well formed with a smooth surface, which indicates that this time the water content is optimal, and the specimen is the most effective in compacting; when the water content is 16%, the specimen is compacted with the best effect, but the specimen is easy to deform after demolding, and there is water seeping out from the bottom of the specimen; and when the water content is 18%, the water content is too high, which leads to serious deformation of the specimen, a slumping phenomenon, and more water seeping out from the bottom. From the above analysis, it can be seen that the construction waste mixture is sensitive to the water content, and the water content of the filler should be strictly controlled in the construction process to achieve the best compaction effect.

The results of the compaction tests have been summarized according to the above steps to obtain the relationship between the dry density and moisture content of coarse and fine aggregates for different mix ratios, as shown in Figure 6. The compaction curves for the mixtures with different ratios exhibit a typical parabolic shape, with the dry density initially increasing and then decreasing as the moisture content increases, which is similar to the compaction characteristics of cohesive soils. The optimal moisture content of the samples ranges from 14.0% to 16.6%, and the maximum dry density ranges from 1.76 g/cm³ to 1.84 g/cm³. A regression analysis of the coarse aggregate content and maximum dry density yields the relationship curve Y = 1.73821 + 0.00132X, with a correlation coefficient R² of 0.98, where X represents the coarse aggregate content. During on-site testing, the formula can be used by substituting different coarse aggregate contents to determine the corresponding maximum dry density of the fill, ensuring the accuracy of the compaction detection on site.

To analyze the effect of different coarse and fine content on the dry density and moisture content, the change curves of the mixes with different coarse and fine content were plotted as shown in Figure 7.

As shown in Figure 7, the content of coarse and fine aggregates in the construction waste mixture has a certain impact on compaction performance. As the coarse aggregate content increases from 20% to 80%, the maximum dry density shows a linear increase, indicating that larger particles gradually provide support to the overall structure. When the coarse aggregate content exceeds 70%, the sample exhibits higher density, with the fine aggregates filling the gaps between the coarse aggregates. However, excessive coarse aggregate content leads to a skeletal-void structure. When the ratio of coarse to fine aggregates is the same, the sample’s optimal moisture content is minimized, meaning the least amount of water is required to achieve the corresponding maximum dry density. When the coarse aggregate content exceeds 70%, the optimal moisture content decreases, likely because the coarse aggregates gradually become the dominant component, absorbing some of the moisture. As the fine aggregate content increases, it provides lubrication for the movement of the coarse aggregates. With the addition of water, the overall sample becomes more prone to deformation, resulting in a decreasing trend in the maximum dry density.

2.2.3. California Bearing Ratio

The California Bearing Ratio (CBR) is an important index to characterize the strength of roadbed fill, which is of great significance for the design and construction of roadbed projects [28]. The CBR value of the material is calculated by the relative percentage of the unit pressure of the specimen at a penetration of 2.5 mm or 5.0 mm to the standard unit pressure, and the standard unit pressures corresponding to a penetration of 2.5 mm and 5.0 mm are 7000 kPa and 10,500 kPa, respectively. Each specimen was compacted 30, 50, and 98 times to study the effect of compaction work on the bearing ratio of the specimens.

(1)

Experimental process

A piece of filter paper is placed on the top surface of the prepared sample, and a porous plate with an adjustable rod is installed on top, with four load plates placed on the porous plate. The sample cylinder and porous plate are sequentially placed into a water tank, and a dial gauge is installed to record the initial reading. Water is added to the tank to ensure the water level is about 25 mm above the top surface of the sample. During soaking, care should be taken not to touch the sample arbitrarily to avoid affecting the accuracy of the results. After the sample has been soaked for four days and nights, the final reading of the dial gauge is recorded, and the sample’s expansion rate is calculated. The experiment is shown in Figure 8.

The sample is removed from the water tank, and the water on the top surface of the sample is poured off. The sample is then allowed to drain for 15 min. Afterward, the additional load, porous plate, and filter paper are removed. The sample is placed on the lift platform of the pavement material strength testing apparatus. The lift platform is adjusted so that the penetration rod makes full contact with the top surface of the sample. Four load plates are placed around the top of the sample, and the testing apparatus is used to slowly apply pressure, causing the penetration rod to penetrate the sample at a speed of 1 mm/min to 1.25 mm/min.

(2)

Experimental results and analysis

The test results of this experiment are based on the bearing ratio when the penetration is 5.0 mm. The bearing ratios of the samples for different compaction cycles are shown in

Table 4. California Bearing Ratio test results.

Typology	Number of Hits	Expansion (%)	Load Ratio Value (%)
Embankment fill	30	0.004	25.11
	50	0.025	31.52
	98	0.091	44.10
Roadbed Fill	30	0.098	50.23
	50	0.033	53.71
	98	0.026	54.15

Note: The compacting work corresponding to 30, 50, and 98 strikes is 819.6 kJ/m³, 1314.9 kJ/m³, and 2677.2 kJ/m³, respectively.

. According to the specifications and standards for highway subgrade fill materials, when construction waste is used as the subgrade fill, the CBR should be ≥8%; when used as the embankment fill, the CBR should be ≥4%; and when used as the lower embankment fill, the CBR should be ≥3%.

As can be seen from Table 4, the bearing ratio of the specimen after water immersion is 25~55%, which fully meets the strength standard of highway roadbed filler. Therefore, the recycled construction waste materials studied in this paper have the applicability of highway roadbed filling and can be directly used as roadbed fillers for backfilling. The expansion of the specimen in the process of water immersion is very small, indicating that the recycled material is less affected by water, which shows that the construction waste has good water resistance. With the increase in the number of compacting, the bearing ratio showed an increasing trend. For the embankment fill, when the number of compacting times was increased to 98 times, the value of the bearing ratio increased by 76% compared with the number of compacting times of 30 times, indicating that the increase in the number of compacting times has a positive effect on the strength of the mixture.

3. Research on the Application of Recycled Construction Waste Filler in Roadbed

3.1. Preparatory Phase

In the field test, the recycled construction waste materials were used to fill the roadbed and embankment, respectively. The optimal construction parameters and construction technology of the test section construction waste filling material were set for roadbed backfilling. Finally, the key indicators of the filling quality of the roadbed after construction were tested. Initially, topsoil was removed from the highway roadbed. Figure 9 shows the test section used for construction waste to fill the roadbed. Next, the construction waste was screened and the material with a particle size of less than 6 cm was used for backfilling the roadbed, while the material with a particle size of more than 6 cm and less than 15 cm was used for backfilling the embankment. In addition, the operation area needs to be prepared with equipment including excavators, sprinklers, dump trucks, bulldozers, graders, foot mills, and vibratory rollers.

3.2. Construction Techniques

After completing the preparation work, the subgrade filling begins, which is divided into two parts: the subgrade and the embankment. The subgrade construction consists of six steps: spreading, initial leveling, first watering, sheep-foot roller compaction, second watering, and vibrating roller finishing (Figure 10). First, the fill material is evenly spread over the soil base of the test section. The loose thickness for the lower embankment and upper embankment is 350 mm, while the loose thickness for the subgrade is 250 mm. After the first watering, a 22t sheep-foot roller is used to compact the material for four passes, with two weak vibration passes and two strong vibration passes, at a speed of 3 km/h. After the second watering, a 26t vibrating roller is used to compact the material for 4 to 8 passes at a speed of 4 km/h, with four passes for the lower embankment, six passes for the upper embankment, and eight passes for the subgrade. During the compaction process, the elevation is observed using a level.

In order to analyze the applicability of recycled aggregate from construction waste for roadbeds, it is necessary to measure the quality of the roadbed. The compaction of the roadbed is an important indicator for measuring the quality of the roadbed. Therefore, two compaction monitoring points are set at the 100 m test section during roadbed construction (as shown in Figure 11).

3.3. Post-Construction Filling Quality Control

3.3.1. Dynamic Deformation Modulus and Compaction

After the subgrade is completed, its compaction is tested. In this study, the dynamic deformation modulus (EVD) of the subgrade surface is detected using a portable falling weight deflectometer (FWD), which reflects the dynamic deflection and deformation modulus of the structural layer. The dynamic deformation modulus is tested according to the JJG198-2008 [38] standard. The detection process can be divided into two stages: the free fall of the hammer and the interaction with the spring. Under the impact load, the bearing plate produces a certain vertical displacement, and pressure and displacement sensors record the load and displacement. The rebound modulus of the subgrade is then determined based on the peak values of pressure and displacement. The related formula is shown in Equation (6):

$$EVD={\displaystyle \frac{\lambda \times \gamma \times \sigma }{S}}$$

(6)

where EVD is dynamic deformation modulus (MPa). $\lambda$ is the influence coefficient of the shape of the load-bearing plate, which is taken as 1.5; $\gamma$ is the radius of the load-bearing plate (mm), which is taken as 150 mm; $\sigma$ is the maximum dynamic stress of the roadbed surface (MPa), which is taken as 0.1 MPa; $S$ is the vertical settlement value of the load-bearing plate (mm).

According to the previously measured number of compaction passes and loose thickness, the construction waste filler is backfilled into different parts of the subgrade to test its post-construction filling quality. The detection results for the subgrade compaction and dynamic deformation modulus are shown in

Table 5. Compaction and dynamic deformation modulus test results.

Testing Location	Number of Compaction Passes	Number of Testing Points	Compaction Test (%)			EVD (MPa)
			Minimum Value	Maximum Value	Average Value	Minimum Value	Maximum Value	Average Value
Lower Embankment	Sheep-foot roller 4 passes + vibrating roller 4 passes	10	93.0	95.3	93.9	32	37	34.2
Upper Embankment	Sheep-foot roller 4 passes + vibrating roller 6 passes	10	94.3	95.9	94.9	45	50	47.9
Subgrade	Sheep-foot roller 4 passes + vibrating roller 8 passes	10	96.1	97.4	96.6	45	52	49.9

.

The construction waste filler is backfilled into different parts of the subgrade according to the number of compaction passes and loose thickness, and its post-construction filling quality is tested. Table 5 shows the test results for the subgrade compaction and dynamic deformation modulus. The EVD of the embankment ranges from 32 to 50 MPa, exceeding the design value of 30 MPa; the EVD of the subgrade ranges from 45 to 52 MPa, exceeding the design value of 40 MPa. The results of the compaction and dynamic deformation modulus tests meet the required design values and even exceed them, indicating that the construction machinery, parameters, and processes determined for the test section are reasonable and have practical guiding significance.

Based on the analysis of Table 5, it can be seen that the test results for compaction and dynamic deformation modulus are relatively consistent. Specifically, the compaction of the subgrade is higher than that of the embankment, and the dynamic deformation modulus of the subgrade is higher than that of the embankment. On this basis, the relationship between traditional testing methods and rapid testing methods is analyzed. Through the field test results from actual engineering applications, as shown in Figure 12, the relationship between compaction (K) and dynamic deformation modulus values is established.

The analysis results in Figure 12 show that for the construction waste filler with coarse aggregate contents of 30%, 40%, 60%, and 70%, the relationship between compaction and dynamic deformation modulus at their respective optimal moisture contents is determined. Through linear regression, the relationship between the compaction and dynamic deformation modulus is derived, as shown in

Table 6. Regression formula between compaction and dynamic deformation modulus.

Construction Waste Filler Type	Relationship Formula	R2
30% Coarse Aggregate	EVD = 15.53e1.501K	0.962
40% Coarse Aggregate	EVD = 9.272e1.825K	0.941
60% Coarse Aggregate	EVD = 6.671e2.245K	0.984
70% Coarse Aggregate	EVD = 2.781e3.312K	0.919

. For construction waste fillers with different coarse aggregate contents, the compaction and dynamic deformation modulus exhibit a consistent exponential relationship. The minimum value of the R² of the regression formula is 0.919, and the maximum value reaches 0.984, which is close to 1. This indicates that there is a strong correlation between the compaction and dynamic deformation modulus for different coarse aggregate contents, allowing the portable deflection measurement method to replace the sand cone method for the large-scale testing of the filling quality of construction waste subgrades.

3.3.2. Static Flexural Modulus

The bending subsidence test is performed in accordance with the standard “Highway Roadbed Pavement Field Test Specification JTG E60-2008” [39]. The specific measurement steps are as follows:

(1)

Measure the grounded area of tires by using duplicate paper or square paper, accurate to 0.1 cm². Carry out sensitivity tests on the percentage meter. The temperature of the road surface is measured using a thermometer to obtain the average temperature in the last 5 days.

(2)

Set measurement points on the roadbed or road surface to be measured, and mark and number the wheel tracks on the roadway to be measured with white paint or chalk. A 300 m long road section is selected on site for bending settlement value detection, and one detection point is set every 10~40 m at the center of the line. Detection points are laid out as shown in Figure 13.

(3)

Insert the deflection gauge into the gap between the rear wheels of the deflection vehicle, ensuring that it is aligned with the direction of the vehicle. Position the deflection gauge at the measurement point, which is 3 to 5 m in front of the center of the wheel gap, while ensuring that the horizontal beam arm does not contact the tire. A dial gauge is installed on the measurement rod of the deflection gauge, and the gauge is zeroed. Gently tap the pressure gauge with a finger to check if the dial gauge is stable.

(4)

The inspector blows the whistle, commanding the vehicle to move forward slowly (with an optimal speed of around 5 km/h). As the road deforms, the dial gauge continues to rotate forward. When the needle reaches its maximum deflection, the initial reading (L₁) is quickly recorded. The vehicle continues to move, and the needle begins to rotate in reverse. When the vehicle has traveled 3 m, the whistle is blown or a red flag is waved to signal the vehicle to stop. After the needle stabilizes, the final reading (L₂) is recorded. The rebound deflection value at the measurement point is calculated using Equation (7).

$$L_T=\left(L_1-L_2\right)\times 2$$

(7)

In the formula, L_T is the rebound value (0.01 mm) at a pavement temperature of T; L₁ is the initial reading (0.01 mm), which corresponds to the maximum deflection when the needle rotates; and L₂ is the final reading (0.01 mm), which is the stable deflection after the needle rotates in reverse.

The deflection values at the measurement points were tested according to the steps outlined above, and the resulting deflection values are shown in

Table 7. Deflection value (0.01 mm) results.

Deflection Value (0.01 mm)	Measurement Point Number
	1	2	3	4	5	6	7	8	9	10
Left Wheel	210	200	140	180	190	230	90	160	40	200
	210	200	220	230	160	200	180	174	180	170
Right Wheel	30	100	52	52	120	140	90	70	80	100
	120	110	90	40	160	60	100	130	68	80

Average (0.01 mm): 102.5; standard deviation (0.01 mm)²: 61.69; deflection representative value (0.01mm): 228.7.

. The measured deflection value should not exceed the design deflection value of 232 mm (0.01 mm). The average value and standard deviation were calculated using Formula (8), with the average value of this test section being 102.5 mm (0.01 mm). The standard deviation was calculated using Formula (9) and found to be 61.69 (0.01 mm)². The representative deflection value was calculated using Formula (10) and found to be 228.7 mm. The results are summarized in Table 7. From Table 7, it can be seen that the deflection values of the left wheel are all greater than those of the right wheel, and there are significant differences between some of the results, with a maximum difference of 180 (0.01 mm). Points with large differences in deflection values should be discarded.

$$\overline{\mathrm{l}} ={\displaystyle \frac{{\displaystyle \sum {\mathrm{l}}_{\mathrm{i}}}}{\mathrm{n}}}$$

(8)

$$S=\sqrt{{\displaystyle \frac{{\displaystyle \sum {\left({\mathrm{l}}_{\mathrm{i}}-\mathrm{l}\right)}^2}}{\mathrm{n}-1}}}$$

(9)

$${\mathrm{l}}_{\mathrm{r}}=\overline{\mathrm{l}}+Z_{\mathrm{a}}S$$

(10)

In the formula, l_r is the calculated representative deflection value; S is the standard deviation; and $Z_a$ is the reliability coefficient, which is generally taken as 2 for highways.

The representative deflection value is 228.7 (0.01 mm), which is less than the design deflection value of 232 (0.01 mm), indicating that the deflection value of the subgrade meets the design requirements. The backfilled subgrade has a high strength and bearing capacity. By substituting the representative deflection value (l_r) into Formula (11), the rebound modulus of the subgrade is calculated to be 40.7 MPa.

$$E_1={\displaystyle \frac{2pr}{l_r}}\left(1-{\mu }^2\right){\alpha }_0$$

(11)

In the formula, E₁ is the rebound modulus; p is the measured average vertical unit pressure of the wheel, with the standard load being 0.7 MPa; r is the radius of the circle, with r = 106.5 mm under the standard load; μ is the Poisson’s ratio, taken as 0.35; and α₀ is the uniform body deflection coefficient, taken as 0.712.

4. Analysis on Deformation Law of Roadbed Backfilled with Construction Waste

4.1. Analysis of Roadbed Settlement and Deformation

4.1.1. On-Site Monitoring of Roadbed Deformation

To better understand the deformation behavior of construction waste used for embankment filling, deformation monitoring of the embankment section is conducted. A third-party service provider was commissioned to establish an automated settlement monitoring system using a static leveling instrument and wireless transmission system. Monitoring points are set within the embankment, and reference points are established outside the embankment to monitor the settlement and deformation of the construction waste embankment section during both the construction period and long-term period.

The monitoring instruments are set up in two layers according to the longitudinal section, with one layer buried at the base of the embankment and the other at the top of the embankment. In the horizontal direction, the monitoring includes the left and right shoulders, left and right lanes, and the center of the road. The monitoring points on the embankment section, along with the reference points, collectively monitor the deformation of the embankment, as shown in Figure 14. The settlement of the embankment itself is calculated by subtracting the settlement at the base of the embankment from the settlement at the top of the embankment.

4.1.2. Monitoring Data Analysis

The monitoring duration lasted for 480 days, and the monitoring process was synchronized with the subgrade embankment construction. Due to time constraints, the monitoring did not include deformations or settlements during the later operational phase of the highway. Therefore, the monitoring data primarily reflects the deformations and settlements caused by the embankment construction during the construction period. The monitoring curve is shown in Figure 15.

The following can be observed in the curve of settlement versus time: The relationship curve between settlement and time shows that after the completion of the roadbed filling, the total settlement is stable at 20~30 mm, the maximum settlement of the right roadbed is 29 mm, the settlement rate is compliant, and the compaction quality is qualified; the settlement data of the right shoulder fluctuates after 300 d, and the slight floating is caused by the micro deformation of the plastic sleeve, which does not affect the overall observation. The settlement suddenly changes at 100 d and 300 d, which may be due to the additional load or humidity changes. The data of the right lane are eliminated due to system problems; Figure 15 shows the change in settlement of the cross-section of the construction waste roadbed over time. During the construction period, the settlement increases with the increase in filling height, and the settlement trends of each point in the cross-section are different. The settlement of the left and right shoulders is large in the early stage, and the left shoulder is stable in the later stage. The settlement of the center line of the roadbed is the largest at 250 d, and the maximum settlement difference compared with other points is 8.34 mm.

4.2. Finite Element Model Establishment

4.2.1. Finite Element Model of Roadbed

Using the ABAQUS (2018) finite element analysis software package, according to the actual situation of the highway, the three-dimensional finite element model of the construction waste roadbed was established with the width of the top surface of the roadbed as 34.5 m, the height of the roadbed as 6m, and the side slope of the roadbed as 1:1.5, and the roadbed mainly includes two parts of the embankment and the roadbed.

Based on ABAQUS finite element simulation, this study uses hexahedral structured grids (C3D8R units) to finely model the construction waste roadbed. The main grid size of the roadbed is 0.5 m (horizontal) × 0.15 m (vertical), the top 0.5 m range is encrypted to 0.1 m, the sub-base and base use 0.1 m layered grids, and the vehicle load action area is further encrypted to 0.05 m. The material constitutive model uses the modified Drucker–Prager model to describe the construction waste filler. The elastic moduli corresponding to compaction degrees of 90%, 93%, and 96% are 80 MPa, 120 MPa, and 180 MPa, respectively. The sub-base and base use a linear elastic model. The reliability of the model is ensured by grid sensitivity verification (error < 3%) and measured settlement comparison (RMSE < 5 mm), and the influence of compaction degree and vehicle load (BZZ-100 standard axle load and overload conditions) on roadbed settlement and shear stress is studied to provide a basis for engineering optimization. Considering the influence of the boundary effect, the width of the original foundation is taken as 150 m, the height is taken as 18 m, and the subgrade is a powdery clay layer, as shown in Figure 16.

During the calculation process, the pavement structure and roadbed materials are regarded as uniform and homogeneous materials, the contact surfaces of the soil layers are completely continuous, and the pressure on each contact surface is evenly distributed; at the same time, horizontal and vertical displacement constraints are imposed on the bottom of the foundation, the foundation and the upper part of the roadbed are free interfaces, and horizontal displacement constraints are set on both sides of the foundation. To match the simulation with the actual roadbed, the boundary conditions of the model are set as follows: the displacement in the X direction is constrained at the left and right ends of the model; the displacement at the bottom surface of the model foundation, i.e., in the Y direction, is fully constrained; and the displacement and forwarding of the cross-section in the Z direction of the model are fully constrained, as shown in Figure 17.

Based on the previous research results, it is concluded that the strength of the construction waste filler can meet the filling requirements of highway roadbeds and has a good working performance. To understand the settlement and deformation characteristics of the construction waste roadbed, based on the basic principles of the ABAQUS finite element analysis and practical engineering applications, the following studies are proposed:

(1)

Deformation simulation during roadbed filling. The simulation results are compared with the actual monitoring settlement results to verify the validity of the model.

(2)

Deformation simulation of different roadbed compaction. The compaction degree is taken as 90%, 93%, and 96%, respectively, to analyze the effect of compaction degree on the settlement and deformation of roadbeds.

(3)

Simulation of different vehicle loads. Based on the finite element model of the roadbed established during the construction period, 0.2 m subgrade and 0.38 m base layer are filled on the roadbed surface according to the design requirements, and the effects of different vehicle loads on vertical displacement and shear stress are studied. Regarding the load form, the vehicle load in this article adopts a steady-state sinusoidal fluctuation load, which can be calculated by using Formula (12):

$$P(t)=P_0+P\sin (\omega t)$$

(12)

In the formula, $P_0$ is the static load of the vehicle, kN; and $\omega$ is the circular frequency of vibration.

According to the actual engineering design data of the roadbed, design specifications, and field test results, the calculation parameters of the analytical model are determined as shown in

Table 8. Simulation calculation parameters of the subgrade.

Simulation Scenario	Materials	Thickness (m)	Density (kg/m3)	Elastic Modulus (MPa)	Poisson’s Ratio	Cohesion (kPa)	Internal Friction Angle (°)
Layered filling simulation	Silty clay (original foundation)	18	1870	15.0	0.35	28.0	18.0
	Roadbed filler	0.80	1810	40.7	0.22	25.0	35.6
	Embankment filling	5.20	1790	36.5	0.22	23.0	30.2
Simulation of different compaction levels	Roadbed filler (compaction degree 96%)	6	1810	0.22	25.0	25.0	35.6
	Roadbed filler (compaction degree 93%)	6	1790	0.22	23.0	23.0	30.2
	Roadbed filler (compaction degree 90%)	6	1720	0.22	20.0	20.0	28.7
Simulation under vehicle load	Cement stabilized base	0.38	2400	800.0	0.25	/	40.0
	Cement stabilized base	0.20	2400	700.0	0.25	/	35.0

.

4.2.2. Constitutive Relations

The Mohr–Coulomb model is an ontological theory that has been certified by a large number of tests and in the field, which can better simulate the shear damage suffered by the roadbed and has been widely used in simulating the roadbed filling problem [40]. Therefore, in this paper, the Mohr–Coulomb model is used as the ontological relationship to simulate the roadbed filled with the construction waste filler. The foundation, roadbed, and roadbed are regarded as elastic-plastic bodies obeying the Mohr–Coulomb model yield criterion, and their yield surface functions are as follows:

$$F=R_{mc}q-p\tan \phi -c=0$$

(13)

In the equation, $\phi$ is the angle of inclination of the Mohr–Coulomb yield surface on the $q-p$ stress plane, which is called the friction angle of the material, $0°\le \phi \le 90°$; c is the cohesion of the material; and $R_{mc}\left(\Theta ,\phi \right)$ is calculated according to Equation (14), which controls the shape of the yield surface in the π plane.

$$R_{mc}={\displaystyle \frac{1}{\sqrt{3}\cos \phi }}\sin \left(\Theta +{\displaystyle \frac{\pi }{3}}\right)+{\displaystyle \frac{1}{3}}\cos \left(\Theta +{\displaystyle \frac{\pi }{3}}\right)\tan \phi$$

(14)

In the equation, $\Theta$ is the polar deflection angle, defined as $\cos \left(3\Theta \right)={\displaystyle \frac{r^3}{q^3}}$, and r is the third bias stress invariant J₃.

4.3. Analysis of Numerical Simulation Results

4.3.1. Simulation of Roadbed Layer Filling

In the simulation process, the stress and strain of the roadbed will be affected by various factors; therefore, without affecting the modeling accuracy, combined with the Mohr–Coulomb yield criterion and the basic principle of finite element, the simulation makes the following assumptions:

(1)

The roadbed fill is an isotropic continuous medium and an ideal elastic-plastic body, and the contact between each soil layer is continuous;

(2)

When the roadbed is filled in layers, the deformation of the roadbed is mainly generated under the influence of its self-weight as the filling height changes;

(3)

The influence of environmental factors such as temperature and humidity on the deformation of the roadbed can be neglected.

The roadbed model is established by the ABAQUS finite element software package and then backfilled according to the roadbed and roadbed, respectively. The simulation of the backfilling process can be regarded as a graded stacking process; i.e., when backfilling the embankment, the bed of the road does not exist. The process is mainly realized by using the birth and death unit in ABAQUS, and the calculation results are shown in Figure 18.

The simulation results show that the maximum settlement of the backfill of construction waste recycled materials occurs in the middle and lower parts, and the load increases due to the increase in filling height. There is no horizontal displacement at the center of the bottom of the roadbed, which mainly occurs at the foot of the slope and extends to the surrounding areas. Therefore, in order to prevent smooth water movement, the roadbed slope should not be too large.

Table 9. Comparison of simulation results and field measurements.

Monitoring Points	Left Shoulder	Left Driving Lane	Subgrade Centerline	Right Driving Lane	Right Shoulder
Simulation results (mm)	16.66	30.66	31.66	30.66	16.66
Monitoring results (mm)	24.29	23.46	25.00	/	29.00
Residual (mm)	7.63	7.20	6.72	/	12.34

compares the roadbed numerical simulation with the on-site monitoring data. Although there is a gap, the deviation is small. Except for the right lane affected by system problems, the other errors are within 13 mm, and the settlement results are consistent. The selected parameters can reasonably reflect the roadbed settlement, and it is feasible to use the constitutive model for extended calculation.

4.3.2. Simulation of Different Roadbed Compaction Degrees

Different degrees of compaction show different construction quality, which will have different effects on the use characteristics of the roadbed. The low strength of compaction construction will lead to the actual loading capacity of the roadbed failing to meet the loading demand of the highway. Therefore, only by compacting the roadbed can the strength, stiffness, and smoothness of the roadbed be ensured, thus improving the service life and service life of the roadbed [41]. In order to obtain the effect of compaction on the deformation of the roadbed, different calculation parameters are substituted into the model to calculate, and the cloud diagram of settlement deformation results is shown in Figure 19.

Combined with Figure 19 and

Table 10. Settlement statistics of key points under different compaction degrees.

Settlement Point Location	Settlement Corresponding to Different Compaction Degrees (mm)
	90%	93%	96%
Maximum settlement value of the model	100.1	99.27	98.52
Maximum horizontal displacement of the model	25.18	25.10	25.05
Subgrade surface center point settlement value	18.15	17.69	17.46
Settlement value at the foot of the roadbed	13.38	13.52	13.61
Horizontal displacement value at the foot of the embankment	13.22	13.29	13.31

above, it can be seen that with the increase in compaction degree, the corresponding parameters in the model also become larger, and then the settlement of the center point of the model surface becomes smaller. This is because the increase in compaction decreases the instantaneous settlement of the roadbed, which is the main component of the total settlement of the roadbed. When the compaction degree increases from 90% to 96%, the maximum settlement value of the model and the settlement value of the center point of the surface of the roadbed decrease by about 2% to 4%. Chen et al. [42] used the PLAXIS software (2023) package to model a highway subgrade and found that the maximum settlement of the model with a compaction degree of 96% was 12.3 mm, while that of the model with a compaction degree of 90% was 12.8 mm, a decrease of about 3.9%, which is consistent with the conclusion of the simulation test. Therefore, in the actual project, the increase in compaction degree has a positive effect on the control of roadbed deformation.

4.3.3. Simulation of Different Vehicle Loads

The design level of the road under study is highway, with high traffic flow every day and a large amount of goods distribution, there are heavy loads, and there are differences in the effects of different loads on the stresses and vertical displacements of the roadbed, so the stresses and vertical displacements of the roadbed under different loads are analyzed. The model used in this section is based on the completion of the roadbed filling, and according to the design documents, 0.2 m sub-base and 0.38 base layers are filled on the upper surface of the roadbed to simulate the effects of the roadbed, base layer, and sub-base layer under different vehicle loads.

(1)

Realization of vehicle load in ABAQUS

Through reviewing the literature, it is understood that the simulation of vehicle loads are mainly static loads, moving loads, and vibration loads, of which vibration loads include sinusoidal loads, triangular loads, and random loads. The static load is assumed to be a stationary vehicle for the constant load, numerically expressed as the weight of the car; the mobile load model is used to enable the static load with the time change to have a mobile process, an optimized static load model, but in the actual situation, there is still a gap. Considering that the car wheels are traveling on the road according to a certain frequency and amplitude jumping on the road surface with periodicity and vibration, this causes the additional stress of the roadbed to be higher than the static load. Combined with the above analysis, this paper adopts the steady-state sinusoidal fluctuation load to approximately characterize the vehicle load [43], ignoring the influence of horizontal load, and its expression (15). The pressure acting on the road surface is mainly transmitted through the wheel tires, the contact track of the tire and the road surface is elliptical, the contact between the tire and the road surface is idealized as a rectangle of 0.2 m × 0.3 m in the simulation, the axle distance on both sides of the wheel is taken to be 1.8m, and the grounded pressure of the tire is taken to be 0.7 MPa. Sinusoidal fluctuating loads of the vehicle are realized in the software application using the DLOAD subroutine.

$$P\left(\mathrm{t}\right)=P_0+\alpha P\sin \left(\omega \mathrm{t}\right)$$

(15)

In the formula, $P_0$ is static load of vehicle; $\alpha$ is the dynamic wheel load coefficient is 0.12; and $P$ is the load amplitude, $P=M_0\mu {\omega }^2$, where $M_0$ is the unsprung mass, which is taken as 250 (N·s²)/m, $\mu$ is the road surface geometric irregularity loss height of the road surface roughness, which is taken as 0.002 m, and $\omega$ is the circular frequency of vibration is affected by the driving speed and vehicle type. $\omega =2\pi \mathrm{v}/L$, and L is 10 m.

The roadbed model is divided according to the six lanes of the actual highway, in which the width of each lane is 3.75 m, and the length is 10 m. The vehicle movement load is applied to the middle lane of the left and right roadbeds, the sinusoidal fluctuating load is applied on the divided loading area, and the Z-axis is used as the traveling direction, as shown in Figure 20. The vehicle speed is selected as 80 km/h, and the loads are 100 kN (standard load), 120 kN, 160 kN, and 200 kN to analyze the effects of the loads on the vertical displacement and shear stress at different depths of the roadbed.

(2)

Grid division of the model

The fineness of the model grid has a large impact on the results: if the grid is too large, although it shortens the running time, it may lead to results that are not accurate enough; if the grid is too dense, it will prolong the software calculation time. Therefore, considering the influence range of the vehicle load, the vehicle load range grid is divided into a finer grid, and the original foundation part is divided into a larger grid. The calculation cell type used for the roadbed part is the C3D8R cell, as shown in Figure 21.

(3)

Analysis of calculation results

The changes in vertical displacement and shear stress of the highway base, sub-base, and roadbed under different loads were simulated, and the calculated deformation cloud diagram is shown in Figure 22.

Since the vehicle load is symmetrically distributed in the plane with the center of the roadbed, the displacement cloud map is also symmetrically distributed with the centerline of the roadbed. Therefore, taking the centerline of the roadbed as the starting point, the relationship curve between the vertical displacement and the distance from the centerline at each place of the top surface of the roadway is plotted, as shown in Figure 23.

As shown in Figure 23, as the distance from the axle center increases (6~8 m), the vertical displacement decreases. The deformation at 1m to the right of the wheel is attenuated by 31%, and the displacement changes near the axle center and the wheel are small. The wheel has small external constraints and small deformation. The displacement at the top of the base is the largest, and the peak displacement is 0.42 mm when the load is 200 kN, which is 56% higher than that at 100 kN, and the load has a significant impact. Wang et al. [44] used the ABAQUS software package to simulate the stress on the base layer. The results showed that the displacement attenuation was about 30% at 1 m away from the load center, and when the load increased from 100 kN to 200 kN, the maximum displacement increase was 55~60%. Therefore, lane monitoring should be selected in actual management to strengthen the management of overloaded vehicles.

As can be seen from Figure 24 below, the main part of the shear stress effect is the top surface of the base layer, which is in direct contact with the wheels. The shear stress increases with the increase in the load, and when the load increases from 100 kN to 200 kN, the shear stress of the top surface of the base layer increases by 35%, which shows that overload has a greater effect on the shear stress. When the vehicle load is certain, the shear stress decreases gradually with the increase in depth, and the change in load has no effect on the shear stress on the top surface of the base layer, and the shear stress is 0.004 MPa. In the subsequent construction process, it can be seen that if you want to improve the shear capacity of the pavement, you can realize it by improving the strength of the base layer.

5. Conclusions

This paper studies the feasibility of using recycled materials from construction waste as roadbed fillers, focusing on the analysis of its road performance and the effects of cement improvement on material strength and water stability. Through the implementation of actual engineering test sections, the optimal construction parameters and processes were determined, and the key indicators of the post-construction filling quality of the roadbed were tested. Finally, combining field monitoring and numerical simulation, the deformation characteristics of the roadbed filled with construction waste were studied. The main conclusions are as follows:

(1)

Recycled construction waste has irregular shapes, wide particle size distribution, rough surface and microcracks. Its natural moisture content is 8.5%, its water absorption is 11.73%, and its apparent density is 2.61 g/cm³. The liquid limit of recycled fine aggregate is 20%, the plasticity index is 5.4, and the relationship curve between coarse material content and maximum dry density is Y = 1.73821 + 0.00132X. Although its physical properties are not as good as natural aggregate, its load ratio (25–55%) and low expansion characteristics meet the requirements of highway subgrade filler specifications.

(2)

The loose paving thickness of the lower embankment, upper embankment, and roadbed is 350 mm, 350 mm and 250 mm, respectively. After multiple rolling, the compaction degree meets the requirements. After 6 to 8 times of vibration compaction of the loose paving thickness of 250 mm, the settlement difference is less than 5 mm and tends to be stable. When the loose paving thickness of the embankment is 350 mm and the compaction thickness after eight times performing vibration compaction is 291~301 mm. According to the results of finite element simulation, the settlement of the roadbed is stable at 20~30 mm. Under different loads, the vertical displacement is inversely proportional to the axle distance, and overloading significantly increases the vertical displacement and shear stress.

(3)

The complex composition and large strength differences in construction waste fillers lead to significant discreteness in their physical and mechanical properties, which may affect the universality of the simulation results. This study did not fully consider the performance differences in construction waste from different sources and lacked a systematic evaluation of long-term service performance. In the future, it is necessary to optimize material properties, improve numerical models, and conduct long-term performance tests.