Intelligent evaluation and improvement method for resilience of earthquake-induced slope disaster traffic network based on cross-scale coupling

Abstract

The application provides a seismic slope disaster traffic network resilience intelligent evaluation and promotion method based on cross-scale coupling, belongs to the technical field of traffic network resilience evaluation under seismic slope disasters, and comprises the following steps: S1, database construction and data integration; S2, slope disaster traffic unit damage evaluation; S3, road network resilience evaluation; S4, intelligent recovery path optimization; and S5, recovery progress evaluation and real-time adjustment. Through the integration of multi-source heterogeneous data, the comprehensive evaluation system integrating physical mechanisms and intelligent algorithms is formed, and full-process integrated decision support is provided for the pre-earthquake prevention, disaster response and post-disaster recovery of the traffic lifeline system.

Description

Technical Field

[0001] This invention relates to the field of transportation network resilience assessment technology under earthquake slope disasters, and in particular to an intelligent assessment and improvement method for transportation network resilience under earthquake slope disasters based on cross-scale coupling. Background Technology

[0002] Earthquake-induced slope hazards are among the major geological disasters that seriously threaten the safety of highway lifelines. The physical damage they cause to local road units, and the resulting disruption of road network function, exhibit significant cross-scale transmission and chain effects. Highway network resilience assessment is a core issue in disaster emergency management and transportation infrastructure safety, and its accuracy and practicality directly affect the efficiency of post-disaster relief and the formulation of recovery strategies.

[0003] Currently, the assessment methods for highway network resilience still have the following main limitations: (1) Scale fragmentation problem: Existing studies mostly focus on single-scale analysis, or only consider the effects of earthquake motion, or only assess the damage of local road sections. They have not yet established a cross-scale coupling mechanism model that can connect "earthquake-slope-road unit-road network function", which makes it difficult to truly reflect the process of road network function degradation and system recovery potential under the action of disaster chain.

[0004] (2) Insufficient recovery modeling: In terms of post-disaster recovery capacity modeling, most methods rely on empirical or static repair sequence arrangements, lacking overall optimization of resource constraints, temporal dependence and dynamic evolution of road network performance during the recovery process. It is difficult to achieve adaptive planning of recovery paths based on real-time damage data, which limits the scientific nature and timeliness of emergency decision-making.

[0005] (3) Limited data fusion capability: Road network resilience assessment involves multi-source heterogeneous data such as seismic motion parameters, slope geological conditions, road structural attributes and traffic operation status. Traditional methods have limited capabilities in data fusion, feature extraction and system behavior modeling, and have not yet formed a comprehensive evaluation system that integrates physical mechanisms and intelligent algorithms. Summary of the Invention

[0006] This invention provides an intelligent assessment and enhancement method for the resilience of transportation networks affected by earthquake slope disasters based on cross-scale coupling. It aims to address the following technical problems in existing technologies: First, it overcomes the scale fragmentation problem in the "earthquake-slope-road unit-network function" disaster chain transmission process in existing assessment methods, achieving quantitative assessment of the entire chain from earthquake input to network function degradation; second, it breaks through the limitations of traditional post-disaster recovery modeling that relies on empirical or static repair sequence arrangements, achieving adaptive optimization of recovery paths based on real-time data, thus improving the scientific rigor and timeliness of emergency decision-making; third, it integrates multi-source heterogeneous data to form a comprehensive evaluation system that combines physical mechanisms and intelligent algorithms, providing integrated decision support for the entire process of pre-earthquake prevention, in-disaster response, and post-disaster recovery of transportation lifeline systems.

[0007] To achieve the above objectives, the present invention adopts the following technical solution: A method for intelligent assessment and improvement of the resilience of transportation networks affected by earthquake slope disasters based on cross-scale coupling includes: S1: Collect earthquake data, slope data, traffic flow and road network data, and historical disaster data; standardize the collected data and establish a database. S2: Based on the database, an earthquake slope hazard assessment model is constructed using an artificial neural network and the Apriori algorithm to obtain high-risk slope units; for the high-risk slope units, a rapid calculation model for earthquake landslide movement is established based on the principle of energy conservation to calculate the sliding distance, sliding velocity, and landslide impact force; a nonlinear mapping relationship is established between the landslide impact force, sliding velocity, sliding distance, and road damage indicators to determine the damage level of each road unit; S3: Based on the damage level of each road unit, assign a traffic capacity weight to each road unit, construct a road network connectivity assessment model, and calculate the connectivity index; based on the connectivity index, combined with the recovery timeliness index, traffic capacity recovery rate, and key facility connectivity rate, conduct a multi-indicator comprehensive resilience score for the highway network to obtain a resilience assessment index. S4: Based on the damage level of each road unit, a genetic algorithm is used to intelligently optimize the recovery path of the post-disaster highway network, construct a comprehensive fitness function that includes a time discount factor and a critical infrastructure connectivity reward, search for the optimal repair sequence, and conduct an expected resilience assessment on the optimal repair sequence to obtain the expected resilience assessment index after repair. S5: During the recovery process, repairs are performed based on the optimal repair sequence. The actual recovery progress is periodically evaluated and compared with the plan. When deviations occur, the subsequent repair sequence is dynamically adjusted. After the repairs are completed, the actual resilience assessment index is calculated and compared with the expected value for verification, thereby achieving flexible response and model self-optimization during the recovery process.

[0008] In this specification, the specific steps of constructing the earthquake slope hazard assessment model using an artificial neural network and the Apriori algorithm in S2 are as follows: The Apriori algorithm is used to mine frequent itemsets in multi-source data, setting the minimum support to 0.2 and the minimum confidence to 0.8 to generate binary association rule features; these binary association rule features, along with earthquake features, slope features, and geological features, form a 15-dimensional input feature vector; an artificial neural network with two hidden layers is used, with the number of hidden layer nodes set to 12 and 8 respectively, employing the ReLU activation function and the Sigmoid output function, trained with the Adam optimizer, a learning rate of 0.001, and an early stopping method to output the slope hazard probability; a probability threshold of 0.7 is set, and slope units with a hazard probability greater than 0.7 are classified as high-risk slope units.

[0009] In this specification, the specific steps for establishing a rapid calculation model of earthquake-induced landslide motion for high-risk slope units in step S2 are as follows: Obtain active fault data at the location of the high-risk slope unit, and assign the maximum potential earthquake magnitude of the active fault as the design magnitude; calculate the shortest distance from the slope unit to the fault plane, and, in conjunction with site condition parameters, calculate the peak ground motion parameters under the maximum credible earthquake scenario through the ground motion attenuation relationship; use the ratio of the peak ground motion parameters to the critical acceleration as the input parameter for ground motion intensity; based on the principle of energy conservation, calculate the reduction in potential energy of the sliding body, earthquake energy, energy dissipated by friction on the sliding surface, topographic energy, deformation energy, and the increase in total kinetic energy of the sliding body, establish an energy balance equation, and solve for the sliding distance, sliding velocity, and landslide impact force.

[0010] In this specification, the specific steps for establishing the nonlinear mapping relationship between landslide impact force, sliding velocity, sliding distance and road damage indicators in step S2 are as follows: using multivariate nonlinear regression analysis, with landslide impact force, sliding velocity, and sliding distance as independent variables and the original data of road damage level as dependent variables, to fit regression coefficients; scaling the regression output results to the range of 0 to 5 using the Sigmoid function, and dividing them into five damage levels: Level 1 minor damage, Level 2 moderate damage, Level 3 severe damage, Level 4 complete interruption, and Level 5 catastrophic damage.

[0011] In this specification, the specific method for assigning capacity weights to each road unit in S3 is as follows: based on the damage level and road grade of the road unit, the capacity weights are determined by referring to Table 2; the road network connectivity assessment model uses the Dijkstra algorithm to simulate post-disaster changes in road network connectivity, and the formula for calculating the connectivity index is: the sum of the products of the capacity weights of all road segments and the importance coefficients of the road segments divided by the sum of the importance coefficients of all road segments, wherein the importance coefficients of the road segments are assigned according to the road grade, with 3 for expressways, 2 for national highways, and 1 for provincial and county roads.

[0012] In this specification, the specific steps for the multi-index comprehensive resilience scoring in S3 are as follows: The weights of each index are determined using the analytic hierarchy process (AHP), where the connectivity index has a weight of 0.35, the recovery timeliness index has a weight of 0.3, the traffic capacity recovery rate has a weight of 0.2, and the critical facility connectivity rate has a weight of 0.15. The recovery timeliness index is the ratio of the expected recovery time to the standard recovery time, and the standard recovery time is determined based on the road unit damage level as shown in Table 1. The traffic capacity recovery rate is the ratio of post-disaster traffic capacity to pre-disaster traffic capacity, and the post-disaster traffic capacity is calculated based on the road unit damage level and road grade as shown in Table 2. The critical facility connectivity rate is the proportion of the number of connected core facilities to the total number of core facilities, and the connectivity status of each core facility is determined based on the damage level of its respective road unit as shown in Table 3.

[0013] In this specification, the parameters of the genetic algorithm in S4 are set as follows: population size 100, crossover probability 0.8, mutation probability 0.05, selection strategy using roulette wheel selection combined with elite retention, retaining the top 10% of individuals, maximum number of iterations 200 generations, and termination condition when fitness improves by less than 1% for 20 consecutive generations; the comprehensive fitness function includes: time-discounted summation of road network performance indicators at each stage of the repair process, plus weighted summation of core facility connectivity rewards at each stage, where the time discount factor is 0.92, and the core facility connectivity reward is 10 points for connectivity and 0 points for non-connectivity.

[0014] In this specification, the road network performance index in S4 is a weighted sum of the connectivity index and the capacity recovery rate. The connectivity index is calculated in real time by the road network connectivity assessment model, and the capacity recovery rate is the ratio of the total road network capacity at that moment to the total road network capacity before the disaster. The core facilities include hospitals, emergency command centers, and material warehouses, and their weights are determined according to the importance level of the emergency response, with hospitals and emergency command centers receiving a weight of 5 and material warehouses receiving a weight of 3.

[0015] In this specification, the specific steps of S5 for dynamically adjusting the recovery path based on real-time feedback are as follows: periodically evaluate the deviation between the actual recovery time of the repaired road segment and the standard recovery time in Table 1. When the deviation exceeds the preset threshold, re-trigger the genetic algorithm optimization of S4, take the current unrepaired road segment set as the initial state, update the start time of the time discount factor in the fitness function, and generate the adjusted repair sequence.

[0016] In this specification, the earthquake data collected in S1 includes epicenter location, focal depth, magnitude, focal mechanism, peak ground motion parameters, spectral characteristic parameters, active fault data, and duration parameters; the slope data includes slope, soil type, vegetation cover, slope geometry, geological conditions, load, acceleration, displacement, and stress distribution; the traffic flow and road network data includes road geometry, structural characteristics of bridges and tunnels, seismic design parameters, traffic flow, road structure, and construction and maintenance data of traffic units; the historical disaster data includes damage type, damage scale, and information on the disaster-causing slope. The above data are standardized using the Min-Max standardization method, missing values ​​are filled using linear interpolation, and road network topology parameters are extracted based on graph theory algorithms.

[0017] In summary, the present invention has at least the following beneficial effects: (1) The ability to quantitatively assess multi-level disaster chains has been improved: By constructing a four-level cross-scale coupled model of “earthquake-slope-road unit-road network function”, the complete disaster chain transmission process from earthquake input, slope instability, road unit damage to road network function degradation has been connected, overcoming the limitations of traditional single-scale or fragmented assessment, and significantly improving the simulation accuracy of road network function evolution under the action of disaster chains.

[0018] (2) Enhanced the scientific nature of post-disaster recovery decision-making: The optimal recovery path optimization method based on genetic algorithm constructed a multi-objective fitness function that includes time discount factor and core facility connectivity reward, realizing dynamic optimization of repair sequence and providing quantitative scientific basis for emergency management decision-making.

[0019] (3) Improved system real-time performance and adaptability: By integrating physical mechanism models with data-driven intelligent algorithms, real-time assessment of the entire chain from seismic input to road network function degradation was realized, supporting dynamic response and adaptive adjustment of recovery path during disasters.

[0020] (4) Provided a technical reference system for the project: A parameterized system covering damage level classification, standard recovery time, traffic capacity weight and core facility connectivity determination was established, providing a quantitative basis that can be directly applied to engineering practice. Attached Figure Description

[0021] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. 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.

[0022] Figure 1This is a schematic diagram of the intelligent assessment and improvement method for the resilience of transportation networks in earthquake-prone slope disasters based on cross-scale coupling involved in this invention. Detailed Implementation

[0023] In the following description, only certain exemplary embodiments are briefly described. As those skilled in the art will recognize, the described embodiments can be modified in various ways without departing from the spirit or scope of the embodiments of the invention. Therefore, the drawings and description are considered to be exemplary in nature and not restrictive.

[0024] The following disclosure provides many different implementations or examples for carrying out different structures of the embodiments of the present invention. To simplify the disclosure of the embodiments of the present invention, specific examples of components and arrangements are described below. Of course, these are merely examples and are not intended to limit the embodiments of the present invention. Furthermore, reference numerals and / or reference letters may be repeated in different examples of the embodiments of the present invention; such repetition is for simplification and clarity and does not in itself indicate a relationship between the various implementations and / or arrangements discussed.

[0025] The embodiments of the present invention will now be described in detail with reference to the accompanying drawings.

[0026] like Figure 1 As shown, this embodiment provides a method for intelligent assessment and improvement of the resilience of transportation networks in earthquake-prone slope disasters based on cross-scale coupling, including: S1: Collect earthquake data, slope data, traffic flow and road network data, and historical disaster data; standardize the collected data and establish a database. S2: Based on the database, an earthquake slope hazard assessment model is constructed using an artificial neural network and the Apriori algorithm to obtain high-risk slope units; for the high-risk slope units, a rapid calculation model for earthquake landslide movement is established based on the principle of energy conservation to calculate the sliding distance, sliding velocity, and landslide impact force; a nonlinear mapping relationship is established between the landslide impact force, sliding velocity, sliding distance, and road damage indicators to determine the damage level of each road unit; S3: Based on the damage level of each road unit, assign a traffic capacity weight to each road unit, construct a road network connectivity assessment model, and calculate the connectivity index; based on the connectivity index, combined with the recovery timeliness index, traffic capacity recovery rate, and key facility connectivity rate, conduct a multi-indicator comprehensive resilience score for the highway network to obtain a resilience assessment index. S4: Based on the damage level of each road unit, a genetic algorithm is used to intelligently optimize the recovery path of the post-disaster highway network, construct a comprehensive fitness function that includes a time discount factor and a critical infrastructure connectivity reward, search for the optimal repair sequence, and conduct an expected resilience assessment on the optimal repair sequence to obtain the expected resilience assessment index after repair. S5: During the recovery process, repairs are performed based on the optimal repair sequence. The actual recovery progress is periodically evaluated and compared with the plan. When deviations occur, the subsequent repair sequence is dynamically adjusted. After the repairs are completed, the actual resilience assessment index is calculated and compared with the expected value for verification, thereby achieving flexible response and model self-optimization during the recovery process.

[0027] In some embodiments, the specific steps of constructing an earthquake slope hazard assessment model using an artificial neural network and the Apriori algorithm in S2 are as follows: The Apriori algorithm is used to mine frequent itemsets in multi-source data, setting the minimum support to 0.2 and the minimum confidence to 0.8 to generate binary association rule features; these binary association rule features, along with earthquake features, slope features, and geological features, form a 15-dimensional input feature vector; an artificial neural network with two hidden layers is used, setting the number of hidden layer nodes to 12 and 8 respectively, using the ReLU activation function and the Sigmoid output function, trained with the Adam optimizer, setting the learning rate to 0.001, and employing early stopping to output the slope hazard probability; a probability threshold of 0.7 is set, and slope units with a hazard probability greater than 0.7 are identified as high-risk slope units.

[0028] In some embodiments, the specific steps of establishing a rapid calculation model for earthquake-induced landslide motion for high-risk slope units in S2 are as follows: acquiring active fault data at the location of the high-risk slope unit, and assigning the maximum potential earthquake magnitude of the active fault as the design magnitude; calculating the shortest distance from the slope unit to the fault plane, and calculating the peak ground motion parameters under the maximum credible earthquake scenario through ground motion attenuation relationship in combination with site condition parameters; using the ratio of the peak ground motion parameters to the critical acceleration as the input parameter for ground motion intensity; and based on the principle of energy conservation, calculating the reduction in potential energy of the sliding body, earthquake energy, energy dissipation from friction on the sliding surface, topographic energy, deformation energy, and the increase in total kinetic energy of the sliding body, establishing an energy balance equation, and solving for the sliding distance, sliding velocity, and landslide impact force.

[0029] In some embodiments, the specific steps for establishing the nonlinear mapping relationship between landslide impact force, sliding velocity, sliding distance and road damage indicators in step S2 are as follows: using multivariate nonlinear regression analysis, with landslide impact force, sliding velocity, and sliding distance as independent variables and the original data of road damage level as dependent variables, to fit regression coefficients; scaling the regression output results to the range of 0 to 5 using the Sigmoid function, and dividing them into five damage levels: Level 1 minor damage, Level 2 moderate damage, Level 3 severe damage, Level 4 complete interruption, and Level 5 catastrophic damage.

[0030] In some embodiments, the specific method for assigning capacity weights to each road unit in S3 is as follows: based on the damage level of the road unit, level 1 is assigned 1.0, level 2 0.8, level 3 0.5, level 4 0.2, and level 5 0.0; the road network connectivity assessment model uses Dijkstra's algorithm to simulate post-disaster changes in road network connectivity, and the connectivity index is calculated as follows: the sum of the products of the capacity weights of all road segments and the importance coefficients of the road segments divided by the sum of the importance coefficients of all road segments, where the importance coefficients of the road segments are assigned according to the road level, with 3 for expressways, 2 for national highways, and 1 for provincial and county roads. In other embodiments, the capacity weights are determined by looking up Table 2.

[0031] In some embodiments, the specific steps of the multi-index comprehensive resilience scoring in S3 are as follows: the weight of each index is determined by the analytic hierarchy process (AHP), wherein the connectivity index weight is 0.35, the recovery timeliness index weight is 0.3, the traffic capacity recovery rate weight is 0.2, and the critical facility connectivity rate weight is 0.15; the recovery timeliness index is the ratio of the expected recovery time to the standard recovery time, and the standard recovery time is determined by referring to Table 1 based on the road unit damage level; the traffic capacity recovery rate is the ratio of the post-disaster traffic capacity to the pre-disaster traffic capacity, and the post-disaster traffic capacity is comprehensively calculated by referring to Table 2 based on the road unit damage level and road level; the critical facility connectivity rate is the proportion of the number of connected core facilities to the total number of core facilities, and the connectivity status of each core facility is determined by referring to Table 3 based on the damage level of its respective road unit.

[0032] In some embodiments, the parameters of the genetic algorithm in S4 are set as follows: population size 100, crossover probability 0.8, mutation probability 0.05, selection strategy using roulette wheel combined with elite retention, retaining the top 10% of individuals, maximum number of iterations 200 generations, and termination condition when fitness improves by less than 1% for 20 consecutive generations; the comprehensive fitness function includes: time-discounted summation of the road network performance indicators at each stage of the repair process, plus a weighted summation of the core facility connectivity rewards at each stage, wherein the time discount factor is 0.92, and the core facility connectivity reward is 10 points when connected and 0 points when not connected.

[0033] In some embodiments, the road network performance index in S4 is a weighted sum of the connectivity index and the capacity recovery rate, wherein the connectivity index is calculated in real time by the road network connectivity assessment model, and the capacity recovery rate is the ratio of the total road network capacity at that moment to the total road network capacity before the disaster; the core facilities include hospitals, emergency command centers and material warehouses, and their weights are determined according to the importance level of the emergency response, with hospitals and emergency command centers taking a weight of 5 and material warehouses taking a weight of 3.

[0034] In some embodiments, the specific steps of dynamically adjusting the recovery path based on real-time feedback in S5 are as follows: periodically evaluate the deviation between the actual recovery time of the repaired road segment and the standard recovery time in Table 1. When the deviation exceeds a preset threshold, re-trigger the genetic algorithm optimization in S4, take the current unrepaired road segment set as the initial state, update the start time of the time discount factor in the fitness function, and generate the adjusted repair sequence.

[0035] In some embodiments, the earthquake data collected in S1 includes epicenter location, focal depth, magnitude, focal mechanism, peak ground motion parameters, spectral characteristic parameters, active fault data, and duration parameters; the slope data includes slope, soil type, vegetation cover, slope geometry, geological conditions, load, acceleration, displacement, and stress distribution; the traffic flow and road network data includes road geometry, structural characteristics of bridges and tunnels, seismic design parameters, traffic flow, road structure, and construction and maintenance data of traffic units; the historical disaster data includes damage type, damage scale, and information on the disaster-causing slope; when standardizing the above data, the Min-Max standardization method is used, missing values ​​are filled using linear interpolation, and road network topology parameters (such as "node-edge adjacency relationship, road segment length, and road grade weight") are extracted based on graph theory algorithms.

[0036] The technical concept of this invention is as follows: S1. Database Construction and Data Integration: First, collect multi-source data including earthquake, slope, traffic flow, and road network structure data, specifically including: Earthquake data: Historical earthquake data is collected, including epicenter location, focal depth, magnitude, focal mechanism, peak ground motion parameters, spectral characteristic parameters, active fault data (fault location, strike, dip, length, maximum potential earthquake magnitude (Mmax)), duration parameters, and composite and derived parameters.

[0037] Slope data: slope gradient, soil type, vegetation cover, slope geometry, geological conditions, load, acceleration, displacement, stress distribution, and other slope response characteristics data.

[0038] Traffic flow and road network data: Collect basic information on the road network in the disaster area, including road geometry, structural characteristics of bridges and tunnels, their seismic design parameters, traffic flow, road structure, and construction and maintenance data of traffic units.

[0039] Historical disaster data: Collect indicators of the impact of historical earthquake slope disasters on the road network, including damage type, damage scale, and information on the disaster-causing slope.

[0040] The collected data were integrated and standardized to establish a database for subsequent modeling. A systematic approach was used to process the multi-source heterogeneous data. In the data preprocessing stage, linear interpolation was used to fill in missing values ​​and remove obvious outliers; graph theory algorithms were used to extract road network topology parameters. In the database integration stage, the Min-Max standardization method was used to unify the data scale. The final model input data consisted of four dimensions: seismic motion parameters (PGA, PGV, Arias intensity, significant duration, etc.), slope geometric and geological parameters (slope, soil type, vegetation cover, elevation, Vs30, fault distance, dip angle, etc.), road network topology parameters (road segment connectivity, road grade, etc.), and basic traffic demand data, providing a complete and reliable data foundation for subsequent resilience assessment modeling.

[0041] S2. Slope Disaster Traffic Unit Damage Assessment: In this step, based on the database of earthquake and slope data collected in S1, an earthquake slope hazard assessment model is constructed using an artificial neural network method (setting the number of hidden layers to 2, the learning rate to 0.001, and the number of iterations to use early stopping (patience value = 20)) and the Apriori algorithm (setting the minimum support to 0.2 and the minimum confidence to 0.8). Based on the model, high-risk slope units are obtained. The model form is as follows: ;

[0042] In the formula, This represents the probability of slope hazard (a continuous value between 0 and 1). The input feature vector consists of 15 dimensions, including 4 types of seismic features (PGA, PGV, Arias intensity, significant duration), 4 types of slope features (slope, soil type, vegetation coverage, elevation), and 3 types of geological features (Vs30, fault distance, stratum dip angle). The input feature vector is a binary association rule feature mined by the Apriori algorithm. The training uses the Adam optimizer (learning rate = 0.001), batch size = 32, maximum iterations of 500 rounds, and an early stopping strategy. : Association rule feature values, based on the Apriori algorithm (minimum support 0.2, minimum confidence 0.8) to mine association rule features, high-risk feature combinations obtained through frequent itemset mining, each rule generates a binary feature to indicate whether the combination exists. The weight matrix and bias terms of the first hidden layer (12 nodes). : The weight vector and bias term of the second hidden layer (8 nodes). The ReLU activation function, ReLU(z) = max(0, z), introduces a nonlinear transformation. Sigmoid function, σ(z) = 1 / (1+e^(z-1)) -z ), which maps the output to the 0-1 probability interval.

[0043] Based on the aforementioned slope hazard assessment model, a probability threshold P0 (P0=0.7) is set, and slope units with output probabilities P>P0 are classified as high-risk slope units. For these high-risk slope units, the nearest active fault is searched in GIS, and the "Mmax" of this fault is assigned as the "design magnitude" of the unit. The shortest distance from the slope unit location to the fault trace is calculated, and considering the fault dip angle, it is converted into the distance to the fault plane (Rrup). Depending on the seismic tectonic zone where the assessment area is located (e.g., eastern / western China), a seismic motion attenuation relationship that has been validated for regional adaptability (e.g., the applicable version in the NGA-West2 series models, or the recommended model in the "China Seismic Ground Motion Parameter Zoning Map") is selected. Input parameters: magnitude = Mmax, distance = Rrup, site condition = the unit's Vs30 value. The PGA-MCE of the unit under the maximum credible earthquake (MCE) scenario is obtained.

[0044] Relating PGA-MCE to the critical acceleration a c (a) c Using the empirical relation F S =1 / (1+βP) converts the hazard probability P output by the ANN model into a static safety factor F. S Where β is the calibration parameter (β is 3.0 for cohesive soil slopes and 2.0 for gravelly or sandy soil slopes), then F S 'a' is calculated using the infinite slope formula. c ( Where g is taken as 9.81 m / s 2 , The ratio of slope unit slope (PGA-MCE / a) c As input parameters for seismic intensity, combined with slope geometry (slope, sliding surface length) and geotechnical physical and mechanical parameters (unit weight), internal friction angle Based on the principle of energy conservation, a rapid calculation model for earthquake-induced landslide movement was established using topographic data and the model was verified in a classic landslide case. The relative error between the sliding distance and the measured value was controlled within 15%, demonstrating good calculation accuracy.

[0045] The formula for calculating energy conservation is: ;

[0046] In the formula, This represents the decrease in the potential energy of the sliding body (potential energy decrease). The energy provided by an earthquake to the sliding body motion system (earthquake energy). This refers to the energy dissipation caused by friction on sliding surfaces (friction energy loss). The energy change caused by topographic relief (topographic energy). Energy dissipation (deformation energy) is generated by the relative motion inside the sliding body. This represents the increase in the total kinetic energy of the sliding body.

[0047] Each energy term is represented as follows: ;

[0048] ;

[0049] In the formula, For the mass of the sliding body, It is the acceleration due to gravity. This represents the vertical displacement of the center of mass of the sliding body. The coefficient of friction of the sliding surface. The residual internal friction angle of the sliding surface is the shear angle. The inclination angle of the sliding surface. This represents the displacement increment along the sliding surface.

[0050] Earthquake energy Calculated based on the input ground motion time history (PGA-MCE and duration); topographic energy Determined by the curvature of the sliding surface and the velocity of the sliding body; deformation energy The calculation is simplified by adopting the uniform deformation assumption. The calculation methods for seismic energy, topographic energy, and deformation energy all employ mature models already published in this field. Specifically, the seismic energy is calculated based on the energy balance method (Kokusho & Ishizawa, 2004), obtaining the total energy contribution of the seismic input to the landslide system by integrating the square of the time history of the input ground motion velocity; the topographic energy is calculated considering the influence of changes in the geometry of the sliding surface on the landslide motion, using the topographic correction method proposed by Xu Guangxing et al. (2010); the deformation energy adopts the uniform deformation assumption (Xu Guangxing et al., 2010; Zhang Yingbin et al., 2024), simplifying the energy dissipation of relative motion within the landslide into overall deformation work. These energy calculation methods have been thoroughly theoretically verified and applied in the field of earthquake landslide motion analysis. This invention integrates them into a cross-scale coupled model; specific calculation formulas can be found in relevant literature. By solving the above energy balance equations, the kinetic energy of the landslide at each moment during its motion can be obtained, thereby obtaining the sliding distance D, sliding velocity V, and landslide impact force F. ;

[0051] ;

[0052] ;

[0053] In the formula, The average deceleration of the sliding body can be obtained from the differential relationship between the change in kinetic energy and the sliding distance.

[0054] Based on historical road damage cases, regression analysis was used to establish a nonlinear mapping relationship between disaster parameters such as landslide impact force, sliding velocity, and sliding distance and road damage indicators (such as crack density and bridge / culvert fracture rate). A correlation model between disaster movement and road unit damage level was also established. The model is as follows: ;

[0055] In the formula, Output raw data for the damage level of road units; the larger the value, the more severe the damage. The regression coefficients were determined using multivariate nonlinear regression analysis based on landslide impact force F, sliding velocity V, sliding distance D, and corresponding road damage level Z data collected from historical earthquake slope disaster cases. The formula for scaling to the (0, 5) interval using the Sigmoid function is: The damage levels are as follows: Level 1: 0.0≤D<1.5 Minor damage; Level 2: 1.5≤D<2.5 Moderate damage; Level 3: 2.5≤D<3.5 Severe damage; Level 4: 3.5≤D<4.5 Complete interruption; Level 5: 4.5≤D≤5.0 Devastating damage.

[0056] S3. Road Network Resilience Assessment: Based on the damage levels of traffic units and road grades (expressways, national highways, provincial / county roads) assessed in S2, the capacity weights of each road unit are determined by referring to Table 2, and a road network connectivity assessment model is then constructed. The Dijkstra algorithm for network analysis is used to simulate changes in road network connectivity under different post-disaster scenarios, quantitatively assessing the mobility of the post-disaster highway network. The road network connectivity assessment model is as follows: ;

[0057] In the formula, The connectivity index is the final output of the model, and its value ranges from 0 to 1. The larger the value, the better the overall connectivity of the road network. A value of 1 represents a completely unobstructed ideal state of the road network, while a value of 0 represents a completely paralyzed road network. The total number of road segments (a positive integer) represents the sum of all road units within the assessment area. This is the road segment index parameter, used as an identifier to traverse and calculate all road segments. Its value ranges from 1 to n, ensuring that each road unit is included in the evaluation system. It is the traffic capacity weight of the i-th road segment. This is the importance coefficient of the i-th road segment (assigned according to road level: 3 for expressways, 2 for national highways, and 1 for provincial / county roads). Among them, expressways are the most important traffic arteries and are assigned a value of 3; national highways are the main traffic arteries and are assigned a value of 2; provincial and county roads are general traffic roads and are assigned a value of 1.

[0058] Building upon this foundation, a resilience assessment system is further constructed, incorporating the connectivity index as one of the core indicators. This system comprehensively scores the resilience of the post-disaster highway network based on multiple indicators, including recovery timeliness, traffic capacity recovery rate, and critical infrastructure connectivity rate. The assessment results can provide quantitative support for post-disaster recovery decision-making, helping decision-makers formulate appropriate recovery plans.

[0059] ;

[0060] in This is a resilience assessment index. For the first The weights of each indicator are 0.35, 0.3, 0.2, and 0.15 respectively (the weights are determined by the Analytic Hierarchy Process (AHP). The road network connectivity index is determined by the road network connectivity assessment model. The (restoration timeliness index) is the ratio of the expected recovery time to the standard recovery time. The expected recovery time can be obtained in advance through the optimization results of the genetic algorithm in S4 (i.e., it is obtained by summing the optimal repair sequence and the standard recovery time of each road segment). The standard recovery time is determined according to the road damage level by referring to Table 1. (Traffic capacity recovery rate) is the ratio of post-disaster traffic capacity to pre-disaster traffic capacity. Post-disaster traffic capacity is calculated by referring to Table 2 based on the road damage level and road grade. (Critical facility connectivity rate) is the proportion of connected core facilities (hospitals, emergency command centers, material warehouses, etc.) to the total number of core facilities. The connectivity status of each core facility is determined by referring to Table 3 based on the damage level of its road unit. The calculation formula can be expressed as: ;

[0061] In the formula, The total number of core facilities, The connectivity status value for the k-th core facility is determined according to Table 3. When the facility is connected... When not connected .

[0062] It is important to note that if there is no expected recovery time initially, the expected recovery time should be replaced with a recovery time estimated based on the damage level, such as a weighted average recovery time estimated in ascending order of damage level, and the initial resilience assessment should be completed independently. After S4 optimization, the "Expected Resilience Assessment Index" should be calculated separately for comparison.

[0063] S4. Intelligent Recovery Path Optimization: Based on the damage level of S2 unit and the resilience assessment results of S3 road network, a genetic algorithm (with the following settings: population size = 100, crossover probability = 0.8, mutation probability = 0.05, selection strategy: roulette wheel selection + elite retention (retaining the top 10% of individuals), maximum number of iterations = 200 generations, and termination condition: fitness improvement of less than 1% for 20 consecutive generations) was used to intelligently optimize the recovery path of the post-disaster road network.

[0064] The post-disaster road network restoration path optimization method based on genetic algorithms aims to intelligently search for the optimal repair scheme by simulating a natural evolutionary process. This method first digitally encodes the damaged road segments to be repaired, with each repair scheme represented as a repair sequence. For example, [3,1,4,2] represents the sequence of repairing segment 3 first, then segment 1, then segment 4, and finally segment 2. Subsequently, a fitness function is used to quantitatively evaluate each repair scheme. This function comprehensively considers the road network performance recovery speed and the priority of critical infrastructure connectivity, assigning a comprehensive score to each scheme. The fitness function used in the genetic algorithm to evaluate the merits of repair schemes is defined as follows: ;

[0065] In the formula, The overall fitness score represents the repair solution (the higher the score, the better the solution). This is the stage number of the repair process, starting from the initial state t=0 and gradually increasing as the repair work progresses until all damaged road sections are repaired. This indicates the total number of damaged road sections that need repair. It is the time discount factor (value 0.92), which reflects the principle of "early recovery of high value". Its value is based on the discount rate commonly used in engineering economic analysis. Total number of critical facilities. The weight of the kth core facility is determined based on the importance level of the facility in emergency response, with hospitals and emergency command centers receiving a weight of 5 (highest priority) and material warehouses receiving a weight of 3 (higher priority). The reward points for the connectivity status of the kth facility are 10 for connectivity and 0 for non-connectivity. The road network performance index (a combined value of connectivity and capacity, ranging from 0 to 10) for stage t is calculated using the following formula: ;

[0066] Where C(t) is the road network connectivity index at time t (calculated by the road network connectivity assessment model). Let be the total capacity of the road network at time t. The total traffic capacity of the road network before the disaster. This represents the traffic capacity recovery rate.

[0067] During the optimization process, the system first randomly generates a batch of initial repair schemes as the first-generation population. Then, it uses a roulette wheel selection method to select the high-scoring, excellent schemes as the parent generation. Next, sequential crossover is used to combine the advantages of different parent generation schemes to generate a new generation of repair schemes. This process is similar to intelligently splicing together the distinctive features of multiple excellent schemes. Simultaneously, the system randomly fine-tunes the schemes with a certain probability, maintaining scheme diversity and avoiding getting trapped in local optima by exchanging or adjusting the repair order of some road segments. Finally, these newly generated schemes, after selection, combination, and fine-tuning, form the next generation population, and the above evolutionary process is repeated. When a predetermined number of iterations is reached or the scheme quality stabilizes and no longer significantly improves, the algorithm terminates and outputs the optimal repair sequence of all generations, providing decision-makers with scientifically sound and reasonable suggestions for post-disaster road network recovery.

[0068] After obtaining the optimal repair sequence, the road network status after the repair is completed (i.e., after all damaged road sections are repaired according to the recovery time in Table 1, the traffic capacity is restored to the good level) corresponding to the sequence is input into the resilience assessment model of S3 to calculate the expected resilience assessment index after repair. Meanwhile, the resilience index at the initial moment after the disaster is recorded as... .like This quantitatively proves the theoretical improvement effect of the optimization scheme, and the improvement range. This can serve as a basis for comparing different candidate solutions. This anticipated assessment can be completed before actual construction, providing decision-makers with a scientific reference.

[0069] S5. Recovery Progress Assessment and Real-time Adjustment: After all the actual repair work was completed, the S3 resilience assessment model was called again to calculate the actual resilience assessment index based on the actual restored road network condition. .Will Compared with the expected resilience index obtained in S4 Compare and calculate the relative deviation. Based on the validation results in S2, where the model prediction error was controlled within 15%, the deviation threshold was set... Set it to 15%. If the accuracy is ≤ 15%, the model prediction is considered accurate and the method is effective; if If the deviation exceeds 15%, deviation analysis is triggered to check whether the actual recovery time exceeds the standard recovery time in Table 1, whether the repair quality meets the standards, and whether secondary disasters have occurred. Deviation information is then fed back to the database for subsequent model iteration and optimization. Furthermore, during the recovery process, the deviation between the actual recovery progress and the plan is periodically evaluated. If the deviation exceeds a preset threshold, the genetic algorithm optimization in S4 is re-triggered to dynamically adjust the subsequent repair sequence. The preset threshold is set at 20% of the standard recovery time.

[0070] Table 1. Correspondence between Road Unit Damage Levels and Standard Recovery Time

[0071] Note: The standard recovery time is based on a comprehensive assessment of engineering repair quotas and historical emergency repair data. The values ​​in the table are typical values ​​and may be adjusted appropriately according to regional resource conditions in actual application.

[0072] Table 2. Correspondence between Road Unit Damage Level and Traffic Capacity Weight

[0073] Note: The capacity weight reflects the ratio of the actual capacity of a road section after a disaster to its pre-disaster capacity. Under the same damage level, the higher the road grade, the higher the priority for emergency repairs and the higher the capacity retention rate.

[0074] Table 3. Determination Table of Road Unit Damage Level and Core Facility Connectivity Status

[0075] Note: Core facilities include key nodes such as hospitals, emergency command centers, supply warehouses, fire stations, and refuge areas. If the location of a core facility or all its connecting roads are impassable, it is considered disconnected. For road units with a damage level of 3, during the restoration process, once the road section is reopened, its connectivity value changes from 0 to 1.

[0076] The embodiments described above are for illustrative purposes only and are not intended to limit the invention. Therefore, any changes in numerical values ​​or substitutions of equivalent elements should still fall within the scope of this invention.

[0077] The above detailed description will enable those skilled in the art to understand that the present invention can indeed achieve the aforementioned objectives and has complied with the provisions of the Patent Law.

[0078] Although preferred embodiments of the invention have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments as well as all changes and modifications falling within the scope of the invention. The above descriptions are merely preferred embodiments of the invention and are not intended to limit the invention. It should be noted that any modifications, equivalent substitutions, and improvements made within the spirit and principles of the invention should be included within the scope of protection of the invention.

[0079] It should be noted that the above description of the process is for illustrative purposes only and does not limit the scope of this specification. Those skilled in the art can make various modifications and changes to the process under the guidance of this specification. However, these modifications and changes remain within the scope of this specification.

[0080] The basic concepts have been described above. Obviously, for those skilled in the art who have read this application, the above disclosure is merely illustrative and does not constitute a limitation of this application. Although not explicitly stated herein, those skilled in the art may make various modifications, improvements, and corrections to this application. Such modifications, improvements, and corrections are suggested in this application, and therefore, such modifications, improvements, and corrections still fall within the spirit and scope of the exemplary embodiments of this application.

[0081] Furthermore, this application uses specific terms to describe its embodiments. For example, "an embodiment," "one embodiment," and / or "some embodiments" refer to a particular feature, structure, or characteristic related to at least one embodiment of this application. Therefore, it should be emphasized and noted that "an embodiment," "one embodiment," or "an alternative embodiment" mentioned twice or more in different positions in this specification do not necessarily refer to the same embodiment. In addition, certain features, structures, or characteristics in one or more embodiments of this application can be appropriately combined.

[0082] Furthermore, those skilled in the art will understand that aspects of this application can be described and illustrated through several patentable types or situations, including any new and useful combination of processes, machines, products, or substances, or any new and useful improvements thereof. Therefore, aspects of this application can be implemented entirely in hardware, entirely in software (including firmware, resident software, microcode, etc.), or a combination of hardware and software. All of the above hardware or software can be referred to as a “unit,” “module,” or “system.” Furthermore, aspects of this application can take the form of a computer program product embodied in one or more computer-readable media, wherein computer-readable program code is contained therein.

[0083] The computer program code required for the operation of each part of this application can be written in any one or more programming languages, including object-oriented programming languages ​​such as Java, Scala, Smalltalk, Eiffel, JADE, Emerald, C++, C#, VB.NET, and Python; general programming languages ​​such as C; Visual Basic, Fortran2103, Perl, COBOL2102, PHP, and ABAP; dynamic programming languages ​​such as Python, Ruby, and Groovy; or other programming languages. This program code can run entirely on the user's computer, or as a standalone software package on the user's computer, or partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In the latter case, the remote computer can be connected to the user's computer via any network, such as a local area network (LAN) or wide area network (WAN), or connected to an external computer (e.g., via the Internet), or in a cloud computing environment, or used as a service such as Software as a Service (SaaS).

[0084] Furthermore, unless expressly stated in the claims, the order of processing elements and sequences, the use of numbers and letters, or other names described in this application are not intended to limit the order of the processes and methods of this application. Although some currently considered useful embodiments of the invention have been discussed in the foregoing disclosure by way of various examples, it should be understood that such details are for illustrative purposes only, and the appended claims are not limited to the disclosed embodiments; rather, the claims are intended to cover all modifications and equivalent combinations that conform to the substance and scope of the embodiments of this application. For example, although the implementation of the various components described above can be embodied in a hardware device, it can also be implemented as a purely software solution, such as an installation on an existing server or mobile device.

[0085] Similarly, it should be noted that, in order to simplify the description of the present application and thus aid in the understanding of one or more embodiments of the invention, the foregoing description of the embodiments of the present application sometimes combines multiple features into a single embodiment, drawing, or description thereof. However, this approach of the present application should not be construed as reflecting an intention that the claimed subject matter requires more features than expressly recited in each claim. Rather, the subject of the invention should possess fewer features than in any single embodiment described above.

Claims

1. A method for intelligent assessment and improvement of the resilience of transportation networks in earthquake-prone slope disasters based on cross-scale coupling, characterized in that, include: S1: Collect earthquake data, slope data, traffic flow and road network data, and historical disaster data; standardize the collected data and establish a database. S2: Based on the database, an earthquake slope hazard assessment model is constructed using an artificial neural network and the Apriori algorithm to obtain high-risk slope units; for the high-risk slope units, a rapid calculation model for earthquake landslide movement is established based on the principle of energy conservation to calculate the sliding distance, sliding velocity, and landslide impact force; a nonlinear mapping relationship is established between the landslide impact force, sliding velocity, sliding distance, and road damage indicators to determine the damage level of each road unit; S3: Based on the damage level of each road unit, assign a traffic capacity weight to each road unit, construct a road network connectivity assessment model, and calculate the connectivity index; based on the connectivity index, combined with the recovery timeliness index, traffic capacity recovery rate, and key facility connectivity rate, conduct a multi-indicator comprehensive resilience score for the highway network to obtain a resilience assessment index. S4: Based on the damage level of each road unit, a genetic algorithm is used to intelligently optimize the recovery path of the post-disaster highway network, construct a comprehensive fitness function that includes a time discount factor and a critical infrastructure connectivity reward, search for the optimal repair sequence, and conduct an expected resilience assessment on the optimal repair sequence to obtain the expected resilience assessment index after repair. S5: During the recovery process, repairs are performed based on the optimal repair sequence. The actual recovery progress is periodically evaluated and compared with the plan. When deviations occur, the subsequent repair sequence is dynamically adjusted. After the repairs are completed, the actual resilience assessment index is calculated and compared with the expected value for verification, thereby achieving flexible response and model self-optimization in the recovery process. The specific steps in S2 for constructing an earthquake slope hazard assessment model using an artificial neural network and the Apriori algorithm are as follows: The Apriori algorithm is used to mine frequent itemsets in multi-source data, with a minimum support of 0.2 and a minimum confidence of 0.8, generating binary association rule features; these binary association rule features are combined with earthquake features, slope features, and geological features to form an input feature vector; an artificial neural network with two hidden layers is used, with the number of hidden layer nodes set to 12 and 8 respectively, employing the ReLU activation function and the Sigmoid output function, trained with the Adam optimizer, a learning rate of 0.001, and an early stopping method to output the slope hazard probability. A probability threshold of 0.7 is set, and slope units with a risk probability greater than 0.7 are classified as high-risk slope units. The specific steps in S2 for establishing a rapid calculation model of earthquake landslide motion for high-risk slope units are as follows: obtain active fault data at the location of the high-risk slope unit, and assign the maximum potential earthquake magnitude of the active fault as the design magnitude; calculate the shortest distance from the slope unit to the fault plane, and calculate the peak ground motion parameters under the maximum credible earthquake scenario by combining site condition parameters and the ground motion attenuation relationship; use the ratio of the peak ground motion parameters to the critical acceleration as the input parameter of ground motion intensity. Based on the principle of energy conservation, the decrease in potential energy of the sliding body, seismic energy, energy dissipation due to friction on the sliding surface, topographic energy, deformation energy, and the increase in total kinetic energy of the sliding body are calculated respectively. An energy balance equation is established and the sliding distance, sliding velocity, and landslide impact force are obtained by solving it. The specific steps for establishing the nonlinear mapping relationship between landslide impact force, sliding velocity, sliding distance and road damage indicators in S2 are as follows: Multiple nonlinear regression analysis is used, with landslide impact force, sliding velocity, and sliding distance as independent variables and the original data of road damage levels as dependent variables, to fit regression coefficients; the regression output is scaled to the 0-5 range using the Sigmoid function, and five damage levels are divided: Level 1: Minor damage; Level 2: Moderate damage; Level 3: Severe damage; Level 4: Complete interruption; Level 5: Devastating damage. The specific steps of S5 for dynamically adjusting the recovery path based on real-time feedback are as follows: periodically evaluate the deviation between the actual recovery time and the standard recovery time of the repaired road segment. When the deviation exceeds the preset threshold, the genetic algorithm optimization of S4 is triggered again. Taking the current unrepaired road segment set as the initial state, the start time of the time discount factor in the fitness function is updated to generate the adjusted repair sequence.

2. The intelligent assessment and improvement method for the resilience of transportation networks in earthquake-prone slope disasters based on cross-scale coupling as described in claim 1, characterized in that, The specific method for assigning capacity weights to each road unit in S3 is as follows: the capacity weights are determined based on the damage level and road grade of the road unit; the road network connectivity assessment model uses the Dijkstra algorithm to simulate post-disaster changes in road network connectivity, and the formula for calculating the connectivity index is: the sum of the products of the capacity weights of all road segments and the importance coefficients of the road segments divided by the sum of the importance coefficients of all road segments, where the importance coefficients of the road segments are assigned according to the road grade, with 3 for expressways, 2 for national highways, and 1 for provincial and county roads.

3. The intelligent assessment and improvement method for the resilience of transportation networks in earthquake-prone slope disasters based on cross-scale coupling as described in claim 2, characterized in that, The specific steps for the multi-indicator comprehensive resilience scoring in S3 are as follows: the weight of each indicator is determined using the analytic hierarchy process (AHP), where the connectivity index has a weight of 0.35, the recovery timeliness index has a weight of 0.3, the traffic capacity recovery rate has a weight of 0.2, and the critical facility connectivity rate has a weight of 0.15; the recovery timeliness index is the ratio of the expected recovery time to the standard recovery time; the traffic capacity recovery rate is the ratio of post-disaster traffic capacity to pre-disaster traffic capacity; and the critical facility connectivity rate is the proportion of the number of connected core facilities to the total number of core facilities.

4. The intelligent assessment and improvement method for the resilience of transportation networks in earthquake-prone slope disasters based on cross-scale coupling as described in claim 1, characterized in that, The parameters of the genetic algorithm in S4 are set as follows: population size 100, crossover probability 0.8, mutation probability 0.05, selection strategy using roulette wheel selection combined with elite retention, retaining the top 10% of individuals, maximum number of iterations 200 generations, and termination condition when fitness improves by less than 1% for 20 consecutive generations; the comprehensive fitness function includes: time-discounted summation of road network performance indicators at each stage of the repair process, plus weighted summation of core facility connectivity rewards at each stage, where the time discount factor is 0.92, and the core facility connectivity reward is 10 points for connectivity and 0 points for non-connectivity.

5. The intelligent assessment and improvement method for the resilience of transportation networks affected by earthquake slope disasters based on cross-scale coupling as described in claim 4, characterized in that, The road network performance index in S4 is a weighted sum of the connectivity index and the traffic capacity recovery rate. The connectivity index is calculated in real time by the road network connectivity assessment model, and the traffic capacity recovery rate is the ratio of the total traffic capacity of the road network at that moment to the total traffic capacity of the road network before the disaster. The core facilities include hospitals, emergency command centers, and material warehouses. Their weights are determined according to the importance level of the emergency response, with hospitals and emergency command centers receiving a weight of 5 and material warehouses receiving a weight of 3.

6. The intelligent assessment and improvement method for the resilience of transportation networks in earthquake-prone slope disasters based on cross-scale coupling as described in claim 1, characterized in that, The earthquake data collected in S1 includes epicenter location, focal depth, magnitude, focal mechanism, peak ground motion parameters, spectral characteristic parameters, active fault data, and duration parameters; the slope data includes slope, soil type, vegetation cover, slope geometry, geological conditions, load, acceleration, displacement, and stress distribution; the traffic flow and road network data includes road geometry, structural characteristics of bridges and tunnels, seismic design parameters, traffic flow, road structure, and construction and maintenance data of traffic units; the historical disaster data includes damage type, damage scale, and information on disaster-causing slopes; the Min-Max standardization method is used to standardize the above data, linear interpolation is used to fill in missing values, and road network topology parameters are extracted based on graph theory algorithms.

Images