Life cycle-based road maintenance active intervention data processing method and system

Abstract

The application provides a road maintenance active intervention data processing method and system based on a whole life cycle, which comprises an Internet of Things sensing module, a digital twin base module and a whole-cycle asset database. The Internet of Things sensing module is composed of a structural health monitoring sensor deployed along a road, a pavement image acquisition device integrated on a patrol vehicle and a data interface connected to an engineering design and construction management system, and is used for automatically collecting asset properties, disease images, loads and environmental data in a whole process from a construction period to an operation period. The digital twin base module is connected with the Internet of Things sensing module, and comprises a data processing unit for cleaning, correlating and spatiotemporal positioning of multi-source data. The whole-cycle asset database stores a design BIM model, construction records, material archives and historical maintenance history based on unified coding correlation. A road three-dimensional twin model realizes synchronous visualization and query of road network level macroscopic display and component level microscopic state through BIM and GIS.

Description

Technical Field

[0001] This invention belongs to the field of road maintenance, and specifically relates to a method and system for proactive intervention data processing in road maintenance based on the entire life cycle. Background Technology

[0002] Currently, with the accelerating pace of urbanization and the ever-expanding transportation networks, the scale and complexity of road infrastructure continue to grow, highlighting the increasing importance of road maintenance. While some information technology has been introduced into road maintenance management, significant shortcomings remain in practical application, hindering improvements in management efficiency and service levels. Existing technological systems are generally fragmented and isolated; maintenance data at different stages is often managed by dispersed systems or departments, lacking unified data standards and integration platforms. This leads to a break in the information chain throughout the entire lifecycle—from planning, design, construction, operation, maintenance, to decommissioning—making it difficult to achieve coherent data traceability and in-depth data mining. In the data collection phase, much work still relies on manual inspections and paper records. This method is not only inefficient and has limited coverage, but also suffers from strong subjectivity, difficulty in guaranteeing accuracy, and a severe lack of real-time capability, failing to capture dynamic changes in road conditions in a timely manner, making preventative maintenance difficult to implement effectively. In terms of decision support, existing methods are mostly based on historical experience or simple statistical analysis, lacking in-depth fusion and intelligent analysis of multi-source heterogeneous data. This makes it impossible to construct accurate road performance degradation models and prediction models, resulting in insufficient scientific rigor in maintenance decisions. This often leads to untimely or excessive maintenance, affecting road lifespan and wasting resources. Resource scheduling and process management are also relatively rudimentary, failing to achieve refined matching and dynamic optimization of personnel, equipment, and materials. Transparent monitoring of the construction process is lacking, and quality and schedule control rely heavily on post-construction inspections, resulting in delayed problem detection and high rectification costs.

[0003] Barriers exist in collaboration among departments and stakeholders, information sharing is inefficient, and business processes are not tightly integrated, resulting in high communication costs, slow response times, and difficulty in forming an efficient collaborative management loop. At the technology application level, the integration of next-generation information technologies such as the Internet of Things, big data, and artificial intelligence with road maintenance operations is still superficial, failing to fully leverage their potential to achieve automatic status perception, intelligent risk warning, automatic solution generation, and optimized resource allocation. In summary, current road maintenance management has significant shortcomings in data integration, intelligent decision-making, process collaboration, and deep empowerment. There is an urgent need to build a data-driven, intelligent, and efficient digital management platform and methodology covering the entire lifecycle to integrate fragmented resources, break down information silos, and improve the precision, intelligence, and scientific level of maintenance management. This will ensure the safe, durable, and efficient operation of road infrastructure and meet the development needs of smart cities and modern transportation systems. Summary of the Invention

[0004] This invention proposes a data processing method and system for proactive intervention in road maintenance based on the entire life cycle. It solves the problems of data fragmentation, decision-making reliance on experience, lack of process traceability, and difficulties in cross-stage collaboration during road maintenance. By constructing a digital twin and intelligent decision-making system for the entire life cycle, it realizes the transformation of maintenance from passive response to proactive prediction and from extensive management to refined closed-loop management.

[0005] The technical solution of this invention is implemented as follows: a road maintenance proactive intervention data processing system based on the entire life cycle, comprising: The Internet of Things (IoT) sensing module consists of structural health monitoring sensors deployed along the road, road surface image acquisition devices integrated on inspection vehicles, and data interfaces that connect to the engineering design and construction management system. It is used to automatically collect asset attributes, defect images, load and environmental data throughout the entire process from the construction phase to the operation phase. The digital twin base module is connected to the IoT sensing module. The digital twin base module includes a data processing unit that cleans, correlates, and spatiotemporally locates multi-source data; a full-cycle asset database that stores design BIM models, construction records, material archives, and maintenance history based on unified coding; and a road 3D twin model that enables synchronous visualization and querying of road network-level macro-display and component-level micro-state through BIM and GIS. The intelligent maintenance decision-making module operates based on a digital twin base module. It includes a disease intelligent identification unit that uses computer vision algorithms to automatically analyze inspection images, identifying diseases such as cracks and potholes and quantifying their location, size, and severity level; a performance prediction and early warning unit that predicts pavement performance index and generates preventative maintenance recommendations based on historical monitoring data and degradation models; and a maintenance plan optimization unit that automatically matches a standard process library and generates a treatment plan containing specific repair methods, material requirements, and construction timing based on disease type, severity, traffic flow, and cost budget. The maintenance operation execution module is used to drive the management closed loop. The maintenance operation execution module includes a visual maintenance command center, which displays the status of maintenance work orders, resource locations and construction progress based on the road 3D twin model; a mobile inspection and operation terminal, which is used by on-site personnel to receive work orders, report defects, record the construction process and acceptance results; and a full-process tracking and feedback unit, which transmits on-site execution data back in real time and updates the asset database and the status of the road 3D twin model.

[0006] Existing road maintenance technologies suffer from numerous systemic flaws, the fundamental problem being the fragmentation of the entire lifecycle data chain and the isolation of management processes. Firstly, at the data level, data from different departments and systems manage each stage of road planning, design, construction, and operation, resulting in inconsistent formats and standards, creating severe "data silos." There is a lack of effective correlation and inheritance mechanisms between BIM models from the design phase, quality inspection records from the construction phase, and testing data from the operation and maintenance phase. This makes it difficult to trace the historical state and performance evolution of road assets, and maintenance decisions lack complete original data support. Secondly, in terms of condition perception and detection, existing technologies heavily rely on periodic manual inspections or simple instrument testing, resulting in low efficiency, limited coverage, and highly subjective and quantifiable results. For pavement defects such as cracks and potholes, traditional methods struggle to achieve rapid identification and location with millimeter-level precision, and cannot continuously monitor hidden conditions such as stress and strain within the structure in real time, leading to delayed defect detection and often missing the optimal intervention window.

[0007] Furthermore, in the maintenance decision-making process, existing methods are mostly based on fixed-cycle planned maintenance or emergency repairs after problems occur, lacking scientific predictive capabilities. The decision-making process relies heavily on the personal experience of engineers, making it difficult to comprehensively consider multiple factors such as the degradation patterns of road structural performance, real-time traffic loads, material performance degradation, environmental impact, and economic costs. This results in maintenance strategies that are either too conservative, leading to resource waste, or too aggressive, accelerating road damage, failing to achieve optimal cost-effectiveness throughout the entire life cycle.

[0008] In terms of maintenance execution and process management, the existing model suffers from a "management black box." From task assignment and on-site construction to quality acceptance, information transmission mainly relies on paper documents or fragmented electronic spreadsheets, resulting in opaque processes, uncontrollable progress, and difficulty in quality traceability. Managers struggle to monitor the location and status of on-site personnel, equipment, and materials in real time, and are unable to effectively supervise whether construction techniques comply with specifications, frequently leading to problems such as substandard maintenance quality, construction delays, and inadequate safety measures.

[0009] In terms of cross-departmental and cross-phase collaboration, the business connections between departments such as planning, design, construction, maintenance, and management are loose, and information sharing mechanisms are lacking. As a result, design defects and construction-related problems are often only discovered during maintenance, making it impossible to form a feedback and optimization closed loop from "construction" to "maintenance".

[0010] To address the aforementioned issues, the specific technical challenges this platform tackles include the challenges of integrating and managing multi-source heterogeneous data throughout its entire lifecycle. This requires designing a unified data coding standard and spatiotemporal indexing framework, overcoming the challenges of automatic cleaning, correlation matching, and dynamic updating of multi-dimensional information such as BIM, GIS, IoT time-series data, and business process data, and constructing a unified digital twin model that can both macroscopically display the road network situation and microscopically reveal the details of components.

[0011] The challenges of accurate identification and intelligent performance prediction of road defects based on artificial intelligence lie in improving the recognition rate, classification accuracy, and quantification accuracy of computer vision algorithms for various road defects under real-world road conditions such as complex lighting and dirt. Simultaneously, it is necessary to construct a hybrid model that integrates multiple factors, including material properties, environmental effects, and load history, to understand road performance degradation mechanisms and drive data, thereby achieving reliable prediction of medium- and long-term performance.

[0012] The dynamic optimization of maintenance plans under multi-objective constraints presents significant challenges. It requires establishing a knowledge base and decision-making model that integrates lifecycle cost analysis, traffic impact simulation, resource constraints, and process standards. Furthermore, it necessitates developing intelligent algorithms capable of rapidly finding the optimal solution from a vast array of possibilities to generate a balanced maintenance plan that considers cost, benefits, timeliness, and impact.

[0013] The challenges lie in achieving real-time transparent control and closed-loop feedback during maintenance operations. This requires enabling reliable operation of mobile terminals in offline and weak network environments, developing AR-based precise spatial positioning and information overlay technology, and designing control logic for automatic acquisition of key construction parameters, real-time anomaly alarms, and automatic locking of business processes to ensure the efficient operation of the "decision-execution-feedback" loop.

[0014] The challenges lie in cross-system, cross-platform, and cross-business process integration and collaboration. It requires building an open and scalable platform architecture, defining standard data interfaces and service buses, breaking down the barriers between existing independent systems, and achieving online and collaborative restructuring of planning, design, construction, maintenance, and management workflows.

[0015] As a preferred implementation, the maintenance scheme optimization unit automatically applies a matching maintenance strategy template based on the specific stage of the road asset's life cycle during actual operation. When the asset is determined to be in the preventive maintenance stage, it automatically recommends preventive processes and materials. When it is in the repair stage, it comprehensively considers structural strength data and traffic flow simulation, prioritizes structural repair schemes, and generates lane-specific and time-specific construction organization suggestions to directly guide the preparation of maintenance production plans.

[0016] As a preferred implementation, the maintenance strategy template is constructed based on road grade, traffic load spectrum and historical data on material durability, and integrates climate impact factor correction coefficients; when matching templates, the design standards, material ratios and maintenance records in the full-cycle asset database are called simultaneously to ensure that the recommended processes and materials are consistent with the original design and historical operating conditions of the asset.

[0017] As a preferred implementation, the logic for generating the lane-specific and time-specific construction organization suggestions is as follows: calling real-time traffic flow data and predicting the congestion index under different closure schemes through simulation models; with the goal of minimizing traffic impact, automatically selecting specific operation modes such as nighttime construction, alternating closures, or using the opposite lane, and simultaneously generating corresponding traffic diversion scheme diagrams and a list of required safety facilities.

[0018] As a preferred implementation, the mobile inspection and operation terminal is embedded with a standardized work order execution program. After receiving the task through the terminal, the on-site personnel use AR navigation to guide them to the exact location of the defect. During the operation, the terminal is forced to record the images and parameters of key processes such as excavation, paving and compaction step by step, and upload them in real time. The management personnel can simultaneously monitor the real-time images and progress of multiple work surfaces on the large screen in the command center.

[0019] As a preferred implementation, the AR navigation uses the geographic coordinates and component information of the road's three-dimensional twin model, combined with the terminal's real-time differential positioning and visual SLAM technology, to achieve centimeter-level on-site projection and three-dimensional overlay guidance of the defect locations, and simultaneously displays the historical maintenance records and internal structural cross-sectional views of the components to be worked on.

[0020] As a preferred implementation, the key process parameters for mandatory recording include material arrival temperature, paving thickness, and number of compaction passes. During the recording process, the terminal compares the real-time data with the threshold range in the preset process standard library. If the data exceeds the standard or the process is missing, an audible and visual alarm is immediately triggered on the terminal interface and the command center screen, and an anomaly report is generated, locking the current process until the management personnel handle it.

[0021] A proactive intervention data processing method for road maintenance based on the entire lifecycle, the method comprising the following steps: The data acquisition process involves automatically collecting multi-source data on asset attributes, structural status, defect images, loads, and environment throughout the entire process from construction to operation, through sensors deployed along the road, mobile inspection devices, and connected external systems. The data modeling and twin construction steps involve cleaning, associating, and spatiotemporally aligning the multi-source data. Based on a unified coding system, a road 3D digital twin model integrating BIM and GIS technologies is constructed and continuously updated. This model integrates and stores planning and design models, construction archives, monitoring data, and maintenance history, enabling traceability and visualization of the asset's status throughout its entire lifecycle. The intelligent analysis and decision-making process, based on the data in the digital twin model, performs the following operations: automatically analyzes inspection images using computer vision algorithms to identify and quantify pavement defects; predicts the degradation trend of key pavement performance indicators and generates early warnings based on historical and real-time data using machine learning models; and, in response to early warnings or planning requirements, invokes a decision-making model with full life-cycle cost, performance, and traffic impact as optimization objectives to automatically match the process library and generate a refined maintenance plan that includes specific repair methods, materials, construction timing, and traffic organization. The task execution and closed-loop feedback steps decompose the maintenance plan into electronic work orders and distribute them to mobile terminals; guide on-site personnel to complete precise positioning construction based on AR navigation through the terminals, and forcibly record key process parameters and images; transmit construction process and acceptance data back in real time, and update the asset status and maintenance records in the digital twin model; The model optimization step involves dynamically correcting the parameters of the performance prediction model and the decision model based on the maintenance effect data from the closed-loop feedback, thereby achieving continuous adaptive optimization of the maintenance strategy.

[0022] As a preferred implementation, in the intelligent analysis and decision-making step, the road section to be maintained is automatically divided into preventive, restorative, or reconstructive maintenance zones based on the current performance data and traffic load spectrum in the digital twin model; for each zone, at least two candidate schemes are matched from the process library, taking into account the material inventory status and cost constraints; the impact of each scheme on the road network at different construction periods is simulated through a traffic simulation model, the scheme with the highest comprehensive score is selected, and a detailed construction organization plan including nighttime operation periods, lane-level closure plans, and temporary signage layout diagrams is generated simultaneously.

[0023] As a preferred implementation, the specific steps of automatically dividing maintenance intervals are as follows: based on the current values ​​and degradation rates of the pavement performance index and structural strength index provided by the digital twin model, combined with preset grading thresholds and traffic load levels, a gridded maintenance interval map containing geographic spatial boundaries, dominant disease types, and recommended maintenance modes is automatically generated; within each interval, based on pavement structure layer detection data and historical maintenance records, the technical necessity score for different intervention depths such as foundation reinforcement, surface function restoration, or structural reconstruction is further calculated, serving as the core input parameter for subsequent scheme matching.

[0024] After adopting the above technical solution, the beneficial effects of this invention are as follows: The benefits brought by this platform are comprehensive and systematic, significantly improving the scientific, accurate, and economical nature of road infrastructure management. Firstly, in terms of management efficiency, the platform achieves "one-map overview and one-code traceability" of road assets by constructing a digital twin spanning the entire lifecycle. Managers can query design drawings, construction records, maintenance archives, and current real-time status of any road section or component at any time, completely changing the previous situation of relying on paper archives and fragmented information, greatly improving the information support level for management decisions and the ability to trace historical issues.

[0025] In terms of scientific maintenance, the platform has achieved a fundamental shift from "experience-driven" to "data and model-driven." AI-based intelligent defect identification significantly improves inspection efficiency and objectivity, enabling more timely defect detection and more accurate descriptions. Meanwhile, performance prediction models can provide early warnings of potential risks, driving the maintenance model from a passive "repair after damage" approach to a proactive "prevention before damage" approach. This allows for early intervention at a lower cost, effectively delaying road performance degradation and extending service life. Furthermore, in terms of decision optimization, the platform's intelligent decision-making module comprehensively considers technical, economic, and social objectives. It can not only recommend optimal repair solutions for individual defects but also optimize multiple maintenance needs from a holistic road network perspective, coordinating construction schedules, closure plans, and resource allocation. This maximizes overall road network performance improvement within a limited budget while minimizing disruption to traffic, enhancing the public's travel experience.

[0026] In terms of process control, the platform achieves "transparent" management of maintenance operations through the linkage between mobile terminals and a visual command center. Managers can remotely monitor the progress, personnel locations, process parameters, and safety conditions at multiple work sites in real time, and promptly correct any non-standard operations. Mandatory recording of process data and image retention create a complete electronic quality archive, making construction quality traceable and assessable, effectively guaranteeing the quality and safety of maintenance projects.

[0027] In terms of economic benefits, the platform's application has brought significant direct and indirect cost savings. Preventative maintenance reduces the occurrence of large-scale structural damage and lowers costly reconstruction expenses; scientific decision-making avoids over-maintenance or ineffective maintenance and optimizes material and manpower allocation; and refined process control reduces rework and project delays. From a life-cycle perspective, the digital assets accumulated during the initial construction phase of the platform provide continuous value for decades of operation and maintenance, achieving the long-term goal of cost reduction and efficiency improvement. This platform not only enhances the technical level and management efficiency of road maintenance itself, but also creates enormous comprehensive benefits for society by extending facility lifespan, ensuring traffic safety, and optimizing resource utilization, strongly supporting the development of smart cities and sustainable transportation infrastructure. Attached Figure Description

[0028] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0029] Figure 1 This is a system flowchart of the present invention. Detailed Implementation

[0030] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0031] Example: Figure 1 As shown, the road maintenance proactive intervention data processing system based on the entire life cycle includes... The Internet of Things (IoT) sensing module consists of structural health monitoring sensors deployed along the road, road surface image acquisition devices integrated on inspection vehicles, and data interfaces that connect to the engineering design and construction management system. It is used to automatically collect asset attributes, defect images, load and environmental data throughout the entire process from the construction phase to the operation phase. The digital twin base module is connected to the IoT sensing module. The digital twin base module includes a data processing unit that cleans, correlates, and spatiotemporally locates multi-source data; a full-cycle asset database that stores design BIM models, construction records, material archives, and maintenance history based on unified coding; and a road 3D twin model that enables synchronous visualization and querying of road network-level macro-display and component-level micro-state through BIM and GIS. The intelligent maintenance decision-making module operates based on a digital twin base module. It includes a disease intelligent identification unit that uses computer vision algorithms to automatically analyze inspection images, identifying diseases such as cracks and potholes and quantifying their location, size, and severity level; a performance prediction and early warning unit that predicts pavement performance index and generates preventative maintenance recommendations based on historical monitoring data and degradation models; and a maintenance plan optimization unit that automatically matches a standard process library and generates a treatment plan containing specific repair methods, material requirements, and construction timing based on disease type, severity, traffic flow, and cost budget. The maintenance operation execution module is used to drive the management closed loop. The maintenance operation execution module includes a visual maintenance command center, which displays the status of maintenance work orders, resource locations and construction progress based on the road 3D twin model; a mobile inspection and operation terminal, which is used by on-site personnel to receive work orders, report defects, record the construction process and acceptance results; and a full-process tracking and feedback unit, which transmits on-site execution data back in real time and updates the asset database and the status of the road 3D twin model.

[0032] In practical implementation, this platform first constructs an all-weather data acquisition network through IoT sensing modules. In the operation and maintenance scenario of an urban expressway, distributed fiber optic sensors and strain gauges deployed in bridge bearings and pavement structural layers continuously monitor structural stress, displacement, and vibration frequency. Automated inspection vehicles equipped with high-definition line-scan cameras and LiDAR regularly patrol, collecting high-resolution surface images and 3D point cloud data of the entire road section. Simultaneously, the platform automatically connects to the design unit's BIM collaboration platform and the construction unit's project management system through standardized interfaces to obtain the road's original design model, construction process records, and material certification archives. This multi-source heterogeneous data from physical monitoring, mobile inspection, and historical systems constitutes a multi-dimensional data stream covering asset attributes, defect characteristics, load effects, and environmental impacts throughout the entire process from construction to operation, providing real-time and continuous raw information input for subsequent processing.

[0033] All collected data is fed into the digital twin base module in real time. Its data processing unit immediately cleans, corrects, and formats the incoming data stream, such as filtering sensor noise and compensating for image distortion. Based on a unified spatiotemporal coding rule, each data segment is precisely associated with a specific road station, lane, and even component. The treated data is synchronously integrated into a full-cycle asset database. This database uses the unique code of each road component as an index to dynamically associate its design geometric parameters, construction and acceptance records, maintenance archives, and current monitoring status, forming a complete digital history. Based on this, the platform constructs and continuously updates a 3D digital twin model of the road integrating BIM and GIS technologies. This model can not only display the overall performance heat map of the road network on an electronic map at a macro level, but also display the tilt history of a specified bridge pier or the crack development process of a certain pavement slab at a micro level through lightweight 3D components, realizing a dynamic, two-way, and refined mapping between the physical road and its virtual mirror.

[0034] Based on the vibrant and integrated data environment provided by the digital twin model, the intelligent maintenance decision-making module initiates the analysis process. The intelligent defect identification unit uses computer vision algorithms to automatically analyze the latest inspection images, accurately identifying defects such as cracks and potholes at specific locations (e.g., K5+230), and quantifying their size, density, and severity level. Simultaneously, the performance prediction and early warning unit integrates years of load, environmental, and historical performance data for this road section, predicting the evolution trend of the pavement service performance index through a machine learning degradation model, and automatically generating preventative maintenance warnings before the performance value approaches the intervention threshold. Subsequently, the maintenance plan optimization unit responds to the warning, comprehensively considering defect details, real-time traffic flow, material inventory, and cost budget, and uses a multi-objective optimization algorithm to select candidate solutions from the standard process library. It also evaluates the traffic impact of different construction plans through traffic simulation, ultimately outputting a refined treatment plan that includes recommended processes (e.g., thin overlay), specific material specifications, nighttime construction periods, lane-level closure plans, and detailed traffic organization schemes.

[0035] The decision-making plan is implemented through the maintenance business execution module. In the visualized maintenance command center, the plan is transformed into specific work orders, which are highlighted on a digital twin model. The work orders are dispatched to the mobile work terminals of maintenance personnel via wireless network. On-site personnel, guided by the augmented reality navigation (AR navigation) provided by the terminal, accurately arrive at the defect locations through the superimposed 3D arrows and highlighted outlines on the screen. During construction, the terminal enforces standard procedures, guiding operators to record or automatically collect key parameters such as mixture temperature, paving thickness, and number of compaction passes, and requires the uploading of images of each completed process. All data is transmitted back in real time, and the command center's large screen can simultaneously monitor the progress, personnel location, process parameters, and on-site images of multiple work areas, achieving transparent and remote supervision of the construction process.

[0036] The end-to-end tracking and feedback unit is responsible for finalization and iteration. After construction is completed and accepted, all process data, quality inspection results, and completion images are systematically transmitted back and updated to the maintenance history of the corresponding road segment in the full-cycle asset database. Simultaneously, this drives the digital twin model to update the road segment's state attributes and geometric representation. Finally, the effectiveness data of this maintenance operation (such as performance improvements monitored over a period after maintenance) is fed back to the intelligent maintenance decision-making module for parameter correction and self-learning of the performance prediction model and decision optimization algorithm. Thus, the platform completes a closed-loop lifecycle from "state perception, intelligent diagnosis, optimized decision-making, and precise execution" to "effect evaluation and model optimization," continuously improving prediction accuracy and decision-making scientific rigor during ongoing operation.

[0037] Specific application scenarios are as follows: This example uses the maintenance process of a section (from chainage K10+500 to K10+600) of an urban arterial road (designed as a Class I highway with six lanes in both directions) that has been in operation for more than 8 years to illustrate the workflow of this platform.

[0038] In the IoT sensing module, the periodic detection data from the deflection sensors deployed on this road section showed a small but continuous upward trend. Simultaneously, images captured by high-definition cameras integrated on inspection vehicles, analyzed by the intelligent defect identification unit, identified continuous and developing fine longitudinal cracks in the third lane of this road section. The digital twin base module synchronized this data and marked the road section as "attention" in the road's 3D twin model. Based on the road section's traffic load spectrum over the past 8 years, environmental temperature cycle data, and current monitoring values, the performance prediction and early warning unit calculated using an embedded asphalt pavement performance degradation model, predicting that its pavement structural strength index would drop below a predetermined threshold within the next 6 months. The platform then automatically generated a "preventive maintenance early warning."

[0039] The maintenance plan optimization unit responded to the warning. First, based on the current performance data and design life (15 years) of the road section, the system determined that it was in the "preventive maintenance stage." Then, it automatically retrieved a maintenance strategy template that matched the road's grade (Class I highway), current dominant traffic load (20% heavy vehicles), and local climate zone (temperate monsoon climate, integrating freeze-thaw cycle influence coefficient). This template clearly stated that the core objectives of this stage were "water sealing, skid resistance, and delaying performance degradation." Simultaneously, the system queried the full-cycle asset database, confirming that the original surface layer design for this road section was 4 cm SMA-13 ​​asphalt mixture, and that there were no structural repair records in the past three years.

[0040] Based on the above analysis, the maintenance scheme optimization unit pre-selected two candidate processes from the standard process library that conform to the preventive maintenance concept: "micro-surfacing" and "ultra-thin wearing layer (UTFC)". Next, the unit used real-time traffic flow data (including average daily traffic volume on weekdays and peak hour traffic volume) and a built-in traffic simulation sub-model to simulate the impact of the two processes on the road segment's service level at different construction times (e.g., closing one lane for 6 hours throughout the day and closing two lanes for 8 hours at night). Simulation calculations showed that although the material cost per square meter of the "ultra-thin wearing layer (UTFC)" process is slightly higher, it requires a shorter construction window (only two nighttime periods, with one lane closed for 5 hours each night), resulting in the lowest impact on daytime peak traffic. The optimization algorithm weighed the "optimal cost-effectiveness throughout the entire life cycle" against the "minimum traffic disruption during construction" as a comprehensive objective, ultimately deciding to adopt the "UTFC-10 type ultra-thin wearing layer" as the recommended process.

[0041] After the decision is made, the system automatically generates a detailed handling plan, which includes: specific repair methods: milling 1 cm of the original road surface, spraying modified emulsified asphalt tack coat, and paving 1.0 cm thick UTFC-10 mixture; material requirements: technical indicators such as the design void ratio, aggregate gradation, and asphalt grade of the mixture; construction timing: it is recommended to carry out the work from 23:00 to 04:00 the following Tuesday and Wednesday night; construction organization suggestions: close the third lane for construction, use part of the width of the second lane to set up a safety buffer zone, and simultaneously generate a traffic diversion plan including a map of the placement of cones and signs, as well as a list of required facilities.

[0042] The generated maintenance work orders are created and displayed on the digital twin model of the visualized maintenance command center. After the dispatcher reviews them, the work orders are distributed to the mobile inspection and operation terminals of the designated maintenance teams via wireless network.

[0043] On Tuesday night, the maintenance team arrived at the site. The workers activated the standard work order execution program on the terminal. The terminal, using integrated real-time differential positioning and visual SLAM technology, combined with the precise geographic coordinates and component models of the road section downloaded from the platform's 3D road twin model, displayed a clear, perfectly matching 3D outline of the construction area's 3D boundary overlaid on the camera's live feed in augmented reality (AR) format. Simultaneously, the terminal's sidebar displayed the last maintenance record for the road section (micro-surfacing work from 3 years ago) and a cross-sectional view of the road structure, providing background reference for the work.

[0044] Before the milling process begins, a mandatory data recording form pops up on the terminal interface. The operator uses the terminal to scan the electronic tag on the UTFC-10 modified asphalt mixture truck that arrives at the site; the mixture type, batch number, and factory temperature are automatically read and entered. During paving, the terminal automatically records the paving thickness (target value 1.0 cm, actual recorded 1.02 cm) via a connected ultrasonic sensor. During compaction, the operator uses a smart terminal installed on each roller to automatically record the number of compaction passes (2 passes for initial compaction, 4 passes for secondary compaction, and 2 passes for final compaction). Simultaneously, the team safety officer uses the terminal to capture and upload panoramic images of the site at key process nodes (such as after tack coat spraying, after paving, and after final compaction).

[0045] All data is transmitted back to the platform in real time via the network. On the command center's large screen, managers can see the progress bar of the work order moving forward in real time, while simultaneously monitoring multiple split screens: one is a real-time video stream from the site, and another is a dynamically updated process parameter dashboard. Once the number of compaction passes meets the standard, the system automatically verifies the paving thickness and material temperature. If an anomaly is detected in the paving thickness data for a certain period (for example, a recorded value of 1.5 cm, exceeding the allowable error range), the system will immediately display a flashing red warning on the terminal screen of the on-site workers and sound a buzzer alarm; simultaneously, on the command center's large screen, the work order status immediately changes from "in progress" to "abnormal," triggering an audible and visual alarm and displaying a "paving thickness exceeds standard" anomaly report.

[0046] Upon receiving the alarm, the on-site technical supervisor immediately checked and confirmed that the false alarm was caused by a temporary sensor error. After calibration, the measurement was repeated and confirmed on the terminal. After management personnel remotely confirmed the handling measures at the command center, the system deactivated the alarm and the work order status, and the work order status was restored. After all construction was completed, the team submitted a "completion application" through the terminal and uploaded the final panoramic photo of the completed work.

[0047] The end-to-end tracking and feedback unit will fully archive all data packages for this maintenance operation—including the plan, actual process parameters, process images, anomaly handling records, and acceptance photos—and update them to the maintenance history of this road section (chainage K10+500 to K10+600) in the full-cycle asset database. Simultaneously, the road's 3D twin model will automatically update the pavement material properties of this section to "UTFC-10 ultra-thin wearing layer," and the surface texture will also be updated accordingly.

[0048] The platform will continuously track the deflection, smoothness, and crack development data of this road section over the next few months. This post-maintenance data will be used to optimize the parameters of the degradation model in the performance prediction and early warning unit, thereby providing a more accurate basis for preventive maintenance decisions for similar roads in the future, completing a closed-loop management process covering the entire lifecycle from intelligent early warning, optimized decision-making, precise execution to effect evaluation and model optimization.

[0049] A proactive intervention data processing method for road maintenance based on the entire lifecycle, the method comprising the following steps: The data acquisition process involves automatically collecting multi-source data on asset attributes, structural status, defect images, loads, and environment throughout the entire process from construction to operation, through sensors deployed along the road, mobile inspection devices, and connected external systems. The data modeling and twin construction steps involve cleaning, associating, and spatiotemporally aligning the multi-source data. Based on a unified coding system, a road 3D digital twin model integrating BIM and GIS technologies is constructed and continuously updated. This model integrates and stores planning and design models, construction archives, monitoring data, and maintenance history, enabling traceability and visualization of the asset's status throughout its entire lifecycle. The intelligent analysis and decision-making process, based on the data in the digital twin model, performs the following operations: automatically analyzes inspection images using computer vision algorithms to identify and quantify pavement defects; predicts the degradation trend of key pavement performance indicators and generates early warnings based on historical and real-time data using machine learning models; and, in response to early warnings or planning requirements, invokes a decision-making model with full life-cycle cost, performance, and traffic impact as optimization objectives to automatically match the process library and generate a refined maintenance plan that includes specific repair methods, materials, construction timing, and traffic organization. The task execution and closed-loop feedback steps decompose the maintenance plan into electronic work orders and distribute them to mobile terminals; guide on-site personnel to complete precise positioning construction based on AR navigation through the terminals, and forcibly record key process parameters and images; transmit construction process and acceptance data back in real time, and update the asset status and maintenance records in the digital twin model; The model optimization step involves dynamically correcting the parameters of the performance prediction model and the decision model based on the maintenance effect data from the closed-loop feedback, thereby achieving continuous adaptive optimization of the maintenance strategy.

[0050] In the intelligent analysis and decision-making steps, based on the current performance data and traffic load spectrum in the digital twin model, the road section to be maintained is automatically divided into preventive, restorative, or reconstructive maintenance zones. For each zone, in combination with material inventory status and cost constraints, no fewer than two candidate schemes are matched from the process library. The impact of each scheme on the road network at different construction periods is simulated through a traffic simulation model, the scheme with the highest comprehensive score is selected, and a detailed construction organization plan including nighttime operation periods, lane-level closure plans, and temporary signage layout diagrams is generated simultaneously.

[0051] The specific steps for automatically dividing maintenance intervals are as follows: based on the current values ​​and degradation rates of the pavement performance index and structural strength index provided by the digital twin model, combined with preset grading thresholds and traffic load levels, a gridded maintenance interval map containing geographic spatial boundaries, dominant disease types, and recommended maintenance modes is automatically generated; within each interval, based on pavement structure layer detection data and historical maintenance records, the technical necessity score for different intervention depths such as foundation reinforcement, surface function restoration, or structural reconstruction is further calculated, serving as the core input parameter for subsequent scheme matching.

[0052] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A road maintenance proactive intervention data processing system based on the entire life cycle, characterized in that, include The Internet of Things (IoT) sensing module consists of structural health monitoring sensors deployed along the road, road surface image acquisition devices integrated on inspection vehicles, and data interfaces that connect to the engineering design and construction management system. It is used to automatically collect asset attributes, defect images, load and environmental data throughout the entire process from the construction phase to the operation phase. The digital twin base module is connected to the IoT sensing module. The digital twin base module includes a data processing unit that cleans, correlates, and spatiotemporally locates multi-source data; a full-cycle asset database that stores design BIM models, construction records, material archives, and maintenance history based on unified coding; and a road 3D twin model that enables synchronous visualization and querying of road network-level macro-display and component-level micro-state through BIM and GIS. The intelligent maintenance decision-making module operates based on the digital twin base module. The intelligent maintenance decision-making module includes a disease intelligent identification unit, which uses computer vision algorithms to automatically analyze inspection images, identify diseases such as cracks and potholes, and quantify their location, size and severity level; and a performance prediction and early warning unit, which predicts the pavement service performance index and generates preventive maintenance suggestions in advance based on historical monitoring data and degradation models. The maintenance plan optimization unit automatically matches a standard process library and generates a treatment plan that includes specific repair methods, material requirements, and construction timing based on the type and severity of the damage, traffic flow, and cost budget. The unit also automatically applies matching maintenance strategy templates based on the specific stage of the road asset's lifecycle. When the asset is determined to be in the preventative maintenance stage, preventative processes and materials are automatically recommended. When it is in the repair stage, structural strength data and traffic flow simulation are used to prioritize structural repair plans and generate lane-specific and time-based construction organization suggestions to directly guide the preparation of maintenance production plans. The maintenance operation execution module is used to drive the management closed loop. The maintenance operation execution module includes a visual maintenance command center, which displays the status of maintenance work orders, resource locations and construction progress based on the road 3D twin model; a mobile inspection and operation terminal, which is used by on-site personnel to receive work orders, report defects, record the construction process and acceptance results; and a full-process tracking and feedback unit, which transmits on-site execution data back in real time and updates the asset database and the status of the road 3D twin model.

2. The road maintenance proactive intervention data processing system based on the entire life cycle as described in claim 1, characterized in that: The maintenance strategy template is constructed based on road grade, traffic load spectrum and historical data on material durability, and integrates climate impact factor correction coefficients. When performing template matching, the design standards, material ratios, and maintenance records in the full-cycle asset database are called up simultaneously to ensure that the recommended processes and materials are consistent with the original design and historical operating conditions of the asset.

3. The road maintenance proactive intervention data processing system based on the entire life cycle as described in claim 1, characterized in that: The logic for generating lane-specific and time-specific construction organization suggestions is as follows: real-time traffic flow data is called, and the congestion index under different closure schemes is predicted through simulation models; with the goal of minimizing traffic impact, specific operation modes such as nighttime construction, alternating closures, or using the opposite lane are automatically selected, and corresponding traffic diversion scheme diagrams and lists of required safety facilities are generated simultaneously.

4. The road maintenance proactive intervention data processing system based on the entire life cycle as described in claim 1, characterized in that: The mobile inspection and operation terminal is embedded with a standardized work order execution program. After receiving the task through the terminal, the on-site personnel use AR navigation to guide them to the exact location of the defect. During the operation, the terminal is forced to record the images and parameters of key processes such as excavation, paving and compaction step by step and upload them in real time. The management personnel can simultaneously monitor the real-time images and progress of multiple work surfaces on the large screen in the command center.

5. The road maintenance proactive intervention data processing system based on the entire life cycle as described in claim 4, characterized in that: The AR navigation system uses the geographic coordinates and component information of the road's 3D twin model, combined with the terminal's real-time differential positioning and visual SLAM technology, to achieve centimeter-level on-site projection and 3D overlay guidance of defect locations, and simultaneously displays the historical maintenance records and internal structural cross-sectional views of the components to be worked on.

6. The road maintenance proactive intervention data processing system based on the entire life cycle as described in claim 4, characterized in that: The key process parameters that the terminal is required to record step by step include the material arrival temperature, paving thickness and number of compaction passes. During the recording process, the terminal compares the real-time data with the threshold range in the preset process standard library. If the data exceeds the standard or the process is missing, the terminal will immediately trigger an audible and visual alarm on the terminal interface and the command center screen, generate an anomaly report, and lock the current process until the management personnel take action.

7. A data processing method for proactive intervention in road maintenance based on the entire life cycle, characterized by: The method includes the following steps: The data acquisition process involves automatically collecting multi-source data on asset attributes, structural status, defect images, loads, and environment throughout the entire process from construction to operation, through sensors deployed along the road, mobile inspection devices, and connected external systems. The data modeling and twin construction steps involve cleaning, associating, and spatiotemporally aligning the multi-source data. Based on a unified coding system, a road 3D digital twin model integrating BIM and GIS technologies is constructed and continuously updated. This model integrates and stores planning and design models, construction archives, monitoring data, and maintenance history, enabling traceability and visualization of the asset's status throughout its entire lifecycle. The intelligent analysis and decision-making process, based on the data in the digital twin model, performs the following operations: automatically analyzes inspection images using computer vision algorithms to identify and quantify pavement defects; predicts the degradation trend of key pavement performance indicators and generates early warnings based on historical and real-time data using machine learning models; and, in response to early warnings or planning requirements, invokes a decision-making model with full life-cycle cost, performance, and traffic impact as optimization objectives to automatically match the process library and generate a refined maintenance plan that includes specific repair methods, materials, construction timing, and traffic organization. The task execution and closed-loop feedback steps decompose the maintenance plan into electronic work orders and distribute them to mobile terminals; guide on-site personnel to complete precise positioning construction based on AR navigation through the terminals, and forcibly record key process parameters and images; transmit construction process and acceptance data back in real time, and update the asset status and maintenance records in the digital twin model; The model optimization step involves dynamically correcting the parameters of the performance prediction model and the decision model based on the maintenance effect data from the closed-loop feedback, thereby achieving continuous adaptive optimization of the maintenance strategy.

8. The data processing method for proactive intervention in road maintenance based on the entire life cycle as described in claim 7, characterized in that: In the intelligent analysis and decision-making steps, the road section to be maintained is automatically divided into preventive, restorative, or reconstructive maintenance zones based on the current performance data and traffic load spectrum in the digital twin model. For each section, considering material inventory status and cost constraints, at least two candidate solutions are matched from the process library; the impact of each solution on the road network at different construction periods is simulated through a traffic simulation model, the solution with the highest comprehensive score is selected, and a detailed construction organization plan including nighttime operation periods, lane-level closure plans, and temporary signage layout is generated simultaneously.

9. The data processing method for proactive intervention in road maintenance based on the entire life cycle as described in claim 8, characterized in that: The specific steps for automatically dividing the road sections to be maintained are as follows: based on the current values ​​and degradation rates of the pavement performance index and structural strength index provided by the digital twin model, combined with preset grading thresholds and traffic load levels, a gridded maintenance interval map containing geographic spatial boundaries, dominant disease types, and recommended maintenance modes is automatically generated; within each interval, based on pavement structure layer detection data and historical maintenance records, the technical necessity score for different intervention depths such as foundation reinforcement, surface function restoration, or structural reconstruction is further calculated, serving as the core input parameter for subsequent scheme matching.

Images