Urban road traffic load adaptive pavement structure design system and method

Abstract

This invention discloses an adaptive pavement structure design system and method for urban road traffic loads, relating to the field of road engineering technology. It includes a load data construction module, a mechanical estimation module, and a pavement structure generation module. The load data construction module is configured to construct a dynamic load spectrum based on a hidden Markov model. The mechanical estimation module is configured to calculate the mechanical parameters of the pavement structure at different locations by combining the dynamic load spectrum and road segment information. The pavement structure generation module is configured to map the mechanical parameters onto pavement design parameters to generate pavement structure schemes. The pavement structure generation module includes a collaborative optimization unit, which includes a multi-level digital twin model to simulate the dynamic evolution of pavement performance. This invention can proactively adapt to constantly changing traffic load characteristics to suit complex urban traffic environments.

Description

Technical Field

[0001] This invention relates to the field of road engineering technology, and in particular to a system and method for designing urban road traffic load-adaptive pavement structures. Background Technology

[0002] With the accelerated pace of urbanization, urban traffic volume continues to grow, especially the increase in heavy vehicles, which has placed unprecedented pressure on urban roads. Traditional pavement design standards and methods are prone to problems such as accelerated pavement damage, increased maintenance costs, and increased traffic safety hazards when dealing with increasingly complex traffic conditions. Current urban road load calculations originate from the highway system but differ from it, exhibiting complex vehicle composition, discontinuous load characteristics, and uneven load growth. Existing pavement structure design programs often struggle to capture these refined load distribution patterns. Their models are typically based on outdated or simplified vehicle model libraries, with limited input conditions, failing to effectively consider the actual full-load state of vehicles, and unable to flexibly combine inputs according to actual traffic composition, resulting in significant deviations between design results and actual service conditions. Summary of the Invention

[0003] In view of the shortcomings of the existing technology, the purpose of this invention is to provide an urban road traffic load adaptive pavement structure design system and method that can actively adapt to the constantly changing traffic load characteristics in order to adapt to the complex urban traffic environment.

[0004] To achieve the above objectives, the present invention is implemented through the following technical solution: In a first aspect, embodiments of the present invention provide an urban road traffic load-adaptive pavement structure design system, comprising: The load data construction module is configured to: construct a dynamic load spectrum based on a hidden Markov model; The mechanical estimation module is configured to calculate the mechanical parameters of the pavement structure at different locations by combining the dynamic load spectrum and road segment information; The pavement structure generation module is configured to map the mechanical parameters onto pavement design parameters to generate a pavement structure scheme; the pavement structure generation module includes a collaborative optimization unit, which includes a multi-level digital twin model to simulate the dynamic evolution process of pavement performance.

[0005] As a further implementation, the road segment information includes standard road segments and characteristic road segments, and the mechanical estimation module includes a multi-layer elastic system element and a finite element analysis element. The multi-layer elastic system element is used to calculate the mechanical parameters of the standard road segment, and the finite element analysis element is used to calculate the mechanical parameters of the characteristic road segment.

[0006] As a further implementation, the multi-level digital twin model includes at least a road structure geometry model layer, a material property model layer, an environmental load model layer, and a construction process model layer. The models at each level interact with each other to form a coupled digital twin.

[0007] As a further implementation, the multi-level digital twin model also includes a monitoring and feedback model layer, which is used to receive measured data and compare the measured data with the prediction results in the design stage to correct and optimize the design system parameters.

[0008] As a further implementation, the pavement structure generation module also includes a damage accumulation and life prediction unit and a decision unit, and the output signals of the collaborative optimization unit and the damage accumulation and life prediction unit are all transmitted to the decision unit. The decision-making unit searches for the optimal solution in the design parameter space based on the constraints and outputs the final design scheme.

[0009] As a further implementation, the damage accumulation and lifetime prediction unit integrates at least two computational models of microstructural damage theories.

[0010] As a further implementation, the decision-making process of the decision-making unit is represented by the following model: ; The constraints are: P ( LC 20 yeares) 0.90; RM Usage 0.20; in, C i Indicates the initial construction cost. C m It is the cost of maintenance and repair. C d It is the cost of traffic delays. P ( LC 20-years indicates the probability that a road surface will reach its 20-year design life. RM Usage This indicates the rate of recycling.

[0011] As a further implementation, the construction of the dynamic load spectrum based on the hidden Markov model includes: Hidden Markov models are used to describe the dynamic transition law of vehicle load state in order to construct a dynamic load spectrum. The Hidden Markov Model includes a load state set, an observation set, an initial state probability distribution, a state transition probability matrix, and an observation probability matrix. The load state set is a combination of key features related to vehicle load, and the observation set consists of vehicle type, axle type, axle load, and traffic volume information obtained through detection.

[0012] As a further implementation, a 3D visualization and BIM / GIS integration module is also included, which is used to overlay mechanical indicators onto the BIM model in the form of a 3D visualized cloud map.

[0013] Secondly, embodiments of the present invention also provide a method for designing urban road traffic load-adaptive pavement structures, based on the aforementioned design system, comprising: Dynamic load spectrum is constructed based on hidden Markov model; Obtain road segment information and calculate the mechanical parameters of the pavement structure at different locations based on the dynamic load spectrum; The mechanical parameters are mapped onto the pavement design parameters, and a collaborative optimization method including multi-level digital twin models is used to simulate the dynamic evolution of pavement performance and generate a pavement structure scheme that adapts to road traffic loads.

[0014] The beneficial effects of this invention are as follows: The urban road traffic load adaptive pavement structure design system of the present invention includes a load data construction module, a mechanical estimation module, and a pavement structure generation module. The load data construction module constructs a dynamic load spectrum based on a hidden Markov model. The mechanical estimation module calculates the mechanical parameters of the pavement structure at different locations by combining the dynamic load spectrum and road segment information. The pavement structure generation module maps the mechanical parameters to the pavement design parameters to generate a pavement structure scheme. By introducing a dynamic load spectrum, it can cope with complex urban traffic environment problems. Attached Figure Description

[0015] The accompanying drawings, which form part of this invention, are used to provide a further understanding of the invention. The illustrative embodiments of the invention and their descriptions are used to explain the invention and do not constitute an improper limitation of the invention.

[0016] Figure 1 This is a detailed architecture diagram of the design system according to one or more embodiments of the present invention; Figure 2 This is a diagram showing the main architecture of the system designed according to one or more embodiments of the present invention. Detailed Implementation

[0017] It should be noted that the following detailed description is illustrative and intended to provide further explanation of the invention. Unless otherwise specified, all technical and scientific terms used in this invention have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0018] Example 1: Research revealed significant differences in traffic load characteristics across different areas within the same city, necessitating that pavement structure design not only considers general applicability but also specific needs. To optimize urban road pavement structure design, this embodiment provides an urban road traffic load-adaptive pavement structure design system, such as... Figure 1 and Figure 2 As shown, it includes: The load data construction module is configured to: construct a dynamic load spectrum based on a hidden Markov model; The mechanical estimation module is configured to calculate the mechanical parameters of the pavement structure at different locations by combining the dynamic load spectrum and road segment information. The pavement structure generation module is configured to map mechanical parameters onto pavement design parameters to generate pavement structure schemes.

[0019] like Figure 1 As shown, the load data construction module obtains data from the dynamic weighing system, video recognition system, and GPS floating car data system to acquire vehicle type information, axle type and axle load information, and traffic volume information.

[0020] A dynamic weighing system (WIM) refers to a system that measures the weight of vehicles on a road, especially determining axle loads. WIM systems are typically deployed on key road sections and can measure the axle load and total weight of passing vehicles in real time. These data are collected at a high frequency (e.g., multiple times per second) to provide accurate instantaneous values ​​of vehicle loads.

[0021] The video recognition system includes high-definition cameras mounted above or to the side of the road. It uses computer vision technology to detect and classify passing vehicles, and can obtain information such as vehicle type, lane occupancy, vehicle speed, and headway. Furthermore, it can estimate the vehicle's load status through image processing algorithms, such as by analyzing the compression degree of the vehicle's suspension system or the vehicle's posture.

[0022] The GPS floating car data system uploads the vehicle's longitude, latitude, speed, and direction information in real time through GPS devices on the vehicle. The data is usually sampled at a low frequency (e.g., once every 5-30 seconds), covering a wide range and providing the vehicle's macroscopic trajectory and speed distribution.

[0023] This embodiment collects, integrates, and processes traffic load information from various data sources. The load data construction module can identify and finely classify at least 15 types of vehicle models, accurately collecting information such as axle type (single axle, steering axle, three-axle, multi-axle), real-time weight of each axle, wheelbase, driving speed, and lateral distribution within the lane (accurate to the probability distribution of wheel tracks) for each type of vehicle. The load data construction module can also connect with traffic flow detectors (such as geomagnetic sensors, coils, and radar) and public transportation operation data systems to obtain more comprehensive traffic volume information.

[0024] Furthermore, the load data construction module preprocesses information and data from multiple sensors or sources. The data preprocessing stage includes data cleaning, noise reduction, format standardization, and preliminary screening to ensure the quality of data in subsequent processing. For example, outlier removal from WIM data can be achieved using wavelet denoising and Kalman filtering algorithms to effectively filter out sensor noise and improve the accuracy of axle load measurement. Blurred or occluded frames are filtered from video data to achieve high-precision vehicle classification and trajectory tracking. Drift correction is performed on GPS data.

[0025] This embodiment utilizes multiple data sources for fusion analysis to construct a refined load spectrum, reflecting vehicle load information on urban roads to a certain extent. However, these technologies still have certain limitations. For example, while the WIM system can provide accurate axle loads, its deployment cost is high and its coverage is limited, making it difficult to form continuous spatial monitoring. The accuracy of video recognition technology is affected by adverse weather or lighting conditions, and it is difficult to directly obtain the full load status of vehicles. Although GPS floating car data has wide coverage, its location accuracy and update frequency may not be sufficient to accurately capture the fine load distribution of vehicles in specific lanes, and the data itself has problems with noise and uneven sampling.

[0026] Therefore, this embodiment further proposes the concept of dynamic load spectrum. Specifically, the load data construction module uses a hidden Markov model to describe the dynamic transfer law of vehicle load state in order to construct a dynamic load spectrum. lambda = (A, B, pi), Here, lambda represents the Hidden Markov Model. The Hidden Markov Model consists of a load state set, an observation set, an initial state probability distribution pi, a state transition probability matrix A, and an observation probability matrix B. The load state set is a combination of key features related to vehicle load, and the observation set consists of vehicle type, axle type, axle load, and traffic volume information obtained through detection.

[0027] A Hidden Markov Model (HMM) is a statistical model used to describe a Markov process with hidden unknown parameters. This embodiment abandons the traditional static probability distribution representation of the load spectrum and uses an HMM to describe the dynamic transition law of vehicle load state.

[0028] First, the load state set of the HMM is defined. A load state is defined as a combination of key features related to vehicle load. These features include: axle load distribution characteristics (e.g., average axle load, axle load skewness, maximum axle load), load factor (which can be estimated from vehicle posture identified by video or WIM data), lateral distribution characteristics (vehicle's position within the lane, which can be estimated from video or GPS trajectory), and driving speed. These continuous feature spaces are discretized into a finite number of discrete states through cluster analysis, such as "lightly loaded car - inner lane - high speed," "heavily loaded truck - outer lane - medium speed - fully loaded," etc., thus obtaining the load state set.

[0029] Secondly, define the observation set of the HMM. Observations are directly obtainable features, including vehicle type information, axle type and axle load information, and traffic volume information. Examples include axle load values ​​measured by WIM, vehicle type identified by video, and speed from GPS.

[0030] Then, the Hidden Markov Model (HMM) is trained using the fused historical data. The training process aims to estimate the parameters of the HMM, namely the initial state probability distribution Pi, the state transition probability matrix A, and the observation probability matrix B. In the state transition probability matrix A, element a_{ij} represents the probability of transitioning from state i to state j, i.e., P(q_{t+1}=j | q_t=i). In the observation probability matrix B, element b_j(k) represents the probability of observing observation k in state j, i.e., P(q_t=k | q_t=j). The initial state probability distribution pi_i represents the probability of being in state i at the initial time step. Iterative optimization methods, such as the Baum-Weltch algorithm, are used to maximize the probability of the observation sequence.

[0031] In this embodiment, the trained Hidden Markov Model (HMM) can accurately quantify the probability of vehicle load states transitioning from one state to another under specific time periods (such as weekday morning rush hour, evening rush hour, off-peak hours, and holidays) and specific environmental conditions (such as sunny, rainy, and snowy days). For example, the model can reveal that during weekday morning rush hour, the probability of fully loaded trucks is significantly lower than during off-peak hours, and their lateral driving position is more inclined towards the outer lane. This description of state transition probabilities makes the load spectrum no longer a static distribution, but a dynamic process that changes with time and environment, which is the core of dynamic load spectrum. This module can also generate independent and differentiated dynamic load spectra based on different road areas (such as main lines, ramps, intersections, bus stops, and bay areas) and different lanes.

[0032] In this embodiment, the mechanical estimation module is used to calculate the mechanical parameters of the pavement structure at different locations by combining the dynamic load spectrum and road segment information. The road segment information is used to represent different road segment areas, including standard road segments and characteristic road segments. Standard road segments are the mainline traffic areas, while characteristic road segments are complex areas such as ramps, intersections, and bus stops.

[0033] like Figure 1 As shown, the mechanical estimation module includes a multi-layer elastic system unit and a finite element analysis unit. The multi-layer elastic system unit is used to calculate the mechanical parameters of standard road sections, and the finite element analysis unit is used to calculate the mechanical parameters of characteristic road sections. The mechanical evaluation module of this embodiment can calculate the spatiotemporal distribution of key control indicators (such as maximum horizontal tensile strain, maximum vertical compressive strain, and maximum shear stress) of the pavement structure at different locations (such as the bottom of the asphalt layer, the top of the base course, and the top of the subgrade) throughout the entire design service life, based on the input dynamic load spectrum.

[0034] This embodiment establishes a hybrid mechanical model that integrates multilayer elastic system theory and refined finite element method. For standard road sections, a multilayer elastic system model is adopted, whose material parameters (elastic modulus, Poisson's ratio) can be dynamically adjusted according to environmental factors such as temperature and humidity, and the strain rate effect of the material is considered. A nonlinear elastic modulus model can be introduced so that the material modulus can be adaptively adjusted under different stress levels. For complex areas such as intersections and bus stops, a three-dimensional refined finite element model is used. This model can simulate the nonlinear behavior of materials, such as the plasticity of asphalt mixtures and the strain softening of the subgrade, and consider the complex stress states generated by vehicle braking, steering, and acceleration, including horizontal tensile stress, shear stress, vertical compressive strain, and the interaction between the structure and underground pipelines.

[0035] For the pavement structure generation module, the pavement design parameters involved include one or more of the following: asphalt pavement thickness, mechanical modulus of base course material, design parameters of bonding layer, and bearing capacity requirements of subgrade. This embodiment uses a multi-layer feedforward neural network as a neural network mapping model to directly map the mechanical parameters to the key pavement design parameters. These design parameters include, but are not limited to, asphalt pavement thickness, mechanical modulus of base course material, design parameters of bonding layer, and bearing capacity requirements of subgrade.

[0036] The pavement structure generation module has a built-in material library containing the mechanical property parameters of various pavement materials (asphalt mixtures, base courses, subbase courses, and subgrades) under different temperatures, humidity levels, and strain rates. These parameters include dynamic modulus, fatigue equations, crack resistance equations, and creep parameters. Using machine learning models, the library is calibrated and predicted in real time based on measured material property data and environmental monitoring data (such as pavement temperature and humidity sensor data), making these parameters non-static. For example, a predictive model can be built to predict the dynamic modulus of asphalt mixtures at different temperatures based on local meteorological data and pavement surface temperature. This predictive model employs a deep neural network (DNN) structure, capable of capturing the nonlinear relationship between temperature, humidity, and material properties. Simultaneously, the module can also establish a nonlinear stress-strain constitutive model of the subgrade based on field surveys and geotechnical test data, considering its variation with factors such as compaction degree and moisture content. For the subgrade model, a modified Cambridge Model or a hardened soil model is used to more accurately describe the elastoplastic behavior of the soil.

[0037] The pavement structure generation module selects matching pavement materials from the material library based on pavement design parameters to generate pavement structure schemes. This design can proactively adapt to constantly changing traffic load characteristics and material properties. However, in actual construction, there are unavoidable material variability, deviations in process control, and the dynamic aging and damage accumulation process of pavement materials over their service life, necessitating further optimization of the pavement structure schemes.

[0038] Therefore, as Figure 1 As shown, the pavement structure generation module in this embodiment includes a collaborative optimization unit, a damage accumulation and life prediction unit, and a decision unit. The output signals of the collaborative optimization unit and the damage accumulation and life prediction unit are both transmitted to the decision unit. The collaborative optimization unit includes a multi-level digital twin model, which includes at least a pavement structure geometric model layer, a material property model layer, an environmental load model layer, and a construction process model layer. The models at each level interact to form a coupled digital twin, simulating the dynamic evolution of pavement performance.

[0039] The multi-level digital twin model is not simply a data integration, but a high-fidelity virtual representation containing multi-scale and multi-dimensional information. It comprises at least four layers: the first layer is the geometric model of the pavement structure, accurately reflecting physical properties such as layer thickness, material type, and joints. The second layer is the material property model, which includes not only initial material parameters from the design phase, such as the viscosity-temperature-time relationship of asphalt and the gradation of aggregates, but also dynamically updates and simulates the evolution of material performance, including asphalt oxidation aging and performance degradation caused by water damage. The third layer is the environmental load model, which receives input from meteorological data and traffic flow monitoring systems (such as inductive loops and video recognition) to simulate the load spectrum and stress-strain response generated by sunlight, cooling, rainwater infiltration, and different types of vehicles (cars, trucks, buses). The fourth layer is the construction process model, which integrates real-time quality inspection data from the construction phase, such as asphalt paving temperature, compaction test data, and interlayer bond strength test data, transforming this data into actual material parameters and interface properties in the digital twin model, quantifying the randomness of construction quality. These hierarchical models interact with each other through a unified interface and data protocol, forming a dynamically updated, highly coupled digital twin to simulate the dynamic evolution of road surface performance.

[0040] In this embodiment, the multi-level digital twin model also includes a monitoring and feedback model layer. The monitoring and feedback model layer is used to connect with the sensor network buried in the built road to receive measured data and compare the measured data with the prediction results in the design stage to correct and optimize the parameters of the design system.

[0041] The sensor network integrates road surface sensors (such as distributed fiber optic sensors, strain sensors, and road condition monitoring vehicles), thermometers, hygrometers, GPS, and inertial measurement units (IMUs) to provide real-time data on the road surface's mechanical response, temperature, humidity, and other parameters under actual service conditions. By comparing these measured data with predictions from the design phase, the accuracy of the performance evolution simulation is calibrated and verified.

[0042] In this embodiment, machine learning algorithms, such as Kalman filtering or Bayesian updates, are used to automatically correct and optimize key parameters in the design system, including localized material mechanics parameters such as asphalt modulus, fatigue equations, subgrade model parameters, and traffic load spectrum models. This continuous model correction process enables the next round of design to be based on more accurate parameters that reflect the actual service environment, thereby achieving continuous design optimization and iterative computational improvement. This module can also perform in-depth analysis of historical monitoring data to discover patterns in pavement performance degradation, providing data support for predictive maintenance.

[0043] This embodiment constructs a highly integrated digital twin model, which can dynamically simulate the performance evolution trajectory of the pavement structure over a 15- to 20-year service life under the combined effects of real construction variability and multiple environmental factors. Based on this, design parameters, such as asphalt layer thickness or binder type, are optimized and adjusted to ensure that the pavement can still achieve the expected service life even with certain construction deviations.

[0044] In this embodiment, the damage accumulation and life prediction unit integrates at least two calculation models. The calculation model can be arbitrarily selected from the asphalt aging model of microstructure damage theory, the fracture mechanics crack propagation model, the stress-strain accumulation rutting model, the temperature crack model, and the water damage model.

[0045] Asphalt material aging models based on microstructure damage theory can simulate the effects of oxidation, thermal fatigue, and other processes on asphalt performance; asphalt pavement crack propagation models based on fracture mechanics can predict the generation and propagation patterns of different types of cracks (surface cracks, semi-deep cracks); asphalt pavement rutting formation models based on stress-strain accumulation can simulate the accumulation of plastic deformation under repeated loading; and asphalt pavement temperature cracking models and water damage models considering temperature gradients and humidity changes are also available. These computational models are integrated into a unified numerical computation framework and combined with material properties, environmental loads, and structural geometry information from multi-level digital twin models to achieve monthly and yearly evolution simulations of key pavement performance indicators such as smoothness IRI, crack rate, and rutting depth.

[0046] For example, when the simulation reaches a specific point in time, if the environmental model predicts continuous low temperatures and rainfall, while the material model shows that the asphalt has aged to a certain extent, it will trigger the evolution calculation of temperature cracks and water damage, and update the overall performance indicators of the pavement structure. Through this setting, the performance changes of the pavement during its 15 to 20-year design service life can be simulated, and the probability of fatigue cracking, rutting and other failure modes of the pavement during its design service life can be predicted, and probability-based life prediction results can be given.

[0047] In one embodiment, an asphalt aging model is simulated based on the oxidation kinetics and thermal fatigue characteristics of asphalt materials. This model considers the chemical degradation processes of asphalt mixtures under long-term exposure to air and high temperatures, such as oxidation and volatilization, as well as the physical aging processes under repeated temperature cycling. For example, the decrease in the stiffness modulus E of asphalt over time t can be expressed by the following formula: ; in, k1 represents the initial stiffness modulus, k2 represents the oxidation aging coefficient, and k2 represents the thermal fatigue coefficient. Let represent the temperature difference during the i-th temperature cycle, and n represent the material constant. This model can predict the decay of the material's stiffness modulus over time and environmental factors (such as temperature and oxygen exposure), thereby affecting the overall stiffness and load-bearing capacity of the pavement.

[0048] Combining simulated stress-strain fields and material aging states, an asphalt pavement crack propagation model calculates the generation and propagation rates of surface cracks and semi-deep cracks. This model is based on fracture mechanics principles, considering stress concentration at the crack tip and the material's fracture toughness. In the fracture mechanics crack propagation model, the crack propagation rate da / dN is described by Paris's law: ; Where 'a' represents the crack length, 'N' represents the number of load cycles, 'C' and 'm' represent material constants, and 'ΔK' represents the range of stress intensity factors. This model can predict the generation and propagation patterns of different types of cracks. For example, when the tensile stress in a local area of ​​the pavement exceeds the tensile strength of the asphalt mixture, cracks begin to initiate and propagate with repeated traffic loads.

[0049] Based on the stress-strain accumulation under repeated loading, the rutting formation model predicts the cumulative plastic deformation of each asphalt mixture layer and calculates the rutting depth. This model typically employs an elastoplastic constitutive relation or an empirical cumulative deformation model. In the stress-strain cumulative rutting model, the rutting depth RD is expressed as: ; Where A, B, and D represent material constants, The vertical stress is represented by , and T represents temperature. This model simulates the accumulation of plastic deformation under repeated loading, thereby predicting the increase in rut depth over time.

[0050] In addition, considering diurnal temperature variations and seasonal temperature changes, a temperature crack model simulates the generation and propagation of low-temperature cracks. This model assesses the tensile stress caused by thermal shrinkage of the pavement under extreme low-temperature conditions; cracks form when the tensile stress exceeds the low-temperature crack resistance of the asphalt mixture. A water damage model assesses the impact of moisture on the strength and interfacial bonding of pavement layers, considering the effects of rainfall and groundwater levels. For example, moisture intrusion can lead to delamination at the asphalt-aggregate interface, thereby reducing the strength and durability of the mixture.

[0051] These models operate within a unified numerical computation framework and are combined with material properties, environmental loads, and structural geometry information from digital twin models to simulate the monthly evolution of key pavement performance indicators (such as inter-rater intensity (IRI), crack rate, and rutting depth). For example, in the fifth year of the simulation, when the environmental model predicts consecutive low temperatures (below -10 degrees Celsius) and the material model indicates that the asphalt aging degree has reached a certain threshold, the simulation engine triggers the calculation of temperature crack evolution and updates the overall IRI and crack rate of the pavement structure. In other implementations, it can also perform scenario simulations to analyze the impact of different traffic volume growth rates, increased heavy vehicle proportions, or extreme weather events (such as prolonged high temperatures) on pavement life.

[0052] In this embodiment, the decision-making unit is used to generate the final design scheme. This module's core objective is to minimize the total lifecycle cost, which includes: initial construction costs (material procurement, construction machinery, and labor costs), maintenance and repair costs throughout the entire lifecycle (based on predicted maintenance timing, processes, and materials), and traffic delay costs caused by pavement distress and maintenance work. Simultaneously, the module also considers constraints such as structural reliability (e.g., the probability of reaching the design life) and environmental impacts (e.g., carbon emissions, and the rate of recycled material utilization).

[0053] In some implementations, intelligent algorithms such as genetic algorithms and particle swarm optimization are employed to search for optimal solutions within a vast space of design parameters (such as the type, thickness, and bonding method of each layer). Genetic algorithms, by simulating natural selection and genetic mechanisms, iteratively generate and evaluate a series of design schemes until the optimal solution that satisfies all constraints and minimizes the total lifecycle cost is found. It can automatically evaluate the performance and economic benefits of different design schemes throughout their lifecycle and output optimal structural combination recommendations. For example, for a specific traffic load and climate condition, the optimization engine might recommend using a substructure made of recycled asphalt material to reduce initial costs and environmental impact, while compensating for potential performance losses by increasing the surface layer thickness, and pre-setting specific preventative maintenance in years 8 and 15.

[0054] In this embodiment, the decision-making process of the decision-making unit is represented by the following mathematical model: ; The constraints are: P ( LC 20 yeares) 0.90; RM Usage 0.20; in, C iIndicates the initial construction cost. C m It is the cost of maintenance and repair. C d It is the cost of traffic delays. P ( LC 20-years indicates the probability that a road surface will reach its 20-year design life. RM Usage This indicates the rate of recycling.

[0055] This embodiment ensures that the design scheme achieves a balance between economy, reliability, and environmental friendliness. It generates an optimal design scheme that takes into account total life cycle cost, structural reliability, and environmental impact, while also pre-recommending maintenance timing, maintenance processes, and maintenance materials.

[0056] like Figure 1 As shown, the design system in this embodiment also includes a 3D visualization and BIM / GIS integration module, which is used to overlay mechanical indicators onto the BIM model in the form of a 3D visualized cloud map.

[0057] The 3D visualization and BIM / GIS integration module seamlessly integrates refined design results into the 3D BIM model. GIS+BIM technology refers to a technical system that combines Geographic Information System (GIS) with Building Information Modeling (BIM). By integrating maps and design data, this technology places building structures within a broader geographical spatial context, supporting full lifecycle management from planning and design to construction and operation.

[0058] Different layers, material types, and thicknesses of the pavement structure are intuitively distinguished in the BIM model using colors and textures. Key mechanical response indicators, such as maximum horizontal tensile strain, maximum vertical compressive strain, and maximum shear stress, are overlaid on the BIM model as 3D visualization cloud maps, allowing users to visually see the most unfavorable stress and strain distribution locations and quickly assess the merits of design schemes. This module can also perform high-precision clash checks with GIS data such as underground pipelines and surrounding buildings, identifying design conflicts in advance and providing suggested solutions. All design parameters, calculation processes, and optimization results are traceable, generating structured 3D BIM models and 2D construction drawings, such as cross-sections, longitudinal sections, and structural details. To ensure BIM model interoperability, this module supports the IFC (Industry Foundation Classes) standard, enabling data exchange with other BIM software. Additionally, this module provides a VR / AR (Virtual Reality / Augmented Reality) interface, allowing users to review design schemes through an immersive experience.

[0059] This embodiment improves the accuracy of pavement life prediction by introducing a dynamic load spectrum when dealing with complex urban traffic environments. Through refined localized load spectrum construction, it accurately considers the spatiotemporal distribution and trends of different vehicle types, axle loads, and frequencies, as well as the performance fluctuations of materials under actual environmental conditions, effectively avoiding premature damage caused by model simplification in traditional methods. Simultaneously, it reduces the number of repairs and traffic interruptions caused by early-stage defects, proactively adapting to constantly changing traffic load characteristics and material properties, rather than simply designing based on static input parameters.

[0060] Example 2: This embodiment provides a method for designing urban road traffic load-adaptive pavement structures, based on the design system described in Embodiment 1, including: Dynamic load spectrum is constructed based on hidden Markov model; The road segment information is obtained, and the mechanical parameters of the pavement structure at different locations are calculated in combination with the dynamic load spectrum. The road segment information includes standard road segments and characteristic road segments. When calculating the mechanical parameters of the pavement structure at different locations, a layered calculation method is adopted: the mechanical parameters of the standard road segment are calculated through multi-layer elastic system units, and the mechanical parameters of the characteristic road segment are calculated through finite element analysis units, so as to achieve accurate adaptation calculation of mechanical parameters of different types of road segments. The mechanical parameters are mapped onto pavement design parameters. Through a collaborative optimization method involving multi-level digital twin models, the dynamic evolution of pavement performance is simulated to generate a pavement structure scheme adapted to road traffic loads. The multi-level digital twin model includes at least a pavement structure geometry model layer, a material property model layer, an environmental load model layer, and a construction process model layer. The models at each level interact in real time to form a coupled digital twin. By simulating the dynamic evolution of pavement performance throughout its entire life cycle, the collaborative optimization of the pavement structure scheme is achieved.

[0061] The above description is merely a preferred embodiment of this application and is not intended to limit this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the protection scope of this application.

Claims

1. A pavement structure design system adaptable to urban road traffic loads, characterized in that, include: The load data construction module is configured to: construct a dynamic load spectrum based on a hidden Markov model; The mechanical estimation module is configured to calculate the mechanical parameters of the pavement structure at different locations by combining the dynamic load spectrum and road segment information; The pavement structure generation module is configured to map the mechanical parameters onto pavement design parameters to generate a pavement structure scheme; the pavement structure generation module includes a collaborative optimization unit, which includes a multi-level digital twin model to simulate the dynamic evolution process of pavement performance.

2. The urban road traffic load adaptable pavement structure design system according to claim 1, characterized in that, The road segment information includes standard road segments and characteristic road segments. The mechanical estimation module includes a multi-layer elastic system element and a finite element analysis element. The multi-layer elastic system element is used to calculate the mechanical parameters of the standard road segment, and the finite element analysis element is used to calculate the mechanical parameters of the characteristic road segment.

3. The urban road traffic load-adaptive pavement structure design system according to claim 1, characterized in that, The multi-level digital twin model includes at least a road structure geometry model layer, a material property model layer, an environmental load model layer, and a construction process model layer. The models at each level interact with each other to form a coupled digital twin.

4. The urban road traffic load adaptable pavement structure design system according to claim 3, characterized in that, The multi-level digital twin model also includes a monitoring and feedback model layer, which is used to receive measured data and compare the measured data with the prediction results in the design stage to correct and optimize the design system parameters.

5. A pavement structure design system for adapting to urban road traffic loads according to any one of claims 1-4, characterized in that, The road structure generation module also includes a damage accumulation and life prediction unit and a decision unit. The output signals of the collaborative optimization unit and the damage accumulation and life prediction unit are all transmitted to the decision unit. The decision-making unit searches for the optimal solution in the design parameter space based on the constraints and outputs the final design scheme.

6. The urban road traffic load adaptable pavement structure design system according to claim 5, characterized in that, The damage accumulation and lifetime prediction unit integrates computational models based on at least two microstructural damage theories.

7. The urban road traffic load adaptable pavement structure design system according to claim 5, characterized in that, The decision-making process of the decision-making unit is represented by the following model: ； The constraints are: P ( LC 20 yeares) 0.90; RM Usage 0.20； in, C i Indicates the initial construction cost. C m It is the cost of maintenance and repair. C d It is the cost of traffic delays. P ( LC (20 years) indicates the probability that a road surface will reach its 20-year design life. RM Usage This indicates the rate of recycling.

8. The urban road traffic load adaptable pavement structure design system according to claim 1, characterized in that, The construction of the dynamic load spectrum based on the hidden Markov model includes: Hidden Markov models are used to describe the dynamic transition law of vehicle load state in order to construct a dynamic load spectrum. The Hidden Markov Model includes a load state set, an observation set, an initial state probability distribution, a state transition probability matrix, and an observation probability matrix. The load state set is a combination of key features related to vehicle load, and the observation set consists of vehicle type, axle type, axle load, and traffic volume information obtained through detection.

9. The urban road traffic load adaptable pavement structure design system according to claim 1, characterized in that, It also includes a 3D visualization and BIM / GIS integration module, which is used to overlay mechanical indicators onto the BIM model in the form of a 3D visualized cloud map.

10. A method for designing urban road traffic load-adaptive pavement structures, characterized in that, Based on the design system as described in any one of claims 1-9, it includes: Dynamic load spectrum is constructed based on hidden Markov model; Obtain road segment information and calculate the mechanical parameters of the pavement structure at different locations based on the dynamic load spectrum; The mechanical parameters are mapped onto the pavement design parameters, and a collaborative optimization method including multi-level digital twin models is used to simulate the dynamic evolution of pavement performance and generate a pavement structure scheme that adapts to road traffic loads.

Images