In Situ Investigation of the Dynamic Response and Settlement in the Expressway Culvert–Subgrade Transition Section Using a Vibration Exciter

Abstract

During the operational phase of the expressway, a significant challenge arises concerning substantial differential settlement in the transition zone connecting the culvert and the general subgrade, affecting its smoothness. In order to address the issue of abrupt stiffness variations within the transition section and to mitigate the occurrence of differential settlement, a gradient pile–reinforced-concrete slab composite foundation was implemented for the first time within an expressway culvert–subgrade transition section. At the same time, an in situ vibration test was conducted through the SBZ30 vibration exciter to comprehensively understand the vertical dynamic responses in the culvert–subgrade transition section under various axle loads and speed conditions. Furthermore, continuous monitoring was conducted to track the long-term settlement of the roadbed. The findings indicate that the utilization of gradient pile–reinforced-concrete slab composite foundations can significantly mitigate the amplitude of the dynamic response parameters. Moreover, dynamic parameters and attenuation coefficients exhibit a gradual reduction as the depth increases. Dynamic stresses, acceleration, and displacements on the roadbed surface exhibited positive correlations with both the axle weight and vehicle speed. However, at deeper depths, the load weight exerted a more pronounced influence. As the speed rose, acceleration decayed faster, affecting a shallower depth. Conversely, the increased load slowed the acceleration decay. The cumulative deformation of the roadbed and the number of excitations followed exponential function characteristics. Settlement values progressively increased while the settlement rate gradually diminished, eventually reaching a stable state, ultimately stabilizing within 4.7 mm. These research outcomes offer valuable guidance and serve as a reference for the implementation of gradient pile–reinforced-concrete slab composite foundations within the culvert–subgrade transition section.

1. Introduction

In the course of highway construction, the integration of subgrade sections with culvert segments is a common occurrence. Nevertheless, the operational phase often witnesses the emergence of substantial differential settlements between the road and the culvert, resulting in abrupt alterations in vehicle behavior at the junction. This phenomenon aligns with the mechanism underlying the “bridgehead bump” phenomenon, and it has been elucidated that the primary driver behind the observed differential settlement lies in the substantial disparity in stiffness between the subgrade structure and the culvert structure [1,2]. This sudden alteration in stiffness has the potential to induce rapid fluctuations in the dynamic response of traffic loads at the juncture, consequently giving rise to noteworthy differential deformation following prolonged interaction. Furthermore, once differential deformation takes root, it can further amplify the dynamic impact of traffic loads. Under the influence of multifrequency traffic load excitation, the differential deformation exhibits a progressive escalation [3].

In soft-soil regions, the management of differential deformation becomes more challenging due to the elevated water content and consolidation settlement characteristics inherent in the soft-soil foundation [4]. Currently, the primary strategy for managing settlement within the transition sections of highways situated in soft-soil regions revolves around reducing settlement through soil improvement or foundation treatment to establish a composite foundation. This approach can be broadly classified into the following three key aspects: (1) Subsoil treatment [5,6,7]: This entails the augmentation of the deformation modulus and bearing capacity of the unsaturated soft-soil foundation. It is achieved through various techniques, such as soft-foundation replacement, drainage consolidation, tamping compaction, and composite-foundation treatment. (2) Subgrade-filling treatment [8,9,10,11]: The management of subgrade filling involves the utilization of well-graded materials, such as sand, gravel, crushed stone, or lime soil, as fill, effectively diminishing the consolidation settlement experienced after construction. Employing lightweight filling materials to substitute the initial fill can reduce the embankment’s self-weight. Furthermore, the incorporation of geocells and geogrids within the roadbed filling layer enhances its overall stability, thereby mitigating the differential settlement deformation within the transition section. (3) Road-surface treatment [12,13,14]: One approach involves the installation of an approach slab between distinct sections to distribute the settlement disparity over the slab’s length. Alternatively, a transition paving technique may be employed during the pavement construction process. While the latter method is straightforward and facilitates easy repairs, it is noteworthy that the pavement may exhibit suboptimal grading during the initial stages of paving. When employed in combination, these methods have demonstrated their effectiveness in effectively managing differential settlement within transition sections located in soft-soil regions [15,16]. Nevertheless, it is important to note that these methods are capable of mitigating the overall settlement, yet they do not provide a fundamental solution to the underlying issue of substantial stiffness disparities.

In comparison to highways, railway engineering imposes more stringent demands on the control of roadbed deformation. The Technical Code for Ground Treatment of Railway Engineering (TB10106-2010) [17] prescribes that the utilization of a pile–net (raft) composite foundation, akin to the pile-supported subgrade, is restricted to addressing typical soft-foundation conditions. In cases of transition sections with deep soft foundations, the recommended foundation-treatment approach entails the implementation of a pile–plate structure. This method has been employed in high-speed railway projects for numerous years, demonstrating its efficacy in effectively resolving the challenge of postconstruction differential settlement within soft-soil foundations subjected to traffic loads [18]. Furthermore, research has revealed that the uneven settlement within the transition section can be mitigated by incorporating a rigid–flexible gradual-transition section [19,20]. Tailoring the pile–slab structure design theory to align with the distinctive attributes of highway roadbed structures and traffic loads, the adoption of the gradient pile–reinforced-concrete slab composite foundation unquestionably represents an innovative approach to tackle the pivotal challenges encountered in the aforementioned project.

The dynamic response of the roadbed serves as a direct indicator of its dynamic stability under the influence of dynamic forces imposed by traffic loads and is contingent upon various influencing factors [21]. The acquisition of dynamic response characteristics constitutes the fundamental groundwork for assessing the functional attributes of the roadbed [22]. The settlement, being the most perceptible manifestation of the roadbed’s behavior, offers invaluable insights into its operational performance. Highways’ dynamic response and settlement behavior are typically analyzed through theoretical analysis, numerical simulations, and field-testing methodologies. Theoretical analysis is a valuable approach for deriving analytical principles. Quan introduced a novel analytical solution for the dynamic response of transversely isotropic viscoelastic pavements employing the wave propagation method [23]. Yang formulated an analytical layer element method specifically designed for the analysis of multilayer transversely isotropic saturated soils subjected to rectangular moving loads [24]. However, obtaining an exact solution for complex highway structural-layer materials can be challenging at times. As a result, numerical modeling was employed to surmount certain complexities inherent in theoretical analysis, providing an objective representation of the comprehensive attributes of the roadbed. Gallego advanced a finite-element model to emulate the behavior of embankment–structural transitions [25]. Shan harnessed the finite-element method to contrast the dynamic response between two distinct transition zones located at the juncture of a bridge and a conventional roadbed [26]. It is imperative to underscore that both theoretical analysis and numerical simulations must be grounded in field testing, with their results necessitating comparison and validation to bestow engineering practicality [27]. This imperative arises from the fact that field-test investigations offer the most objective and realistic portrayal of the working state within the embankment [28]. Numerous studies have been undertaken concerning field tests conducted in railway-sector transition sections. Mei scrutinized the attributes of peak dynamic displacements occurring along the track and transition subgrade under the influence of heavy-haul trains through field dynamic testing [29]. Hu undertook field dynamic tests within the culvert-embankment–culvert transition zone, encompassing train operating speeds spanning 5–360 km/h, and revealed that the train axle weight and travel direction exerted a more pronounced influence on the sections and layers within the immediate vicinity [30]. Huang executed field-monitoring tests within the context of the foam-concrete roadbed–culvert transition zone in a ballast railway and derived the relationship between the thickness of the foam-concrete layer and the dynamic response parameters [31]. Nevertheless, the existing body of literature on the dynamic characteristics of expressway transition sections remains relatively limited. Furthermore, the gradient pile–reinforced-concrete slab composite foundation treatment is used for the first time in the highway field, and there is a dearth of pertinent research concerning the dynamic response and settlement characteristics of the roadbed when employing this foundation-treatment approach.

In this investigation, a gradient pile–reinforced-concrete slab composite foundation is applied for the first time within an expressway culvert–subgrade transition section. To assess the road-base performance postfoundation treatment, an in situ vibration test is conducted using the SBZ30 vibration exciter, enabling a comprehensive evaluation of the vertical dynamic responses within the culvert–subgrade transition section under varied axle loads and speed conditions. Furthermore, continuous monitoring is being conducted to track the long-term settlement of the roadbed. This study marks the first instance of conducting on-site excitation tests within a highway transition section. The findings of this study furnish essential original data that can serve as a cornerstone for in-depth investigations into the dynamic characteristics and settlement patterns of road–culvert transition sections treated with gradient pile–reinforced-concrete slab composite foundations.

2. Problem Formulation

The transition section project, situated between the road and culvert, is positioned in Deqing, Zhejiang Province. This region is characterized by a local soil quality typical of a lacustrine and limnetic plain area, featuring a predominantly flat terrain. The prevailing engineering geological factor encountered along the project alignment predominantly comprises poor subsoil conditions. Notably, the depth of the soft-soil layer exhibits significant variation, spanning a range from 18.6 to 43.7 m. The primary physical and mechanical characteristics of the soft-soil layer are detailed in

Table 1. Mechanical parameters of the soft soil.

Layer Number	Soil Layer Name	Density/(g∙cm−3)	Coefficient of Consolidation/(10−3 cm2∙s−1)	Modulus of Compression/MPa	Direct Shear Test		Bearing Capacity/kPa	Static Cone Penetration Test
					Cohesion/kPa	Internal Friction Angle/°		Cone Resistance/kpa	Sidewall Resistance/kPa
②3	Mucky clay	1.76	0.31	3.19	10.1	4.8	55	530	12.9
②4	Silty clay	1.84	0.45	5.26	14.5	11.0	90	800	19.8
③3	Mucky clay	1.77	0.31	3.36	10.7	5.5	55	870	13.5
③4	Silty clay	1.82	0.45	14.64	21.5	10.5	100	1240	25.0

.

In response to the substantial settlement challenges posed by the deep soft ground, the foundation-treatment approach employed within this road–culvert transition section entails the utilization of a gradient pile–reinforced-concrete slab spanning a total length of 31.2 m.

The longitudinal section design diagram for this transition section is depicted in Figure 1. On the left side of the transition section, the highway culvert segment underwent treatment involving cement mixing piles spaced at 1.4 m intervals, featuring a pile length of 20 m, and the roadbed height was set at 3.5 m. On the right side of the transition section, the standard roadbed area was reinforced with 45 m long prestressed pipe piles spaced 2.7 m apart to address the soft-ground foundation. The roadbed height in this segment was set at 4.1 m. Additionally, a solidified-layer process was conducted as a preparatory step before commencing the foundation treatment.

To ensure the alignment’s continuity and smoothness at the junction where the cement-mixing pile section on the left end of the transition segment interfaced with the pipe-pile treatment section on the right end, the pile length on the left end of the variable-length pile–concrete slab structure was set at 21 m, while the pile length on the right end was extended to 45 m. Within this range, the thickness of the soft-soil layer exhibited a progressive increase. To facilitate a linear transition of the settlement in this section, the pile spacing was set at 2.6 m, with each successive row along the longitudinal direction incrementally extending the pile length by 2 m. The tubular piles employed in this arrangement boasted a diameter of 0.5 m. A total of 11 horizontal tubular piles and 12 longitudinal tubular piles were integrated for this purpose. A reinforced-concrete slab (31.2 (Length) × 33.5 (Width) × 0.5 m (Height)) was placed on top of the pile. The upper surface of the slab was equipped with a reinforcing mesh (23 × 22φ20 mm), while the lower surface of the slab was fitted with a reinforcing mesh (25 × 24φ16 mm). Additionally, several hoop bars were integrated into the design. The design of the gradient pile–reinforced-concrete slab composite foundation is shown in Figure 2.

3. In Situ Test with Vibration Exciter

3.1. Test Device

Appropriate test equipment is essential to ensure the seamless execution of the experiment [32]. The excitation system comprised a shaker, frequency converter, counterweight block, and water-cooling system, as depicted in Figure 3. This instrument encompasses the functionalities of variable frequency, variable torque, and variable direction. Tongji University independently developed this system, enabling the simulation of dynamic loads generated by trains at various axle loads and speeds. The counterweight is constructed from cement concrete and serves a dual purpose. Firstly, the contact area at the base of the counterweight block is employed to regulate the dynamic stress imparted by the shaker onto the roadbed surface. Secondly, it acts as a safeguard against equipment displacement caused by the shaker’s excessive force, preventing any unwanted upward movement. The counterweight was affixed to the exciter using steel plates and long bolts. The contact areas between the counterweight and the pavement were configured as two rectangular regions (0.32 × 0.22 m²), positioned at a separation distance of 1.8 m.

The data-acquisition system comprises a transducer, a dynamic collector, an amplifier, and a test computer. The voltage signal produced by the transducer in response to the vibrational load undergoes amplification through the amplifier; following which, the data are collected by the dynamic collector.

The sensor calibration information is of paramount importance, as emphasized by previous research [33]. In this investigation, the JMYJ-1410 m dynamic earth pressure cell sensor was employed with a 2.0 MPa range and a calibration coefficient of 0.0009648. Additionally, the CA-YD-117 accelerometer was used, featuring a range of 1500 m·s⁻², a sensor voltage sensitivity of 0.01319 V·s⁻²/m, and a charge sensitivity of 61.72 pC/(m/s). The JMDL-5105Y vibration pickup had a range of 5.2 mm, a velocity sensitivity of 34.788 V/m, and a displacement sensitivity of 10,469.34 V/m.

3.2. Sensor Buried

Currently, research on roadbed excitation tests is more prevalent in the railway sector, while highways typically employ real vehicle tests. To investigate the dynamic characteristics of the road–culvert transition section after the gradient pile–reinforced-concrete slab composite foundation treatment, on-site excitation tests were conducted for the first time in the field of highway engineering, with section B of the transition section being selected for the study. The dynamic response under the excitation load is acquired through the use of a sensor positioned at the location indicated in Figure 4. Two vibration pick-ups and dynamic earth pressure cells were positioned 2 m apart on the top of the asphalt surface to gauge the dynamic displacement, vibration acceleration, and vertical dynamic stress of the upper surface layer. Eight dynamic earth pressure cells and accelerometers were deployed to monitor the vertical dynamic stresses and vibration accelerations within the embankment at depths of 0.9 m, 1.7 m, and 2.5 m from the surface.

To account for the long-term working stability of the roadbed, settlement monitoring was deemed necessary. A total of five settlement observation points were strategically positioned for this test: four on the top surface of the high-strength bolts at the four corners of the excitation system, and one on the top surface of the shaker. The settlement value is defined as the difference between the settlement observations before and after the excitation test.

The sensor was buried as follows: firstly, prior to burying the sensor, site formation is performed using an excavator. Subsequently, in accordance with the sensor burial program, the sensor wire protection trench is excavated. This protection trench is sequentially connected, forming a progressively wider and deeper groove based on the actual hole position of the sensor. Following the completion of the protective trench construction, the sensor is buried. By employing fiber-optic sensors, losses associated with low-frequency pickups can be effectively mitigated [34]. One of the signal cable protections involves the use of PVC pipe. It is important to note the configuration of different PVC pipe segments, with the outermost PVC pipe extending beyond the roadbed to prevent potential drainage damage. The final step of the burial process includes the initial measurement reading and backfilling with mulch to ensure proper instrument functionality.

3.3. Test Scheme

The parameters of the test conditions are detailed in

Table 2. Excitation test parameters.

Group	Excitation Frequency/Hz	Simulated Axle Load/t	Simulated Vehicle Speed/km∙h−1	Cumulative Number of Vibrations	Simulated Operating Hours/Day
1	9.5	35	105	4.0 × 103	3.8
2	9	30	100	7.0 × 103	6.6
3	8.5	30	90	1 × 104	9.5
4	9.33	40	100	1.4 × 104	13.2
5	8.67	35	90	1.7 × 104	16.1
6	9.33	40	100	2 × 104	18.9
7	8.17	30	90	2.3 × 104	21.7
8	8.5	35	90	2.6 × 104	24.5
9	9.5	40	100	3 × 104	28.3
10	8.5	35	95	3.3 × 104	31.1
11	8.83	40	95	3.6 × 104	33.9
12	7.83	30	80	3.9 × 104	36.8
13	8.5	40	90	4.2 × 104	39.6
14	7.6	35	80	4.5 × 104	42.4
15	7.98	35	90	4.8 × 104	45.2
16	8.5	40	90	5.1 × 104	48.1
17	9	45	95	5.4 × 104	50.9
18	9.5	50	90	5.8 × 104	54.6
19	9.83	50	100	6.1 × 104	57.5
20	7.6	35	80	1.6 × 106	1507.6

, which includes parameters such as simulated load, axle weight, vehicle speed, and simulation duration. These parameters need to be determined in advance. According to the actual loading condition of the expressway, the designed test scenario entails axle weights ranging from 35 to 50 t, with vehicle speeds varying between 80 and 105 km/h [35]. The excitation frequency is obtained by the following equation:

f = v/3.6l

(1)

where v is the vehicle speed and l is the vehicle wheelbase.

Groups 1–19 represent short-term excitation conditions, aimed at acquiring the dynamic response characteristics of the roadbed. Group 20, on the other hand, represents the long-term excitation condition, specifically designed to measure the long-term settlement of the roadbed.

4. Analysis of Test Results

4.1. Dynamic Stress

4.1.1. Variation of Vertical Dynamic Stress

Figure 5 displays the variation of the dynamic stress along the depth direction under dynamic load excitation with specific loads and different speeds. To align with practical engineering applications, a straight line representing 0.2 times the self-gravity–stress relationship is also included in the figure. The depth of the working area of the roadbed is indicated by the longitudinal coordinates where the self-gravitational stress line intersects with the dynamic stress change line.

The results uncover a pattern of dynamic stress variation characterized by a gradual decrease with depth, displaying a typical exponential decline. The loaded road layer and subgrade exhibit a notably higher rate of decay in the initial and middle stages compared to the later stages, reflecting the dissipation of a substantial amount of dynamic loading energy. In the measurement data obtained under the 30/40 t load, it is observed that the decay rate of dynamic stress in the upper layer of the roadbed is greater than that in the middle and lower layers. Furthermore, an increase in vehicle speed results in an increase in surface dynamic stress, but this does not necessarily correspond to a proportional increase in dynamic stress in the middle and lower layers.

Additionally, under varying working conditions involving different axle loads and vehicle speeds, the depth of the roadbed’s working area follows an inconsistent pattern of change. Specifically, an increase in vehicle speed leads to an expansion of the depth of the working area, while an increase in the axle load may cause a reduction in the depth of the roadbed’s working area.

The variation of dynamic stresses along the depth for different axle-load conditions at 90/100 km/h vehicle speeds is presented in Figure 6.

For different axle-load conditions at the same vehicle speed, the impacts of dynamic stresses are primarily concentrated in the surface layer. Larger axle loads result in higher dynamic stress responses, but there is also a greater degree of attenuation in the pavement layer. At higher vehicle speeds, its influence on the dynamic parameters of the lower layer is nearly negligible. For instance, in the test results obtained at 100 km/h with a 30–45 t load, the depth of the working zone remains primarily around 2.3 m.

4.1.2. Relationship between Dynamic Stress and Velocity and Load

The variation pattern of dynamic stress amplitude with vehicle velocity is depicted in Figure 7. The curves in the graphs all exhibit a relatively consistent trend, where dynamic stress amplitude tends to increase with the rising vehicle speed, indicating a positive correlation. However, the rate of growth is not particularly pronounced.

The dynamic stress amplitude at locations with the greater burial depth exhibits an inflection point with increasing vehicle speed. This behavior is primarily attributed to variations in the compaction degree of the roadbed fill material below, induced by high-frequency excitation, which subsequently impacts the vibration characteristics. As a result, dynamic stress amplitude fluctuates, although the overall trend of a positive correlation remains evident.

The variation in the dynamic stress amplitude at different depths as the load is altered is illustrated in Figure 8. Two speed groups, namely 90 km/h and 100 km/h, can be distinguished based on differences in the axle load.

The findings indicate that, under the influence of load, there is a relatively consistent trend in the dynamic stress. Specifically, the surface dynamic stress increases in proportion to the load. However, as the depth increases, a negative correlation between dynamic stresses at each point and the load becomes evident. This implies that, overall, the amplitude of dynamic stresses tends to decrease with the increasing load at greater depths within the roadbed.

It is evident that the correlation between the dynamic stress, speed, and load depends on the burial depth conditions. When the burial depth is shallow, increases in the speed and load result in higher dynamic stress amplitudes. However, when the burial depth is deeper, specifically below the road base, an increase in speed leads to an overall rise in dynamic stress, with some fluctuations, while an increase in the load leads to an overall decrease in dynamic stress.

4.2. Vibration Acceleration

4.2.1. Variation of Vibration Acceleration

The acceleration and attenuation coefficient along the depth direction under the influence of various vehicle speeds for a 30 t axle load are shown in Figure 9.

The trend in acceleration with depth exhibited similarities to the dynamic stress attenuation pattern, showing an exponential decrease with increasing depth. Furthermore, the acceleration decay coefficients exhibit a consistent trend, decreasing gradually with increasing speed along the depth direction. While dynamic parameters in the pavement layer experience rapid growth at higher vehicle speeds, the response level in the road-base layer, after considering the attenuation effect of the pavement layer, remains relatively small.

The changes in acceleration and attenuation coefficients at various depths under different loads at the same velocity are depicted in Figure 10 and Figure 11. Under the same velocity conditions, axle load exerts a noticeable impact on the acceleration level. As the axle load increases, the acceleration level rises, and the rate of decay decreases, resulting in an increased depth of the working zone. This observation highlights the dominant role of the axle load in the overall parameter variation when acceleration is employed as a control parameter to analyze the dynamic response pattern of the roadbed.

The acceleration increment associated with varying load increments at different depths is illustrated in Figure 12. The figure reveals a nonlinear trend in the impact of load changes on acceleration. As the load increment increases, the acceleration growth demonstrates a corresponding acceleration in growth.

4.2.2. Relationship between Vibration Acceleration and Velocity and Load

The changes in acceleration magnitude at various depths concerning vehicle speed are depicted in Figure 13.

It is evident that an increase in vehicle speed leads to a substantial augmentation in the acceleration magnitude. Furthermore, Figure 14 demonstrates the alterations in the acceleration magnitude with the load at different depths, illustrating that an increase in the axle load similarly results in a continuous increase in the acceleration magnitude. These findings emphasize the significant impact of both the load and speed on the acceleration variation across various depth ranges.

The trends in acceleration attenuation coefficients with respect to vehicle speed and load are demonstrated in Figure 15 and Figure 16, respectively. These figures reveal that the acceleration attenuation coefficient diminishes as the speed increases, signifying an acceleration of attenuation with higher speeds, consequently reducing the depth of influence. Conversely, in the case of the load, an increase in the load results in a deceleration of attenuation, leading to an expansion of its influence depth.

4.3. Dynamic Displacement

4.3.1. Variation of Dynamic Displacement

Figure 17 displays the distribution of the dynamic displacement amplitude and its attenuation coefficient along the depth for various operating conditions obtained from excitation tests. The figure reveals that the dynamic displacement magnitude gradually diminishes with greater depth, and the rate of this decrease gradually decreases. Furthermore, the dynamic displacement attenuation coefficient decreases more rapidly as the vehicle speed increases. In other words, higher vehicle speeds lead to a faster decay of the dynamic displacement amplitude in the depth direction compared to slower vehicle speeds.

In Figure 18 and Figure 19, the distribution of the dynamic displacement along the depth under different loads is displayed. It is evident that the dynamic displacement magnitude at various depths increases as the load weight increases. Furthermore, the growth of the dynamic displacement decay coefficient corresponds with the increase in the load weight, signifying that greater load weights result in a slower decay of the dynamic displacement amplitude and a deeper influence.

4.3.2. Relationship between Vibration Acceleration and Velocity and Load

The change in dynamic displacement amplitude at different depths with respect to vehicle speed is depicted in Figure 20. The figure reveals that the surface dynamic displacement increases with higher velocities. In contrast, the deep-layer dynamic displacement remains relatively unaffected by velocity changes. This phenomenon occurs because, at a depth of 0.9 m, the dynamic displacement amplitude has already undergone significant attenuation, resulting in a smaller magnitude after attenuation, thus rendering the changes less pronounced.

The variation of the dynamic displacement magnitude with the load at different depths is illustrated in Figure 21. It is evident from the figure that the dynamic displacement magnitude at both the surface and deeper depths experiences a substantial increase with the rising load when the vehicle speed is set at either 90 km/h or 100 km/h. In the deeper layers of the roadbed, the impact of speed on the dynamic displacement magnitude change is relatively minor. However, at greater depths within the roadbed, there is a pronounced positive correlation between the load and dynamic displacement magnitude. Hence, the load exerts a more substantial influence on the dynamic displacement magnitude than velocity.

The variation in the dynamic displacement attenuation coefficient with vehicle speed is depicted in Figure 22. With a 30 t axle load, the dynamic displacement attenuation coefficient at depths of 0.9 m and 1.6 m decreases as the vehicle speed increases. This phenomenon is attributed to the faster decay of high-frequency vibrations in the medium. However, with a 40 t axle load, this trend becomes less prominent, suggesting that the load exerts a more significant influence on the dynamic displacement attenuation characteristics in the roadbed.

Figure 23 displays the change in the dynamic displacement attenuation coefficient with the load. In the figure, it is evident that, at speeds of 90 km/h and 100 km/h, the dynamic displacement attenuation coefficient for each layer in the roadbed increases with the load. This indicates that a greater load results in a slower attenuation of the dynamic displacement amplitude and a deeper impact.

4.4. Monitoring of Long-Term Settlement

After averaging the settlement values obtained from the four settlement observation points under long-term excitation test conditions, the relationship between vertical settlement and cumulative vibration cycles, reflecting the overall settlement of the roadbed, is depicted in Figure 24.

It is evident that the cumulative plastic settlement of the roadbed increases with the number of vibration cycles and eventually reaches a state of convergence. The settlement follows an exponential development pattern, with the final value of the cumulative settlement under dynamic loading ranging within 4.5 mm. This suggests that the on-site foundation-treatment method is highly effective in controlling settlement. It is important to emphasize that soil properties inherently possess uncertainty. The influence of rainfall and excitation loading can bring about variations in soil attributes, such as cohesion, friction angle, and permeability, ultimately resulting in alterations to the settlement value curve [36]. Furthermore, rainfall causes a reduction in the suction of the unsaturated soil matrix, alongside a decrease in the pile-bearing capacity, subsequently contributing to an increase in settlement [37]. Therefore, it is essential to incorporate the influence of rainfall in the design of the foundation treatment. From Figure 24, the cumulative deformation stage of the roadbed can be categorized into three distinct stages:

• Stage I—Rapid Accumulation of Rainfall Infiltration Deformation (0–4.7 × 10⁵ vibration cycles): During this phase, rainfall infiltration causes an increase in the water content of the loose residual roadbed material. This, in turn, reduces the friction between particles, increases lubrication, and makes the material more prone to compaction under the excitation force;
• Stage II—Slow Increase in Natural-Condition Deformation (4.7 × 10⁵–1.2 × 10⁶ vibration cycles): During this phase of long-term cyclic vibration, the residual roadbed gradually compacts, leading to a decrease in the rate of cumulative settlement;
• Stage III—Cumulative Settlement Stabilization (1.2 × 10⁶–1.6 × 10⁶ vibration cycles): During this phase, the cumulative settlement exhibits minimal change and has converged, reaching a stable stage.

5. Conclusions

In this study, a gradient pile–concrete slab composite foundation was employed for the initial time in the expressway culvert–subgrade transition section located in the soft-soil region of China. This treatment aimed to address the issue of significant differential settlement resulting from the abrupt stiffness transition at the junction of the transition section. Additionally, a field excitation test was conducted using SBZ30 variable-frequency torque vibrators to establish the vertical distribution patterns of dynamic responses under various vehicle axle loads and speeds. Long-term roadbed settlement was also monitored. The primary findings are summarized as follows:

(1)

The gradient pile–concrete slab composite foundation has demonstrated its ability to effectively decrease the amplitude of dynamic response parameters and roadbed settlement. It offers valuable guidance and serves as a reference for settlement control and foundation treatment in similar expressway-transition-section projects;

(2)

Dynamic stress exhibits exponential attenuation characteristics, with the pavement and base layers dissipating a significant amount of dynamic load energy. When the burial depth is shallow, an increase in the vehicle speed and axle load results in higher dynamic stress amplitudes. However, when the burial depth is deeper (beneath the road base), higher speeds lead to an overall increase in the dynamic stress, albeit with some fluctuations, while an increase in the axle load generally leads to a reduction in the dynamic stress;

(3)

Acceleration amplitudes tend to increase with higher axle loads and speed. As the vehicle speed increases, acceleration decay accelerates, resulting in a shallower depth of influence. Conversely, as the load increases, acceleration decay slows down, leading to a greater depth of influence;

(4)

Dynamic displacement amplitude decreases with depth, with slower decay at greater depths. Surface-layer displacement increases with the velocity and load, but deep-layer displacement is largely unaffected by velocity changes. The load has a more significant impact on the displacement amplitude, particularly in deeper subgrade layers, compared to velocity;

(5)

The cumulative deformation of the roadbed and the number of excitations exhibit characteristics consistent with exponential functions, ultimately stabilizing within 4.5 mm. The cumulative deformation process can be divided into three stages: rapid accumulation (0–4.7 × 10⁵ vibration cycles), slow increase (4.7 × 10⁵–1.2 × 10⁶ vibration cycles), and settlement stabilization (1.2 × 10⁶–1.6 × 10⁶ vibration cycles).

In addition, numerous additional factors influence the dynamic response and settlement values of roadbeds. In future research, it becomes imperative to conduct field tests encompassing a broader spectrum of load configurations and rainfall conditions. Furthermore, the integration of numerical modeling can be employed to comprehensively investigate the overall performance of the roadbed.