Meteorological factor classification and quantification system and method for roadbed climate resilience

Abstract

This invention discloses a meteorological factor classification and quantification characterization system and method for roadbed climate resilience. The system includes a multi-source heterogeneous data intelligent fusion module, a multi-scale meteorological factor classification and quantification module, and a roadbed resilience application interface module connected in sequence. The multi-source heterogeneous data intelligent fusion module is used to access and collaboratively process multi-source heterogeneous data, outputting roadbed digital environment base data with high spatiotemporal integrity. The multi-scale meteorological factor classification and quantification module is used to analyze, classify, and quantify the features of the roadbed digital environment base data, generating a standardized dataset. The roadbed resilience application interface module is used to encapsulate and output the quantified standardized dataset according to the needs of downstream engineering application scenarios. This invention achieves a deep transformation from "weather data" to "engineering stress signals," enabling analysis to move from the statistical correlation level to the physical mechanism level, providing more targeted quantitative basis for roadbed engineering analysis.

Description

Technical Field

[0001] This invention relates to the fields of transportation infrastructure engineering, plateau meteorology and disaster prevention and mitigation, and specifically to a meteorological factor classification and quantitative characterization system and method for roadbed climate resilience. Background Technology

[0002] The harsh climate and unique meteorological conditions in special regions such as plateaus, including strong radiation, drastic temperature differences, and complex precipitation patterns, are the main driving forces behind roadbed subsidence, frost heave, and cracking. Improving the climate resilience of roadbeds requires, first and foremost, the precise quantification of these meteorological drivers.

[0003] Current mainstream methods have significant limitations: First, data application is superficial. Engineering designs often directly use raw data or simple statistical values ​​from meteorological stations. Such data is not linked to the thermodynamic and hydraulic response processes of roadbed materials, and cannot distinguish the essential differences in the impact of meteorological phenomena of different durations and intensities on the roadbed. Second, regional characteristics are lacking. General meteorological indicators are difficult to effectively characterize the engineering hazards of complex and distinctive meteorological events unique to the plateau, such as "short-term strong convective hail," "spring snowmelt floods," and "freeze-thaw cycles accompanied by wind erosion." Third, the model input barrier is high. Existing meteorological data products do not match the input format required by roadbed water-thermal-mechanical coupled numerical models or machine learning early warning models in terms of spatiotemporal resolution and parameter types, forming a "data gap."

[0004] Existing patent CN115019474A discloses a method for early warning of meteorological risks of geological disasters, which provides early warning by analyzing the relationship between rainfall and geological disasters. However, it does not involve the mechanistic classification and quantification of multiple meteorological factors for the internal physical processes of engineering structures, and in particular, it does not consider the specific impacts of plateau-specific meteorological events on roadbed structures. Therefore, developing a dedicated characterization system that can bridge meteorological science and roadbed engineering and deeply quantify regional climate stress has significant theoretical value and urgent engineering need. Summary of the Invention

[0005] The purpose of this invention is to address the shortcomings of existing technologies by providing a classification and quantitative characterization system and method for meteorological factors related to roadbed climate resilience. This invention primarily constructs a classification and quantitative system oriented towards engineering physical processes. Through the definition and calculation of "engineering impact equivalent," it breaks through the traditional model of directly referencing raw meteorological data, achieving a deep transformation from "weather data" to "engineering stress signals." This allows the analysis to delve from the statistical correlation level to the physical mechanism level, providing a more targeted quantitative basis for roadbed engineering analysis.

[0006] To achieve the above objectives, the present invention adopts the following technical solution:

[0007] A meteorological factor classification and quantification characterization system for roadbed climate resilience includes a multi-source heterogeneous data intelligent fusion module, a multi-scale meteorological factor classification and quantification module, and a roadbed resilience application interface module connected in sequence.

[0008] The multi-source heterogeneous data intelligent fusion module is used to access and collaboratively process data from ground-based meteorological stations, highway traffic monitoring stations, multi-source satellite remote sensing data, UAV aerial survey data, and numerical model reanalysis data. The multi-source heterogeneous data intelligent fusion module has built-in data correction algorithms for complex plateau terrain and a data fusion model based on spatiotemporal attention mechanism, and outputs road area digital environment base data with high spatiotemporal integrity.

[0009] The multi-scale meteorological factor classification and quantification module is used to analyze, classify and quantify the features of road area digital environment base data, and generate a standardized dataset.

[0010] The roadbed resilience application interface module is used to encapsulate and output the standardized dataset generated by quantification according to the requirements of downstream engineering application scenarios.

[0011] In this invention, the data correction algorithm built into the multi-source heterogeneous data intelligent fusion module for complex plateau terrain is based on the characteristics of the plateau region's dramatic terrain undulations and strong spatial heterogeneity of meteorological elements. It adopts a multi-source data cross-validation strategy and constructs a dynamic correction model by fusing terrain elevation data and historical meteorological statistics. Compared with traditional data correction algorithms, it introduces the influence weight coefficients of terrain slope and aspect on meteorological elements, and performs special corrections for special meteorological phenomena such as temperature inversion layers and valley winds common in the plateau, effectively eliminating data bias caused by terrain shading and vertical gradient changes.

[0012] The multi-source heterogeneous data intelligent fusion module incorporates a spatiotemporal attention-based data fusion model specifically designed for complex plateau terrain. This model addresses the challenge of uneven spatiotemporal distribution of data collection in plateau regions by constructing a three-layer attention mechanism architecture. In the temporal dimension, a Long Short-Term Memory (LSTM) network is used to capture the dynamic changes in meteorological elements. In the spatial dimension, a Graph Attention Network (GAT) is used to analyze the propagation relationships of meteorological elements under complex terrain. A terrain feature enhancement module is also added, embedding Digital Elevation Model (DEM) data into the attention calculation process. Compared to conventional spatiotemporal fusion models, this model can adaptively allocate weights from different data sources at various spatiotemporal nodes, significantly improving the completeness and accuracy of data fusion under complex terrain conditions.

[0013] As a further explanation of the present invention, the multi-scale meteorological factor classification and quantification module includes a mechanistic classification unit, an engineering impact equivalent calculation unit, a regional characteristic event quantification sub-module, and a composite climate stress construction sub-module.

[0014] The mechanistic classification unit constructs a three-dimensional classification framework of "dynamic-thermal-moisture" based on the main physical processes by which meteorological factors affect the roadbed structure. This framework categorizes original meteorological elements (including thermal, dynamic, and moisture elements, as well as other elements) into meteorological factor categories. Ground-based meteorological station data provides high-precision real-time meteorological element observations, directly supporting the classification of thermal, dynamic, and moisture elements. Data from dedicated highway traffic monitoring stations acquires meteorological environmental parameters closely related to the roadbed, aiding in the classification of other elements. Multi-source satellite remote sensing data and UAV aerial survey data supplement the monitoring of the roadside meteorological environment at both macro and micro scales, providing spatial distribution characteristic data for various meteorological factors within the framework. Numerical model reanalysis data provides long-term meteorological element simulation results, facilitating the complete construction of a meteorological factor classification system over time. Finally, roadside digital environment base data, serving as a fundamental support, integrates information such as topography and geological conditions, correlating with the meteorological factor classification results. This allows for a more accurate assessment of the impact of meteorological factors on the roadbed structure, achieving deep integration of meteorological data and roadside environmental data.

[0015] The engineering impact equivalent calculation unit calculates the "engineering impact equivalent" for various classified meteorological factors. The specific calculation process is as follows: collect meteorological factors related to the roadbed engineering; and determine the weight of each meteorological factor's impact on the roadbed engineering based on historical monitoring data and mechanical tests. The raw meteorological factor data were normalized to obtain standardized values. Standardized values ​​of each meteorological factor With corresponding weights After multiplying and summing, the equivalent value of the engineering impact is obtained; the specific calculation formula is expressed as follows:

[0016] ,

[0017] in: Equivalent to the engineering impact; For the first The weights of each meteorological factor and ; For the first Standardized values ​​of each meteorological factor; This represents the total number of meteorological factors.

[0018] The regional characteristic event quantification submodule identifies and quantifies typical meteorological events on the Tibetan Plateau based on knowledge rules and pattern recognition algorithms. At the knowledge rule level, a structured rule base is established by integrating unique meteorological theories, historical disaster archives, and local climate zoning standards specific to the plateau region. For example, based on the plateau freeze-thaw cycle, a diurnal temperature range exceeding 15°C for three consecutive days with daily average temperatures fluctuating around 0°C is set as a triggering condition for freeze-thaw disasters. Using research findings on permafrost degradation in the Three-River-Source region, a rule for determining snowmelt-type floods based on a combination of sudden drops in snow depth and abnormal soil moisture content is formulated. The pattern recognition algorithm employs a spatiotemporal convolutional neural network to extract features from multi-source remote sensing data and ground meteorological station observation sequences, and learns the temporal evolution patterns of meteorological elements through a long short-term memory network. The algorithm training set includes nearly 20 years of ECMWF reanalysis data, ground station monitoring data, and disaster survey records from the Qinghai-Tibet Plateau region. Transfer learning is used to optimize the model's recognition accuracy under complex terrain conditions.

[0019] The composite climate stress construction submodule identifies and quantifies typical composite events based on association rule mining and causal graph models. "Based on association rule mining and causal graph models" refers to using association rule mining technology to find potential correlations between different meteorological factors from a large amount of meteorological data, discovering frequently occurring combinations of meteorological factors; and using causal graph models to further analyze the causal logic behind these correlations, clarifying the influence mechanism of each meteorological factor on the occurrence of typical composite events, thereby achieving accurate identification and quantification of typical composite events.

[0020] As a further explanation of the present invention, the meteorological factors under the "dynamic-thermal-moisture" three-dimensional classification framework are categorized as follows: dynamic meteorological factors include wind-driven erosion factors and hydrodynamic erosion factors; thermal meteorological factors include persistent thermal stress factors, drastic thermal stress factors, and radiative thermal stress factors; and moisture meteorological factors include liquid water intrusion factors, solid water action factors, and gaseous water transport factors.

[0021] As a further explanation of the present invention, the typical meteorological events of the plateau identified by the regional characteristic event quantification submodule include plateau hail events and spring snowmelt flood potential; the quantification index of the plateau hail event is hail kinetic energy flux, which is calculated by hail size distribution obtained by radar reflectivity factor and dual-polarization radar parameter inversion; the quantification index of the spring snowmelt flood potential is snowmelt flood index, which is a weighted function of previous snow water equivalent, warming rate and potential precipitation.

[0022] The specific calculation process for hail kinetic energy flux is as follows:

[0023] 1) Obtain radar reflectivity factor and dual-polarization radar parameters: Use a meteorological radar system to collect radar reflectivity factor data, differential reflectivity, and differential propagation phase shift rate within the target area;

[0024] 2) Inversion of hail size distribution: based on radar reflectivity factor Hail size Experience relationship:

[0025] ,

[0026] in: Let be the hail size spectral distribution function. and These are the minimum and maximum values ​​for hail size, respectively;

[0027] By combining dual-polarization radar parameters, the hail size spectrum distribution can be solved using iterative or optimization algorithms. ;

[0028] 3) Calculate the hailfall speed: based on the hail size. With falling speed Empirical formula:

[0029] ,

[0030] in: and This is an empirical coefficient used to calculate the corresponding falling speed for hailstones of different sizes;

[0031] 4) Calculate hail kinetic energy flux: hail kinetic energy flux The calculation formula is:

[0032] ,

[0033] in: air density;

[0034] The obtained hail size spectrum distribution and the obtained falling speed Substituting into the formula and integrating over the hail size, the kinetic energy flux of the hail is finally obtained. .

[0035] The functional expression for the snowmelt flood index is:

[0036] ,

[0037] in: Indicates the snowmelt flood index. This is the equivalent of the water from the accumulated snow in the previous period. For the heating rate, For potential rainfall, , , These are the weight coefficients of the corresponding variables, and they satisfy... .

[0038] As a further explanation of the present invention, the typical composite event identified by the composite climate stress construction submodule is the "freeze-thaw-scour" event. The identification logic is that a valid rainfall event exceeding a threshold occurs within a specified time window after a complete freeze-thaw cycle. The quantification index is a composite function of freeze-thaw depth, number of cycles, and subsequent rainfall intensity. The expression of the composite function is as follows:

[0039] ,

[0040] in, Indicates the freeze-thaw depth. Indicates the number of loops. Indicates the intensity of subsequent rainfall. These are the weighting coefficients.

[0041] As a further explanation of the present invention, the data correction algorithms built into the multi-source heterogeneous data intelligent fusion module for complex plateau terrain include elevation correction algorithms, slope correction algorithms and aspect correction algorithms for solar radiation data.

[0042] As a further explanation of the present invention, the standardized dataset output by the roadbed toughness application interface module includes a toughness design parameter package, a real-time risk feature vector, and a disease case-climate stress comparison database.

[0043] The present invention also provides a characterization method for a meteorological factor classification and quantification characterization system for roadbed climate resilience as described above, comprising the following steps:

[0044] S1. Construct the road domain digital environment base data of the target road segment through a multi-source heterogeneous data intelligent fusion module;

[0045] S2. Use the mechanistic classification unit to perform mechanistic classification on the original meteorological elements;

[0046] S3. Calculate the "engineering impact equivalent" of various meteorological factors through the engineering impact equivalent calculation unit;

[0047] S4. Identify and quantify typical meteorological events on the plateau through the regional characteristic event quantification submodule;

[0048] S5. Identify and quantify typical composite events by constructing a submodule based on composite climate stress;

[0049] S6. Integrate the results of S3, S4, and S5 to form a standardized dataset;

[0050] S7. Output the encapsulated standardized dataset through the roadbed toughness application interface module.

[0051] In the method described in this invention, step S6 specifically comprises:

[0052] Outlier removal and missing value imputation are performed in steps S3, S4, and S5 to unify the time resolution and spatial coordinate system;

[0053] Based on the three-dimensional classification framework of "dynamic-thermal-moisture", the data of each module are mapped to the corresponding dimension to form a structured matrix;

[0054] Standardization: The Z-score standardization method is used to make each index dimensionless. The calculation formula is as follows:

[0055] ,

[0056] in: The original data, The mean, The standard deviation is used to ensure data comparability.

[0057] Dataset presentation: The final data is stored in CSV format, and the file structure includes timestamps, meteorological factor indicators under the three-dimensional classification dimension, and regional labels.

[0058] Further explanation of the method described in this invention: In S3, the calculation of "engineering impact equivalent" adopts a hybrid model based on a combination of physical mechanisms and statistical analysis; firstly, the intrinsic relationship between meteorological factors and roadbed engineering response is analyzed through physical equations, then historical engineering data is used for parameter calibration and verification, and finally the meteorological observation values ​​are transformed into dimensionless or scaled values ​​with clear engineering significance; in S6, the standardized dataset includes engineering impact equivalent, quantitative results of regional characteristic events, and quantitative results of composite climate stress.

[0059] The hybrid model (i.e., semi-empirical-physical model) based on the combination of physical mechanisms and statistical analysis in step S3 of this invention is designed for the characteristics of roadbed climate resilience research. During the model parameter calibration process, multi-source heterogeneous meteorological data and roadbed performance monitoring data are introduced for joint optimization, making it more suitable for the actual needs of roadbed engineering under complex climatic conditions. At the same time, the empirical coefficients in the existing model are regionally adjusted, taking into account the geological and climatic differences in different regions, forming a regional coefficient correction system, thereby significantly improving the accuracy and applicability of the calculation results for assessing the impact of roadbed engineering.

[0060] Further explanation of the method described in this invention: the characterization method is used to drive a high-resolution fully coupled numerical model of roadbed water-thermal-mechanical system, or to generate a climate risk zoning map of the roadbed of a plateau highway network.

[0061] Compared with the prior art, the beneficial effects of the present invention are:

[0062] 1. This invention constructs a classification and quantification system for engineering physical processes. By defining and calculating the "equivalent engineering impact", it breaks through the traditional mode of directly referencing raw meteorological data and realizes a deep transformation from "weather data" to "engineering stress signal". This allows the analysis to go from the statistical correlation level to the physical mechanism level, providing a more targeted quantitative basis for roadbed engineering analysis.

[0063] 2. This invention achieves refined quantification of meteorological events unique to the plateau region. Through a dedicated sub-module and algorithm for quantifying regionally unique events, it effectively characterizes the engineering hazards of plateau-specific meteorological events such as hail and snowmelt floods, solves the adaptability problem of general methods in special areas, and improves the regional specificity of meteorological factor quantification.

[0064] 3. This invention bridges the "data gap" between meteorological data and engineering applications. Through a standardized roadbed toughness application interface module, it directly outputs data products adapted to downstream scenarios such as toughness design, intelligent early warning, and performance evaluation, significantly improving the utilization efficiency of meteorological information in engineering practice and strengthening its support for engineering decision-making.

[0065] 4. This invention possesses powerful capabilities for identifying and quantifying complex events. By constructing a sub-module based on complex climate stress, it accurately captures complex disasters formed by the synergistic effects of multiple meteorological factors, thus overcoming the limitations of single-factor quantification and providing more comprehensive technical support for the overall prevention and control of roadbed disasters. Attached Figure Description

[0066] Figure 1 This is a system module diagram of the present invention.

[0067] Figure 2 This is a schematic diagram of the multi-source heterogeneous data intelligent fusion module of the present invention.

[0068] Figure 3 This is a schematic diagram of the multi-scale meteorological factor classification and quantification module of the present invention.

[0069] Figure 4 This is a schematic diagram of the roadbed toughness application interface module of the present invention. Detailed Implementation

[0070] The invention will be further described below with reference to the accompanying drawings.

[0071] Example: See Figure 1-4 A meteorological factor classification and quantification characterization system for roadbed climate resilience is proposed, comprising a multi-source heterogeneous data intelligent fusion module, a multi-scale meteorological factor classification and quantification module, and a roadbed resilience application interface module connected in sequence.

[0072] The multi-source heterogeneous data intelligent fusion module is used to access and collaboratively process data from ground-based meteorological stations, highway traffic monitoring stations, multi-source satellite remote sensing data, UAV aerial survey data, and numerical model reanalysis data. The multi-source heterogeneous data intelligent fusion module has built-in data correction algorithms for complex plateau terrain and a data fusion model based on spatiotemporal attention mechanism, and outputs road area digital environment base data with high spatiotemporal integrity.

[0073] The multi-scale meteorological factor classification and quantification module is used to analyze, classify and quantify the features of road area digital environment base data, and generate a standardized dataset.

[0074] The roadbed resilience application interface module is used to encapsulate and output the standardized dataset generated by quantification according to the requirements of downstream engineering application scenarios.

[0075] Furthermore, the multi-scale meteorological factor classification and quantification module includes a mechanistic classification unit, an engineering impact equivalent calculation unit, a regional characteristic event quantification sub-module, and a composite climate stress construction sub-module.

[0076] The mechanistic classification unit constructs a three-dimensional classification framework of "dynamic-thermal-moisture" based on the main physical processes by which meteorological factors affect the roadbed structure, and classifies the original meteorological elements into meteorological factor categories under this framework. The meteorological factor categories under the "dynamic-thermal-moisture" three-dimensional classification framework are as follows: dynamic meteorological factors include wind-driven erosion factors and hydrodynamic erosion factors; thermal meteorological factors include persistent thermal stress factors, drastic thermal stress factors, and radiative thermal stress factors; and moisture meteorological factors include liquid water intrusion factors, solid water action factors, and gaseous water transport factors.

[0077] The engineering impact equivalent calculation unit calculates the "engineering impact equivalent" for various classified meteorological factors. The specific calculation process is as follows: collect meteorological factors related to the roadbed engineering; and determine the weight of each meteorological factor's impact on the roadbed engineering based on historical monitoring data and mechanical tests. The raw meteorological factor data were normalized to obtain standardized values. Standardized values ​​of each meteorological factor With corresponding weights After multiplying and summing, the equivalent value of the engineering impact is obtained; the specific calculation formula is expressed as follows:

[0078] ,

[0079] in: Equivalent to the engineering impact; For the first The weights of each meteorological factor and ; For the first Standardized values ​​of each meteorological factor; This represents the total number of meteorological factors.

[0080] The regional characteristic event quantification submodule identifies and quantifies typical plateau meteorological events based on knowledge rules and pattern recognition algorithms. Typical plateau meteorological events include plateau hail events and spring snowmelt flood potential.

[0081] The quantitative index of the plateau hail event is the hail kinetic energy flux, which is calculated by inverting the hail size distribution using radar reflectivity factor and dual-polarization radar parameters; the specific calculation process is as follows:

[0082] 1) Obtain radar reflectivity factor and dual-polarization radar parameters: Use a meteorological radar system to collect radar reflectivity factor data, differential reflectivity, and differential propagation phase shift rate within the target area;

[0083] 2) Inversion of hail size distribution: based on radar reflectivity factor Hail size Experience relationship:

[0084] ,

[0085] in: Let be the hail size spectral distribution function. and These are the minimum and maximum values ​​for hail size, respectively;

[0086] By combining dual-polarization radar parameters, the hail size spectrum distribution can be solved using iterative or optimization algorithms. ;

[0087] 3) Calculate the hailfall speed: based on the hail size. With falling speed Empirical formula:

[0088] ,

[0089] in: and This is an empirical coefficient used to calculate the corresponding falling speed for hailstones of different sizes;

[0090] 4) Calculate hail kinetic energy flux: hail kinetic energy flux The calculation formula is:

[0091] ,

[0092] in: air density;

[0093] The obtained hail size spectrum distribution and the obtained falling speed Substituting into the formula and integrating over the hail size, the kinetic energy flux of the hail is finally obtained. .

[0094] The quantitative index for the potential of spring snowmelt floods is the snowmelt flood index, which is a weighted function of the previous snowmelt water equivalent, the warming rate, and the potential rainfall; the function expression is:

[0095] ,

[0096] in: Indicates the snowmelt flood index. This is the equivalent of the water from the accumulated snow in the previous period. For the heating rate, For potential rainfall, , , These are the weight coefficients of the corresponding variables, and they satisfy... .

[0097] The composite climate stress construction submodule identifies and quantifies typical composite events based on association rule mining and causal graph models. Typical composite events are "freeze-thaw-scour" events. The identification logic is based on valid rainfall events exceeding a threshold occurring within a specified time window after a complete freeze-thaw cycle. The quantification index is a composite function of freeze-thaw depth, cycle number, and subsequent rainfall intensity, expressed as follows:

[0098] ,

[0099] in, Indicates the freeze-thaw depth. Indicates the number of loops. Indicates the intensity of subsequent rainfall. These are the weighting coefficients.

[0100] Furthermore, the standardized dataset output by the roadbed toughness application interface module includes a toughness design parameter package, a real-time risk feature vector, and a disease case-climate stress comparison database.

[0101] The characterization method of the meteorological factor classification and quantification characterization system for roadbed climate resilience in this embodiment includes the following steps:

[0102] S1. Construct the road domain digital environment base data for the target road segment through a multi-source heterogeneous data intelligent fusion module; specifically:

[0103] Through the multi-source heterogeneous data intelligent fusion module, multi-source heterogeneous data such as ground-based meteorological stations, highway traffic monitoring stations, multi-source satellite remote sensing, UAV aerial surveys and numerical model reanalysis are accessed. After being corrected by the built-in data correction algorithm and fused by the data fusion model based on the spatiotemporal attention mechanism, the road domain digital environment base of the target road segment is constructed.

[0104] S2. Use mechanistic classification units to perform mechanistic classification on the original meteorological elements; specifically:

[0105] The mechanistic classification unit in the multi-scale meteorological factor classification and quantification module is invoked. Based on the three-dimensional classification framework of "dynamic-thermal-moisture", the original meteorological elements in the road area digital environment base are classified mechanistically to obtain various refined meteorological factors.

[0106] S3. Calculate the "engineering impact equivalent" of various meteorological factors using the engineering impact equivalent calculation unit; specifically:

[0107] A hybrid model combining physical mechanisms and statistical analysis (semi-empirical-physical model) is adopted to transform the observed values ​​of various fine meteorological factors obtained in S2 into corresponding "engineering impact equivalents". Compared with existing technologies, the improved semi-empirical-physical model is tailored to the characteristics of roadbed climate resilience research. During the model parameter calibration process, multi-source heterogeneous meteorological data and roadbed performance monitoring data are introduced for joint optimization, making it more suitable for the actual needs of roadbed engineering under complex climatic conditions. At the same time, the empirical coefficients in the existing model have been regionalized, taking into account the geological and climatic differences in different regions, forming a regional coefficient correction system, thereby significantly improving the accuracy and applicability of the calculation results for roadbed engineering impact assessment.

[0108] S4. Identify and quantify typical plateau meteorological events through the regional characteristic event quantification submodule; specifically:

[0109] The regional characteristic event quantification submodule in the multi-scale meteorological factor classification and quantification module is activated. Based on knowledge rules and pattern recognition algorithms, typical plateau meteorological events in the road area digital environment base are identified. The engineering impact equivalent calculation unit is called to quantify the identified typical plateau meteorological events and obtain the regional characteristic event quantification results.

[0110] S5. Identify and quantify typical composite events through a sub-module constructed using composite climate stress; specifically:

[0111] By constructing a sub-module for complex climate stress in the multi-scale meteorological factor classification and quantification module, based on association rule mining and causal graph model, a complex event formed by the synergistic or progressive effects of multiple single factors is identified, and the complex event is quantified to obtain the quantification result of complex climate stress.

[0112] S6. Integrate the results of S3, S4, and S5 to form a standardized dataset; that is, integrate the engineering impact equivalent of S3, the regional characteristic event quantification results of S4, and the composite climate stress quantification results of S5 to form a standardized dataset for the target road segment; specifically:

[0113] Outlier removal and missing value imputation are performed in steps S3, S4, and S5 to unify the time resolution and spatial coordinate system;

[0114] Based on the three-dimensional classification framework of "dynamic-thermal-moisture", the data of each module are mapped to the corresponding dimension to form a structured matrix;

[0115] Standardization: The Z-score standardization method is used to make each index dimensionless. The calculation formula is as follows:

[0116] ,

[0117] in: The original data, The mean, The standard deviation is used to ensure data comparability.

[0118] Dataset presentation: The final data is stored in CSV format, and the file structure includes timestamps, meteorological factor indicators under the three-dimensional classification dimension, and regional labels;

[0119] S7. The encapsulated standardized dataset is output through the subgrade toughness application interface module, specifically as follows:

[0120] Through the roadbed toughness application interface module, the standardized dataset generated in S6 is encapsulated according to the input requirements of downstream engineering application scenarios, and the toughness design parameter package, real-time risk feature vector or disease case-climate stress comparison database is output.

[0121] Application Example 1: Application on a permafrost slope section of Highway G214

[0122] 1. System Deployment and Data Preparation

[0123] A road section with typical differences in slope orientation and a history of shallow landslides was selected for deployment of the characterization system of this embodiment. Through a multi-source heterogeneous data intelligent fusion module, hourly meteorological observation data from the past 15 years, a high-resolution digital elevation model, historical Sentinel-1 / 2 series satellite imagery, and seasonal UAV multispectral and thermal infrared remote sensing data from the road section and surrounding area were integrated. After processing by the module's built-in data correction algorithm (elevation and slope-aspect correction for solar radiation data) and spatiotemporal attention mechanism fusion model, a "road area digital environment base" with a spatial resolution of 30 meters and a daily temporal resolution was generated, including key field data such as air temperature, precipitation, soil moisture, and surface temperature.

[0124] 2. Classification and Quantification of Meteorological Factors and Generation of "Climate Stress Spectrum"

[0125] The system calls the multi-scale meteorological factor classification and quantification module for in-depth processing:

[0126] Mechanistic classification and equivalent calculation: After classifying the temperature data, the average annual number of freeze-thaw cycles on the sunny slope and the average annual number of freeze-thaw cycles on the shady slope of the road section were calculated, as well as the freeze-thaw index equivalent for each cycle; after classifying the precipitation data, it was divided into short-duration heavy precipitation and long-duration weak precipitation, and the Green-Ampt infiltration model was used to quantify it into the surface saturated infiltration depth equivalent and the deep progressive wetting depth equivalent, respectively.

[0127] Quantification of regionally specific events: The system identifies a typical snowmelt-type flood potential event in May 2018 through knowledge rules and pattern recognition algorithms, calculates the snowmelt-flood index, and, combined with the rainfall data during that period, determines that this is a high-risk "snowmelt + rainfall" combined disaster event.

[0128] Composite event construction: The system automatically retrieves historical data, constructs multiple "freeze-thaw-scour" event templates, identifies the event in April 2020 where 15mm of precipitation occurred within 48 hours after a strong freeze-thaw cycle, and calculates the composite stress index of the event.

[0129] By integrating the above analysis results, the system automatically generates a standardized dataset for the road section that includes more than 20 quantitative indicators over the past 15 years.

[0130] 3. Engineering Application and Effect Verification

[0131] The "climate stress spectrum" can be output for the following scenarios through the roadbed toughness application interface module:

[0132] Retrospective analysis of disease mechanism: Correlation analysis between "climate stress spectrum" and historical slope deformation monitoring data (InSAR) verifies the strong correlation between the quantitative indicators of this system and the engineering response.

[0133] Improved early warning model: The "composite stress index" and "snowmelt flood index" were added as new features to the machine learning early warning model for slope instability of this section of the road, which improved the accuracy of early warning and reduced the false alarm rate.

[0134] Supporting resilience design: Based on the statistical results of the "climate stress spectrum", differentiated drainage and insulation design suggestions were proposed for the sunny and shady slopes of this section, which improved the pertinence of the protection design.

[0135] Application Example 2: Application on long and steep slopes in high-altitude seasonally frozen soil regions (targeting wind-freeze-thaw combined effects)

[0136] 1. Application Scenarios and Problems

[0137] The implementation plan was applied to a long, steep slope section (average slope > 5%) at an altitude of approximately 3800 meters in the eastern part of a plateau. This area is a typical high-altitude seasonal permafrost region, without perennial permafrost, but with long, cold winters and intense freeze-thaw cycles. The main engineering problems are: ① Severe snow and ice accumulation on the road surface in winter, accompanied by strong winds forming "blown snow," causing traffic disruptions and structural frost heave; ② During the spring thaw, the surface soil of the slope becomes loose due to repeated freeze-thaw cycles, making it prone to shallow landslides under the influence of gravity and meltwater; ③ Strong diurnal temperature differences and solar radiation accelerate the aging of asphalt pavement. Traditional maintenance relies on experience and regular inspections, which cannot accurately predict the coupled disasters of "wind-freeze-thaw" cycles.

[0138] 2. System Implementation and Data Processing

[0139] The system focuses on a 10-kilometer stretch of road, integrating the following data into a multi-source heterogeneous data intelligent fusion module: automatic weather stations deployed along the road (measuring wind speed, wind direction, temperature, and humidity), road surface condition sensors (measuring temperature and ice thickness), satellite remote sensing surface temperature products, and high-resolution wind field reanalysis data from regional meteorological models. After fusion processing, a "roadside digital environment base" with a spatial resolution of 100 meters and a temporal resolution of 1 hour is generated, primarily including wind speed and direction fields, road surface temperature, snow depth, and soil temperature profile data.

[0140] 3. Classification and Quantification and Generation of "Climate Stress Spectrum"

[0141] The system calls the multi-scale meteorological factor classification and quantification module for processing:

[0142] Quantification of wind erosion factors: Based on wind speed and direction frequency, the annual sand transport potential and blowing snow potential index of each road segment unit are calculated, and the mountain pass area in the middle of the road segment is identified as a high-risk area for blowing snow.

[0143] Quantification of dramatic thermal stress: The number of effective freeze-thaw cycles per year in the region was calculated, "harmful freeze-thaw cycles" were defined and their cycle depth was quantified, and it was found that the number of "harmful freeze-thaw cycles" was more on sunny slopes than on shady slopes.

[0144] Composite event construction—wind-driven freeze-thaw erosion: Construct a composite event model of "wind-driven freeze-thaw erosion" and define the "wind-freeze erosion index", which is a function of freeze-thaw cycle intensity, wind speed and soil moisture content.

[0145] 4. Engineering Applications and Effects

[0146] Customized products can be output through the roadbed toughness application interface module:

[0147] Precise prevention and control of blowing snow disasters: Based on the spatial distribution map of blowing snow potential, the layout of snow protection nets and wind deflectors was optimized, reducing the time of traffic interruption in winter.

[0148] Dynamic early warning of slope stability: By integrating the "wind frost erosion index" into the slope monitoring system of the maintenance unit, multiple potential shallow landslide areas were successfully warned, improving the efficiency of maintenance resource allocation.

[0149] Road material selection guidelines: Based on the "radiative heat stress equivalent", high viscoelastic modified asphalt is recommended to improve the fatigue life of road materials.

[0150] Application Example 3: Application in road sections affected by rainstorm-debris flow disaster chains in deep river valleys

[0151] 1. Application Scenarios and Problems

[0152] The example was applied to a highway section in a deeply incised river valley. This area has steep terrain, abundant rainfall, and the chain of debris flows triggered by torrential rain is the primary threat to roadbed safety. The core challenges are: ① Traditional regional rainstorm warnings cannot accurately reflect the distribution of localized heavy rainfall in complex terrain; ② It is difficult to quantify the contribution of the combination of prior soil moisture conditions and short-term heavy rainfall to the initiation of debris flows; ③ The roadbed structure not only faces flood erosion but is also subjected to burial and impact from debris flow deposits.

[0153] 2. System Implementation and Data Processing

[0154] For a 15-kilometer stretch of river valley road, the system integrates quantitative precipitation estimation products from meteorological radar, high-resolution topographic data, historical geological disaster data, MODIS vegetation index, and soil moisture simulation data from a distributed hydrological model. The intelligent fusion module of multi-source heterogeneous data generates a refined "road area digital environment base" with a spatial resolution of 50 meters and a temporal resolution of 1 hour through downscaling technology. The core variables include areal rainfall, soil saturation, topographic slope, and vegetation cover.

[0155] 3. Classification and Quantification and Generation of "Climate Stress Spectrum"

[0156] The system calls the multi-scale meteorological factor classification and quantification module for analysis:

[0157] Refined quantification of liquid water intrusion factors: Rainfall is decomposed into accumulated rainfall and short-term intensified rainfall. Hydrological models are used to transform rainfall sequences with different spatiotemporal distributions into runoff generation and soil pore water pressure changes for each slope unit.

[0158] Hydrodynamic erosion factors and debris flow initiation potential: Combine topographic data to calculate the comprehensive topographic index of each confluence unit, construct a debris flow rainstorm critical curve model, and dynamically calculate the "debris flow initiation proximity" of each unit.

[0159] Construction of composite disaster chains: Define the disaster chain event of "rainstorm-flood-debris flow", identify slope units that reach the "debris flow initiation proximity" threshold, predict the movement path of the debris source, and assess the flood erosion intensity and debris flow impact / burial risk level of the roadbed.

[0160] 4. Engineering Applications and Effects

[0161] The following decision support products are generated through the roadbed resilience application interface module:

[0162] Mechanism-based real-time debris flow risk warning: The spatial distribution map of "debris flow initiation proximity" is updated and released every hour, achieving accurate warning and preventing vehicles from entering high-risk areas.

[0163] Optimized design of roadbed protection structure: Based on the spatial distribution map of "debris flow impact risk level" and "flood scour intensity", differentiated protection engineering design is carried out to optimize the project cost while ensuring the reliability of protection.

[0164] Dynamic planning of maintenance and inspection routes: After a disaster, the optimal inspection and assessment route is automatically generated based on the simulation results of the impact range of the "disaster chain" event, which improves the efficiency of emergency response.

[0165] Obviously, the above embodiments are merely examples for clearly illustrating the present invention, and are not intended to limit the implementation of the present invention. Those skilled in the art will recognize that other variations or modifications can be made based on the above description; it is neither necessary nor possible to exhaustively list all possible implementations; however, obvious variations or modifications derived therefrom are still within the scope of protection of the present invention.

Claims

1. A meteorological factor classification and quantitative characterization system for roadbed climate resilience, characterized in that: It includes a multi-source heterogeneous data intelligent fusion module, a multi-scale meteorological factor classification and quantification module, and a roadbed resilience application interface module, which are connected in sequence. The multi-source heterogeneous data intelligent fusion module is used to access and collaboratively process data from ground-based meteorological stations, highway traffic monitoring stations, multi-source satellite remote sensing data, UAV aerial survey data, and numerical model reanalysis data. The multi-source heterogeneous data intelligent fusion module has built-in data correction algorithms for complex plateau terrain and a data fusion model based on spatiotemporal attention mechanism, and outputs road area digital environment base data with high spatiotemporal integrity. The multi-scale meteorological factor classification and quantification module is used to analyze, classify and quantify the features of road area digital environment base data, and generate a standardized dataset. The roadbed resilience application interface module is used to encapsulate and output the standardized dataset generated by quantification according to the requirements of downstream engineering application scenarios.

2. The meteorological factor classification and quantitative characterization system for roadbed climate resilience according to claim 1, characterized in that: The multi-scale meteorological factor classification and quantification module includes a mechanistic classification unit, an engineering impact equivalent calculation unit, a regional characteristic event quantification sub-module, and a composite climate stress construction sub-module. The mechanistic classification unit constructs a three-dimensional classification framework of "dynamic-thermal-moisture" based on the main physical processes by which meteorological factors affect the roadbed structure, and classifies the original meteorological elements into meteorological factor categories under this framework; The engineering impact equivalent calculation unit calculates the "engineering impact equivalent" for various classified meteorological factors. The specific calculation process is as follows: collect meteorological factors related to the roadbed engineering; and determine the weight of each meteorological factor's impact on the roadbed engineering based on historical monitoring data and mechanical tests. ; The raw meteorological factor data were normalized to obtain standardized values. Standardized values ​​of each meteorological factor With corresponding weights After multiplying and summing, the equivalent value of the engineering impact is obtained; The regional characteristic event quantification submodule identifies and quantifies typical meteorological events on the plateau based on knowledge rules and pattern recognition algorithms. The composite climate stress construction submodule identifies and quantifies typical composite events based on association rule mining and causal graph modeling.

3. The meteorological factor classification and quantitative characterization system for roadbed climate resilience according to claim 2, characterized in that: The meteorological factors under the "dynamic-thermal-moisture" three-dimensional classification framework are categorized as follows: dynamic meteorological factors include wind-driven erosion factors and hydrodynamic erosion factors; thermal meteorological factors include persistent thermal stress factors, drastic thermal stress factors, and radiative thermal stress factors; and moisture meteorological factors include liquid water intrusion factors, solid water action factors, and gaseous water transport factors.

4. The meteorological factor classification and quantitative characterization system for roadbed climate resilience according to claim 2, characterized in that: The regional characteristic event quantification submodule identifies typical plateau meteorological events including plateau hail events and spring snowmelt flood potential. The quantification index for plateau hail events is hail kinetic energy flux, which is calculated by inverting hail size distribution through radar reflectivity factor and dual-polarization radar parameters. The quantification index for spring snowmelt flood potential is snowmelt flood index, which is a weighted function of previous snow water equivalent, warming rate and potential precipitation.

5. The meteorological factor classification and quantitative characterization system for roadbed climate resilience according to claim 2, characterized in that: The typical composite event identified by the composite climate stress construction submodule is the "freeze-thaw-scour" event. The identification logic is that an effective rainfall event exceeding a threshold occurs within a specified time window after a complete freeze-thaw cycle. The quantitative index is a composite function of freeze-thaw depth, number of cycles, and subsequent rainfall intensity.

6. The meteorological factor classification and quantitative characterization system for roadbed climate resilience according to claim 1, characterized in that: The intelligent fusion module for multi-source heterogeneous data includes data correction algorithms for complex plateau terrain, such as elevation correction, slope correction, and aspect correction algorithms for solar radiation data.

7. The meteorological factor classification and quantitative characterization system for roadbed climate resilience according to claim 1, characterized in that: The standardized dataset output by the roadbed toughness application interface module includes a toughness design parameter package, real-time risk feature vectors, and a disease case-climate stress comparison database.

8. A characterization method for a meteorological factor classification and quantitative characterization system for roadbed climate resilience as described in any one of claims 1-7, characterized in that... Includes the following steps: S1. Construct the road domain digital environment base data of the target road segment through a multi-source heterogeneous data intelligent fusion module; S2. Use the mechanistic classification unit to perform mechanistic classification on the original meteorological elements; S3. Calculate the "engineering impact equivalent" of various meteorological factors through the engineering impact equivalent calculation unit; S4. Identify and quantify typical meteorological events on the plateau through the regional characteristic event quantification submodule; S5. Identify and quantify typical composite events by constructing a submodule based on composite climate stress; S6. Integrate the results of S3, S4, and S5 to form a standardized dataset; S7. Output the encapsulated standardized dataset through the roadbed toughness application interface module.

9. The characterization method of the meteorological factor classification and quantitative characterization system for roadbed climate resilience according to claim 8, characterized in that: The calculation of "engineering impact equivalent" in S3 adopts a hybrid model based on a combination of physical mechanisms and statistical analysis. First, the intrinsic relationship between meteorological factors and roadbed engineering response is analyzed through physical equations. Then, historical engineering data is used for parameter calibration and verification. Finally, meteorological observation values ​​are transformed into dimensionless or scaled values ​​with clear engineering significance. The standardized dataset in S6 includes engineering impact equivalent, quantitative results of regional characteristic events, and quantitative results of composite climate stress.

10. The characterization method of the meteorological factor classification and quantitative characterization system for roadbed climate resilience according to claim 8, characterized in that: The characterization method is used to drive a high-resolution fully coupled numerical model of roadbed water-thermal-mechanical system, or to generate a climate risk zoning map of the roadbed of the plateau highway network.

Images