An infrastructure flood resilience assessment method, device, equipment and medium

Abstract

The application discloses a kind of infrastructure flood resilience evaluation methods, comprising: using flood simulation coupling model to carry out the whole process simulation of urban flood in multiple scenarios to target area;From the simulation process, the spatial attribute index of blue-green-gray infrastructure is extracted as the independent variable, the flood dynamic index of different flood stages is extracted as the dependent variable, to form a sample data set, and the independent variable is divided into multiple independent variable groups according to function;Standardization is carried out on the independent variable and the dependent variable respectively, and data enhancement is carried out;Bayesian mixed effect model is constructed, prior distribution is set for model parameters, posterior distribution of parameters is obtained based on the enhanced sample data set using sampling algorithm, and the evaluation result is output according to the posterior distribution.The application can clearly reveal the dynamic action law of blue-green-gray infrastructure in initial rising period, flood peak period and recession period, output more stable and reliable phased dynamic resilience evaluation result, and significantly improve the safety guarantee level of urban flood control and drainage.

Description

Technical Field

[0001] This invention relates to the field of flood resilience assessment technology, and in particular to a method, apparatus, equipment and medium for assessing the flood resilience of infrastructure. Background Technology

[0002] In the field of urban flood resilience assessment, existing technologies typically employ multiple linear regression or traditional machine learning models (such as random forests and support vector machines) to establish a mapping relationship between urban spatial attributes (such as green space ratio and pipeline density) and flood outcomes (such as inundation depth and water volume). Some studies have also begun to utilize numerical simulation software (such as LISFLOOD and SWMM) to conduct scenario simulations to obtain data support.

[0003] However, existing technologies still have shortcomings: First, they ignore the dynamic time-varying characteristics and stage-specific contribution differences of the entire flood process, and only obtain a single resilience result through simple weighted summation, which masks the dynamic contribution changes of infrastructure at different stages such as the initial rise, peak flood, and receding flood, resulting in the assessment results being unable to accurately guide the precise regulation of cities during specific rainfall periods; Second, they are difficult to effectively handle the small sample dilemma and overfitting problem in hydrodynamic simulation. When faced with the combination and superposition of multiple "spatial schemes-climate scenarios", the number of effective samples generated is often extremely limited, which makes the prediction accuracy and generalization ability of the assessment model significantly decrease when facing unseen climate scenarios, making it difficult to provide robust resilience assessment conclusions. Summary of the Invention

[0004] This invention provides a method for assessing the flood resilience of infrastructure, which can clearly reveal the dynamic effects of blue, green, and gray infrastructure during the initial flood initiation, peak flood, and receding flood periods, and output more robust and reliable phased dynamic resilience assessment results, significantly improving the level of urban flood control and drainage safety.

[0005] In a first aspect, embodiments of the present invention provide a method for assessing the flood resilience of infrastructure, comprising: A flood simulation coupled model was used to simulate the entire process of urban flooding in the target area under multiple scenarios. Spatial attribute indicators of blue, green, and gray infrastructure were extracted as independent variables during the simulation process, and dynamic flood indicators of different flood stages were extracted as dependent variables to form a sample dataset. The independent variables were then divided into multiple groups according to their functions. The flood stages included the initial stage, the peak flood stage, and the receding flood stage. The independent and dependent variables are standardized respectively, and a phased resampling strategy is used for data augmentation to obtain an augmented sample dataset. A Bayesian mixture effect model is constructed, and a prior distribution is set for the parameters of the Bayesian mixture effect model. Based on the enhanced sample dataset, a sampling algorithm is used to obtain the posterior distribution of the parameters. According to the posterior distribution, the phased dynamic evaluation results of the flood resilience of blue-green-grey infrastructure are output.

[0006] Furthermore, the use of a flood simulation coupling model to simulate the entire process of urban flooding in the target area under multiple scenarios includes: Obtain historical hydrological and meteorological data of flood events in the target area, as well as predicted vertical data of sea level rise, and construct various hydrological and climatic flood scenarios based on the above data; For each flood scenario, corresponding boundary conditions are set, including water level, rainfall intensity, typhoon intensity, sea level rise prediction data, and land subsidence prediction data. The MIKE FLOOD coupled model was used to simulate the entire process of urban flooding; the coupled model includes the MIKE 11 one-dimensional river hydrodynamic module, the MIKE 21 two-dimensional surface runoff module, and the MIKE URBAN municipal pipe network drainage module.

[0007] Furthermore, the spatial attribute indicators of the blue-green-gray infrastructure include at least one of the following: water surface ratio, river network density, and water connectivity of blue infrastructure; green space ratio, sunken green space ratio, and green space connectivity of green infrastructure; average embankment crest elevation, municipal drainage network density, and average drainage pipe diameter of gray infrastructure; ecological embankment ratio, drainage outlet and overflow weir density, and average surface permeability of the degree of facility coupling; and the flood dynamic indicators include at least one of the following: water accumulation in built-up areas, water storage capacity of green infrastructure, water storage capacity of pipe networks, and inundated area where the product of water depth and flow velocity is ≥0.4 m² / s.

[0008] Furthermore, the standardization of the independent and dependent variables, respectively, and the data augmentation using a phased resampling strategy, yield an augmented sample dataset, including: The independent variables are standardized using Z-score. The dependent variable is subjected to Yeo-Johnson transformation to make it approximately normally distributed, and the transformed value is scaled to a preset range using Min-Max. The sample dataset was divided into three initial single-stage sample datasets according to the stages of flooding. Calculate the Pearson correlation coefficient between the independent and dependent variables in each initial single-stage sample dataset, and classify the independent variables into weakly correlated features and strongly correlated features based on the correlation strength. Each initial single-stage sample dataset is resampled with replacement, and adaptive Gaussian noise is injected according to the strength of the features during the resampling process to obtain an enhanced single-stage sample dataset. The three enhanced single-stage sample datasets are then merged to obtain the enhanced sample dataset.

[0009] Furthermore, the prior distribution for setting the parameters of the Bayesian mixture effect model includes: A Laplace prior distribution is applied to the fixed effects coefficients of the Bayesian mixed effects model, a normal prior is set for the intercept, and a semi-normal prior is set for the standard deviation of the random effects. The fixed effects are the deterministic contributions of the spatial attributes of blue-green-gray infrastructure to flood resilience, while the stochastic effects are the environmental uncertainties caused by differences in hydrological and climatic scenarios.

[0010] Furthermore, based on the enhanced sample dataset, a sampling algorithm is used to obtain the posterior distribution of the parameters. According to the posterior distribution, a phased dynamic assessment result of the flood resilience of blue-green-grey infrastructure is output, including: Based on the enhanced sample dataset, the Markov chain Monte Carlo sampling algorithm is used to sample the model parameters, and the posterior distribution of all parameters is obtained by fitting the sampling results. Extract the posterior mean and the 95% highest density interval of the parameters from the posterior distribution; Based on the 95% highest density range, the significance of the independent variables and independent variable groups on flood resilience was determined. Based on the posterior mean of the parameters, calculate the stage contribution of each independent variable and each group of independent variables to flood resilience at different flood stages; Based on the stage contribution value and significance determination results, output the phased dynamic assessment results of the flood resilience of blue, green and gray infrastructure.

[0011] Furthermore, the stage contribution value is the sum of the average base contribution over the entire flood cycle and the additional differentiated contribution in the corresponding flood stage.

[0012] Secondly, embodiments of the present invention provide an infrastructure flood resilience assessment device, comprising: The multi-scenario flood simulation module is used to simulate the entire process of urban flooding in a target area under multiple scenarios using a flood simulation coupling model. The feature extraction module is used to extract spatial attribute indicators of blue-green-gray infrastructure as independent variables and extract flood dynamic indicators of different flood stages as dependent variables from the simulation process to form a sample dataset. The independent variables are divided into multiple independent variable groups according to their functions. The flood stages include the initial stage, the flood peak stage, and the receding stage. The data processing and enhancement module is used to standardize the independent variable and the dependent variable respectively, and to perform data enhancement using a phased resampling strategy to obtain the enhanced sample dataset. The model building and evaluation output module is used to build a Bayesian mixture effect model, set a prior distribution for the parameters of the Bayesian mixture effect model, obtain the posterior distribution of the parameters based on the enhanced sample dataset using a sampling algorithm, and output the phased dynamic evaluation results of the flood resilience of blue-green-grey infrastructure according to the posterior distribution.

[0013] Thirdly, embodiments of the present invention provide an electronic device, comprising: Memory, used to store computer programs; A processor for executing the computer program; Wherein, when the processor executes the computer program, it implements the infrastructure flood resilience assessment method described in any of the first aspects above.

[0014] Fourthly, embodiments of the present invention provide a computer-readable storage medium storing a computer program that, when executed, implements the infrastructure flood resilience assessment method described in any of the first aspects above.

[0015] Compared with existing technologies, the present invention provides a method for assessing the flood resilience of infrastructure, which has the following advantages: It uses a flood simulation coupled model to simulate the entire process of urban flooding in a target area under multiple scenarios; it extracts spatial attribute indicators of blue-green-gray infrastructure as independent variables and extracts dynamic flood indicators of different flood stages as dependent variables from the simulation process, forming a sample dataset; and it divides the independent variables into multiple groups according to function; wherein the flood stages include the initial stage, the peak flood stage, and the receding stage; it standardizes the independent and dependent variables respectively, and uses a phased resampling strategy for data augmentation to obtain an augmented sample dataset; it constructs a Bayesian mixture effect model, sets a prior distribution for the parameters of the Bayesian mixture effect model, and uses a sampling algorithm to obtain the posterior distribution of the parameters based on the augmented sample dataset; and outputs the phased dynamic assessment results of the flood resilience of blue-green-gray infrastructure according to the posterior distribution. This invention can clearly reveal the dynamic role of blue-green-gray infrastructure in the initial flood stage, the peak flood stage, and the receding stage, outputting more robust and reliable phased dynamic resilience assessment results, and significantly improving the level of urban flood control and drainage safety. Attached Figure Description

[0016] To more clearly illustrate the technical features of the embodiments of the present invention, the drawings used in the embodiments of the present invention will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0017] Figure 1 This is a flowchart illustrating a method for assessing the flood resilience of infrastructure provided in an embodiment of the present invention; Figure 2 This is a schematic diagram of the structure of an infrastructure flood resilience assessment device provided in an embodiment of the present invention; Figure 3 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present invention. Detailed Implementation

[0018] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0019] It should be noted that although functional modules are divided in the device schematic diagram and a logical order is shown in the flowchart, in some cases, the steps shown or described may be performed in a different order than the module division in the device or the order in the flowchart. The terms "first," "second," etc., in the specification, claims, and the aforementioned drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence.

[0020] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains. The terminology used herein is for the purpose of describing embodiments of the invention only and is not intended to limit the invention.

[0021] In a first aspect, embodiments of the present invention provide a method for assessing the flood resilience of infrastructure, see [link to previous document]. Figure 1 This is a flowchart illustrating an embodiment of an infrastructure flood resilience assessment method provided by the present invention.

[0022] like Figure 1 As shown, the method includes the following steps: S1: Use a flood simulation coupled model to simulate the entire process of urban flooding in the target area under multiple scenarios; The flood simulation coupling model used can be a one-dimensional hydrodynamic model, a two-dimensional hydrodynamic model, or a one- or two-dimensional coupled hydrodynamic model, including but not limited to: urban stormwater models and hydrodynamic models such as MIKE URBAN, SWMM, and FLO-2D.

[0023] In terms of constructing multi-scenario simulations, multiple sets of hydro-climate scenarios can be formed by combining different rainfall recurrence periods, different rainfall durations, different typhoon paths, different sea level rises, different tidal levels, different topographic conditions, and different land use types.

[0024] S2: Extract spatial attribute indicators of blue, green, and gray infrastructure from the simulation process as independent variables, extract flood dynamic indicators of different flood stages as dependent variables, form a sample dataset, and divide the independent variables into multiple independent variable groups according to function; wherein, the flood stages include the initial period, the flood peak period, and the receding period; Among them, the spatial attribute indicators of blue, green and gray infrastructure can be extracted from three dimensions: spatial distribution, structural characteristics and functional parameters, by combining spatial data (such as vector layers and raster data) output by simulation models and infrastructure survey data of the target area with geographic information processing tools such as ArcGIS and QGIS.

[0025] Extracting dynamic indicators of floods at different stages requires first defining the time nodes of the three flood stages—initial stage, peak stage, and receding stage—based on dynamic data such as water depth, water range, and water flow velocity simulated throughout the entire flood process. Then, for each flood stage, dynamic indicators that can characterize flood features and reflect flood resilience levels are extracted from the simulation data.

[0026] The grouping of independent variables can be based on the functional types and mechanisms of blue, green, and gray infrastructure. All extracted spatial attribute indicators can be divided into multiple groups of independent variables with clear physical meanings. The specific grouping method can be flexibly set. Preferably, they can be divided into blue infrastructure group, green infrastructure group, gray infrastructure group, and coupling layer group.

[0027] S3: Standardize the independent variable and the dependent variable respectively, and use a phased resampling strategy to perform data augmentation to obtain the augmented sample dataset; Specifically, the independent variables are standardized to eliminate the dimensional differences of spatial attribute indicators of different types of infrastructure, such as the inconsistency of the dimensions of percentage indicators, length indicators, and elevation indicators, so as to avoid the model parameter estimation bias caused by dimensional differences. Specific standardization methods include, but are not limited to, Z-score standardization, Min-Max standardization, and range standardization.

[0028] The dependent variable is standardized to correct the skewed distribution of flood dynamic indicators, such as the tendency of indicators like the water accumulation area and water storage capacity in built-up areas to have extreme values ​​and exhibit a clear right-skewed distribution, thereby improving the numerical stability of subsequent Bayesian sampling.

[0029] Combining the three flood stages defined in step S2—the initial period, the peak period, and the receding period—a phased resampling strategy was adopted to augment the data and expand the sample size in a targeted manner, providing high-quality data support for the parameter estimation of the subsequent Bayesian mixed effects model.

[0030] S4: Construct a Bayesian mixture effect model, set a prior distribution for the parameters of the Bayesian mixture effect model, use a sampling algorithm to obtain the posterior distribution of the parameters based on the enhanced sample dataset, and output the phased dynamic evaluation results of the flood resilience of blue-green-grey infrastructure according to the posterior distribution.

[0031] Specifically, in line with the needs of flood resilience assessment of blue, green and gray infrastructure, fixed effects and random effects are introduced to balance the deterministic contribution of the infrastructure itself and the interference of environmental uncertainty. A Bayesian mixed effects model is constructed. The model can be constructed using Bayesian modeling tools such as Stan, PyMC3, and JAGS. The model structure can be flexibly adjusted according to the grouping of independent variables and the characteristics of flood stages.

[0032] Among them, fixed effects are used to quantify the deterministic contribution of the spatial attributes of blue, green and gray infrastructure to flood resilience (including the basic contribution of a single independent variable and the stage contribution of different flood stages), while random effects are used to quantify the environmental uncertainty interference caused by differences in hydrological and climatic scenarios, so as to ensure that the model can accurately capture the dual impact of infrastructure role and environmental disturbance, and adapt to the assessment needs of multiple scenarios and stages.

[0033] For the parameters of the Bayesian mixture effect model, a prior distribution is set. By combining the physical meaning of the model parameters with engineering practice, reasonable prior constraints are provided for the parameters to avoid parameter estimation divergence and overfitting. Through Bayesian posterior inference, the posterior distribution of all model parameters is accurately estimated by combining sample data and prior distribution. The specific sampling algorithm can be a Markov chain Monte Carlo (MCMC) algorithm, including but not limited to NUTS (Non-Rotating Sampler) algorithm, Metropolis-Hastings algorithm, etc.

[0034] Finally, by analyzing the posterior distribution characteristics, quantifying the contributions of independent variables and groups of independent variables, and outputting the phased dynamic assessment results of the flood resilience of blue-green-gray infrastructure, the core infrastructure types and optimization directions at different stages are analyzed in conjunction with the characteristics of flood stages, forming a visualized and quantitative assessment report, providing a scientific and accurate decision-making basis for urban flood control planning and optimization of the layout of blue-green-gray infrastructure.

[0035] In summary, this invention, by extracting dynamic flood indicators in stages, constructing a Bayesian mixture effect model, and quantifying the contribution of each stage, can clearly reveal the dynamic role of blue, green, and gray infrastructure during the initial flood initiation, peak flood, and receding flood periods. It outputs more robust and reliable staged dynamic resilience assessment results, which can provide refined guidance for urban flood control scheduling and facility regulation during different rainfall periods, significantly improving the city's emergency response capabilities for flood disasters and the level of flood control and drainage safety.

[0036] In one optional implementation, the step of using a flood simulation coupling model to simulate the entire process of urban flooding in the target area under multiple scenarios includes: Obtain historical hydrological and meteorological data of flood events in the target area, as well as predicted vertical data of sea level rise, and construct various hydrological and climatic flood scenarios based on the above data; For each flood scenario, corresponding boundary conditions are set, including water level, rainfall intensity, typhoon intensity, sea level rise prediction data, and land subsidence prediction data. The MIKE FLOOD coupled model was used to simulate the entire process of urban flooding; the coupled model includes the MIKE 11 one-dimensional river hydrodynamic module, the MIKE 21 two-dimensional surface runoff module, and the MIKE URBAN municipal pipe network drainage module.

[0037] Specifically, the process involves analyzing historical hydrological data on flood events in the target area, breaking down flood, tide, and waterlogging causative factors, clarifying the impact range and intensity characteristics of each causative factor, obtaining hydrological and meteorological data such as water level, rainfall, and typhoons by querying monitoring data from local water resources departments and reviewing relevant literature, and combining this with existing literature on future sea level rise (SLR) vertical data to set up gradient hydrological and climate sequence files and construct flood disaster scenarios with increasing risk.

[0038] Boundary conditions were set for each flood scenario, including water level, rainfall intensity, typhoon intensity, sea-level rise prediction data, and land subsidence prediction data. The methods for obtaining and setting the parameters for each boundary condition are as follows: Water level data was obtained by querying monitoring data from local water resources departments and reviewing relevant literature to ensure that the data reflects the characteristics of river water levels at different return periods; Rainfall intensity data was determined based on relevant standards for the study area, including the design storm return period and rainfall duration, and the rainfall distribution ratio of typical local rainfall patterns was selected to obtain the rainfall intensity for each time period; Typhoon data was obtained from typhoon monitoring information released by meteorological departments, including typhoon location, wind speed, wind direction, and air pressure, to simulate the impact of typhoons on surface runoff and river flow; Sea-level rise (SLR) prediction data was obtained based on the IPCC AR6 high-emission scenario. AR6, or the Sixth Assessment Report of the Intergovernmental Panel on Climate Change (IPCC), contains authoritative and scientifically sound sea-level rise predictions under high-emission scenarios. It projects a sea-level rise rate (SLR) of approximately 10.5 cm by 2050 and 28.0 cm by 2100. Land subsidence (LS) rates are based on data obtained from local geological surveys. Land subsidence leads to a decrease in topographic elevation and exacerbates flooding; it projects an LS rate of approximately 30.0 cm by 2050 and 80.0 cm by 2100.

[0039] For example, see Table 1 below, which shows examples of hydrometeorological boundary conditions for 12 flood scenarios: Table 1 Hydrological and Meteorological Boundary Conditions Understandably, in the table, "5a water level" and "10a water level" represent water levels corresponding to the return periods (once in 5 years and once in 10 years). The return period refers to the average number of years a certain hydrological phenomenon occurs once in a long period of observation, and it is the core parameter for setting flood scenarios. "5a6h design rainfall" represents the design rainfall intensity that occurs once in 5 years and lasts for 6 hours. SLR is sea level rise, and LS is land subsidence. The predicted SLR for 2050 is about 10.5cm and LS is about 30.0cm, and the predicted SLR for 2100 is about 28.0cm and LS is about 80.0cm.

[0040] The MIKE FLOOD coupled model was used to simulate the entire process of urban flooding. This model was implemented by coupling the Mike 11 one-dimensional river module, the Mike 21 two-dimensional surface runoff module, and the Mike Urban pipe network module in the Mike Flood software. The three modules work together to simulate the river flood discharge, surface runoff, and pipe network drainage processes, respectively.

[0041] Specifically, the construction process of Mike 11's one-dimensional river module is as follows: (1) Construction of River Network (.nwk 11) File: First, input the river network centerline file, which is used to define the spatial distribution of the river network in the target area; in the “Network” module of the MIKE 11 software, set the upstream / downstream connection relationship between the river channels according to the branch name (river channel name) and chainage (mileage) data to ensure the integrity of the river network topology and avoid river channel connection errors; in the “Structures” module, define the location of the sluice gate, the bottom of the gate (the elevation of the riverbed where the sluice gate is located) / the top of the gate (the elevation of the corresponding levee section), the width of the sluice gate and the control rules. In this embodiment, the water level is used to control the gate height, that is, the gate opening degree is automatically adjusted according to the change of the river water level to realize flood control and ensure that the sluice gate can operate flexibly according to the flood situation and play a flood control role.

[0042] (2) Construction of river cross section (.xns 11) files: Input batch river cross section data, check the cross section sequence, cross section shape at river connection, closure and elevation benchmark to ensure the accuracy and consistency of cross section data, and avoid deviation of simulation results due to cross section data errors; river roughness is assigned according to cross section. In this embodiment, Manning's n=0.025 is selected. Manning's roughness coefficient is used to reflect the roughness of the inner wall of the river and affects the water flow resistance. This value is consistent with the actual roughness characteristics of the river in the Pearl River Estuary area, ensuring the accuracy of water flow resistance simulation.

[0043] (3) Boundary condition (.bnd 11) file construction: The downstream uses the tidal process line of different return periods of Sanshakou Tide Gauge (Sanshakou Tide Gauge Station) as the water level boundary; the upstream uses a small flow inflow (0.01 m³ / s) as the stability boundary to avoid numerical instability during the simulation.

[0044] (4) Construction of hydrodynamic parameters (.hd 11) file: Initial conditions are set to Global Values, with Water Level set to 0.01m to ensure the stability of the initial state of the simulation; the other hydrodynamic parameters use the default parameters of the MIKE 11 software and do not need to be adjusted.

[0045] (5) Output MIKE11: Input the above-constructed river network file (.nwk 11), river cross section file (.xns 11), boundary condition file (.bnd 11), and hydrodynamic parameter file (.hd 11) into the simulation file (.sim11) of the MIKE 11 module, and set the initial condition type to HD: Parameter File (hydrodynamic parameter file); set the simulation time step to 2s to ensure the refinement of the simulation and accurately capture the dynamic changes of water level and flow; set the total duration to 24h to cover the entire flood process and keep it consistent with the simulation duration of other subsequent modules.

[0046] (6) Output results: Run the MIKE 11 module to simulate the changes in water level and flow in the river. Output a result file every 15 minutes. This result file will serve as the dynamic receiving condition for the MIKE 21 two-dimensional surface runoff module and the MIKE URBAN municipal pipeline drainage module, realizing the simulation of hydraulic exchange between the river and the surface and pipeline, and ensuring the continuity of the entire flood simulation.

[0047] The construction process of Mike 21's two-dimensional surface runoff module is as follows: (1) Construction of a 2D surface module (.m21) based on a structured grid: In the MIKE 21 software, “Module Selection” is set to “Hydrodynamic only” and includes the “Inland Flooding” option to adapt to urban flooding simulation scenarios; “Bathymetry” is input as the grid file (.dfs2) generated by integrating the DEM, and “Map projection” is set to UTM-49 to maintain consistency with the coordinate system of the MIKE 11 module and ensure spatial alignment of the simulation area; the simulation time step is also 2s, with a total duration of 24h, and a result file is output every 15min to ensure synchronization with the river channel simulation results; “Boundary” is always a closed boundary to prevent flood overflow from the simulation area and ensure the rationality of the simulation; “Flood and Dry” function is enabled, with Drying depth set to 0.005m and Flooding depth set to 0.005m. The depth (inundation depth) is 0.05m; the "Resistance" input is the surface roughness Manning value grid file (.dfs2) generated based on land use, and it is ensured that it is aligned with the Bathymetry to the same grid system. The Manning values ​​corresponding to different land use types are shown in Table 2 below; the hourly wind speed / direction time series file (.dfs0) is input into "Wind Condition" to simulate the impact of typhoons on surface runoff; in "Structure"—"Dikes", the dike Crest variation data (.xyz) is input one by one. The dike data is used to simulate the constraint effect of dikes on flood overflow, ensuring that the simulation fits the actual flood scenario.

[0048] Table 2 Manning values ​​corresponding to different land use types (2) Output results: Run the MIKE 21 module to simulate the dynamic changes of surface water accumulation. Output a result file every 15 minutes, including core data such as water depth, flow velocity and inundation range. These data will be used to calculate the subsequent dynamic response index (dependent variable) of flood, characterize the surface flood state at different flood stages, and provide data support for subsequent resilience assessment.

[0049] The construction process of the MIKE URBAN municipal drainage module is as follows: (1) Pipeline and node data organization: Create a new MIKE URBAN project and enable the Working Mode of MOUSE / SWMM. Both MOUSE and SWMM are mature pipeline simulation models. Enabling this mode can improve the accuracy of pipeline simulation. The Coordinate System is UTM-49, which is consistent with the MIKE 11 and MIKE 21 modules to ensure spatial alignment. Import drainage system topology data, including vector data of rainwater wells, drainage outlets, and pipe segments. These data are used to define the connection relationship of the municipal pipeline network. Use the "Project check tool" to check and correct topology errors, such as incorrect pipe segment connections and duplicate nodes, to ensure the integrity and accuracy of the pipeline network topology and avoid deviations in simulation results due to topology errors.

[0050] (2) Catchment area and impermeability parameterization: The study area after removing the river surface was imported as the basic range of the catchment area; the "Catchment Delineation Wizard" was used to delineate the catchment area to ensure that the catchment area range matched the actual topography and pipeline layout; the "Catchment Connection Wizard" was used to connect the storm drains to the sub-catchment areas to ensure that the surface runoff of the sub-catchment areas could be smoothly incorporated into the storm drains and enter the pipeline system; based on land use surface files (.shp, Table 3) with different impermeability, they were connected to the catchment area through "Catchment Processing" to reflect the influence of underlying surface changes on runoff generation. The impermeability of the underlying surface for different land use types is shown in Table 3 below: Table 3 Impermeability of underlying surfaces for different land use types (3) Set rainfall boundary conditions: Use the rainstorm pattern formula to generate hourly rainfall intensity process lines as MOUSE rainfall input, drive the entire process of runoff generation in the catchment area—well entry—pipeline discharge—node overflow, and realize the coordinated simulation of surface runoff and pipeline drainage.

[0051] (4) Output results: Run the MIKE URBAN module to output core data such as node water level, overflow, pipe section flow and full pipe degree, and outlet outflow process. These data are used to identify drainage bottlenecks and contributions to waterlogging.

[0052] The coupling settings for MIKE FLOOD are as follows: (1) Connect MIKE 11-MIKE 21, namely, Lateral links: In the MIKE FLOOD software, set the connection elevation to the top elevation of the embankment. When the river water level is higher than the top elevation of the embankment, the flood in the river overflows to the two-dimensional surface through the lateral connection. When the two-dimensional surface water level is higher than the river water level, the surface water flows back to the river through the lateral connection, realizing the two-way hydraulic exchange between the river and the surface.

[0053] (2) Connecting MIKE URBAN-MIKE 11, i.e. River / Urban links: The connection point is set at the drainage outlet of the pipe network. This connection allows the outflow of the pipe network to be dynamically controlled by the river water level, thereby depicting the drainage attenuation and intensification of waterlogging caused by the high tide level. Specifically, when the river water level rises, it will raise the boundary water level of the drainage outlet, reduce the water flow capacity of the pipe network, and cause poor drainage of the pipe network and intensified overflow at the nodes. When the river water level falls, the boundary water level of the drainage outlet decreases, the drainage capacity of the pipe network is restored, and the surface water is driven to be discharged through the pipe network, realizing the hydraulic coupling between the pipe network and the river.

[0054] (3) Connecting MIKE URBAN-MIKE 21, i.e. Urban links: establish an exchange relationship with the two-dimensional grid through rainwater wells, in which the rainwater wells connected to the sunken green space are set as overflow weirs; when the surface water reaches a certain depth, the surface runoff flows into the pipe network through the rainwater wells and is discharged by the pipe network; when the pressure in the pipe network is too high, the water in the pipe network overflows to the surface through the rainwater wells, forming or aggravating waterlogging, realizing bidirectional hydraulic exchange between the surface and the pipe network.

[0055] After completing the construction of all the MIKE FLOOD coupled models, the acquisition of hydrological and meteorological data, the setting of boundary conditions and coupling settings, the model was run to complete the 24-hour full-process simulation of flooding under 12 hydrological and climate scenarios.

[0056] This embodiment establishes three types of bidirectional hydraulic connections by coupling the MIKE FLOOD with the MIKE 11, MIKE 21, and MIKE URBAN modules, thus fully recreating the entire flood process from river channel to surface to pipe network. Combined with 12 gradient hydrological and climate scenario settings, it comprehensively covers flood scenarios of different levels of danger in the target area. At the same time, through refined boundary condition settings, it improves the accuracy and reliability of the simulation results, solving the technical problems of traditional hydrodynamic simulation being unable to achieve multi-module collaboration and incomplete scenario coverage.

[0057] In one optional implementation, the spatial attribute indicators of the blue-green-gray infrastructure include at least one of the following: water surface ratio, river network density, and water connectivity of blue infrastructure; green space ratio, sunken green space ratio, and green space connectivity of green infrastructure; average embankment crest elevation, municipal drainage network density, and average drainage pipe diameter of gray infrastructure; ecological embankment ratio, drainage outlet and overflow weir density, and average surface permeability of the degree of facility coupling; and the flood dynamic indicators include at least one of the following: water accumulation in built-up areas, water storage capacity of green infrastructure, water storage capacity of pipe networks, and inundated area where the product of water depth and water flow velocity is ≥0.4 m² / s.

[0058] Specifically, the Blue Infrastructure (BI) indicator system is primarily used to characterize the spatial layout, scale, and connectivity of blue infrastructure, reflecting its flood control and storage capacity. It includes three indicators: (1) Water surface ratio WSR , The area of ​​open surface water bodies (such as rivers, lakes, canals, wetlands, etc.) within the study area is used to characterize the area supply level and potential water storage capacity of blue space. The specific calculation formula is as follows: ; in, The total water surface area within the study area. This represents the total area of ​​the study region.

[0059] (2) Channel network density CND , The value represents the total length of the river centerline per unit area, reflecting the spatial density of the blue water system and the linear supply intensity of the discharge channels. A larger value indicates a denser water network and a richer potential drainage path. The specific calculation formula is as follows: ; in, The total length of all river centerlines. This represents the total area of ​​the study region.

[0060] (3) Blue infrastructure connectivity BIC , ), based on graph theory The index characterizes the connectivity and closed-loop degree of the blue network structure, reflecting the redundancy, number of alternative paths, and connectivity stability of the river network. A higher value indicates a greater degree of loop formation and stronger structural redundancy in the water network, meaning it still possesses multi-path transport capabilities even when locally blocked. The specific calculation formula is as follows: ; in, This represents the total number of connections (such as rivers, canals, etc.) in the Blueway network. This represents the total number of nodes in the network (such as tributary confluences, lake outlets, etc.).

[0061] Furthermore, the Green Infrastructure (GI) indicator system is primarily used to characterize the spatial supply, retention capacity, and connectivity of green infrastructure, reflecting its interception, infiltration, and flood storage capabilities. Specifically, it includes three indicators: (1) Green space ratio GSR , The green space (such as parks, lawns, and vegetated areas) is the proportion of green space (such as parks, lawns, and vegetated areas) to the total area of ​​the study region. It is used to characterize the overall supply level of green space and its potential contribution to infiltration, retention, and ecological regulation. The specific calculation formula is as follows: ; in, This represents the total green space area within the study area. This represents the total area of ​​the study region.

[0062] (2) Sunken green space ratio SGSR , The value represents the proportion of sunken green space to the total green space area. It is used to characterize the proportion of space in a green space system that has micro-topographical retention functions. The higher the value, the greater the proportion of green space units that can provide on-site retention and delay confluence. The specific calculation formula is as follows: ; in, To study the area of ​​sunken green spaces within the region, This represents the total area of ​​all green spaces within the study area.

[0063] (3) Green infrastructure connectivity GIC , The degree of green space patch connectivity, based on patch adjacency relationships (eight-neighborhood), represents the proportion of actual adjacent connection pairs to the theoretical maximum possible number of connection pairs. It is used to measure the spatial connectivity of green space patches. The higher the value, the more complete the physical adjacency relationship between green space patches and the stronger the continuity of potential ecological and hydrological processes. The specific calculation formula is as follows: ; in, This represents the number of adjacent patch pairs connected in the green space. Let be the theoretical maximum possible number of connections, where , This represents the total number of green space patches.

[0064] Furthermore, the gray infrastructure (GyI) indicator system is primarily used to characterize the flood control and drainage capabilities of gray infrastructure, reflecting its core role in resisting floods and draining accumulated water. Specifically, it includes three indicators: (1) Average dike crest elevation ADCE , The average elevation of the dike crest, weighted by the length of the dike segment, is used to characterize the overall elevation control level of flood control projects along the coast of the region. A higher value generally means a stronger ability to resist the risk of backwater or overtopping. The specific calculation formula is as follows: ; in, For the first The top elevation of the embankment section, For the first Length of the dike section This represents the total number of sections of the dike.

[0065] (2) Municipal drainage network density DSD , The value represents the total length of the centerline of the municipal drainage network per unit area. It is used to characterize the coverage of the drainage network and the space supply for drainage channels. The higher the value, the denser the drainage facilities are. The specific calculation formula is as follows: ; in, The total length of the centerline of all municipal drainage pipe networks. This represents the total area of ​​the study region.

[0066] (3) Average drain pipe diameter ADPD , The average pipe diameter, weighted by pipe segment length, is used to characterize the overall scale and potential water conveyance capacity of the drainage network. A larger value generally indicates a larger main pipe diameter and a higher theoretical upper limit of water conveyance capacity. The specific calculation formula is as follows: ; in, For the first The diameter of the drainage pipe section For the first The length of the drainage pipe section This represents the total number of drainage pipe sections.

[0067] Furthermore, the blue-green-grey infrastructure coupling degree (CD) index system is primarily used to characterize the degree of synergy among the three types of infrastructure, reflecting their joint ability to resist floods. Specifically, it includes three indicators: (1) Ecological shoreline ratio ESR , The value represents the proportion of shoreline length for ecological revetments or buffer zones (such as dike setbacks with reserved waterfront ecological zones) to the total shoreline length. It is used to characterize the degree of synergy between gray protection projects and blue-green spaces. The higher the value, the greater the proportion of shoreline integrating "engineering protection and ecological buffering". The specific calculation formula is as follows: ; in, The length of the embankment using ecological revetment. The total length of the waterfront embankment within the study area.

[0068] (2) Outlet and weir density OWD , The total number density of "pipeline network - river drainage outlet" and "pipeline network - LID unit overflow weirs such as sunken green space" per unit area is used to characterize the hydraulic connection strength between the gray drainage system and the receiving water body / blue-green storage unit. The higher the value, the denser the cross-system exchange interface and the more potential diversion and scheduling paths. The specific calculation formula is as follows: ; in, and These refer to the number of drainage outlets from the municipal pipe network to the river and the number of overflow weirs from the municipal pipe network to sunken green spaces (rain gardens, etc.). This represents the total area of ​​the study region.

[0069] (3) Average surface perviousness ASP , Based on land use type permeability parameters and area-weighted regional average permeability level, this value is used to comprehensively characterize the overall control capacity of the underlying surface over the rainfall infiltration-runoff process. A higher value indicates a lower overall impermeability and better potential in-situ infiltration and peak shaving conditions. The specific calculation formula is as follows: ; in, For the first The permeability of different land use types was set as follows (based on the actual situation of the study area, the permeability parameters for different land use types are as follows: green space / water surface is 0.95, public facilities / commercial / public facility is 0.25, residential land is 0.45, road is 0.2, and industrial / warehouse / traffic facility is 0.35). For the first Area of ​​various land use types This represents the total area of ​​the study region.

[0070] Furthermore, the dynamic indicators of flooding include at least one of the following: water accumulation in built-up areas, water storage capacity of green infrastructure, water storage capacity of pipe networks, and inundated area where the product of water depth and flow velocity is ≥0.4 m² / s. The specific analysis steps, calculation formulas, and meanings of each indicator are as follows: (1) Water volume of constructed area (WVCA): The two-dimensional surface inundation water depth time series data within the constructed area during the simulation process were extracted from the MIKE FLOOD simulation results using ArcGIS software and defined as WVCA. This index is a negative index; the larger the value, the more severe the flooding in the constructed area. The calculation formula is as follows: ; in, For time points in the flood process, Let be the submerged water depth of the i-th grid cell at time t. The area of ​​each grid cell, This is a grid cell affiliation identifier used to filter grid cells within the built-up area.

[0071] (2) Storage volume of Green Infrastructure (SVBG): The storage volume of green infrastructure (BGI) is determined by reading the MIKEFLOOD simulation results from ArcGIS. It is defined as the time series quantity of the water volume stored, which is represented by the surface inundation depth within the BGI spatial range, during the entire flood process. This index is a positive indicator; the larger the value, the stronger the storage contribution provided by BGI. The calculation formula is as follows: ; in, This is the grid cell affiliation identifier, used to filter grid cells within the BGI range.

[0072] (3) Peak storage volume of sewer system (SVSS): The peak value of the instantaneous water storage volume inside the drainage sewer system during the entire flood process is defined as the MIKE FLOOD simulation results obtained by ArcGIS. This index is a positive index. Under certain boundary conditions, the stronger the system's temporary storage capacity, the more beneficial it is. The calculation formula is as follows: ; in, Let be the total water volume of all pipe sections in the drainage network at time t. This represents the total water volume stored at all nodes of the drainage network at time t.

[0073] (4) The flooded area with an inundation depth × water velocity ≥ 0.4 m² / s (FADV) was determined using the MIKE FLOOD simulation results obtained from ArcGIS. Throughout the flooding process, the built-up area met the hazard threshold conditions. The area covered is defined as FADV, used to characterize the high risk of flooding dynamics that pedestrians and vehicles may face. This indicator is negative; the larger the area, the higher the risk exposure. The calculation formula is as follows: ; in, Let be the submerged water depth of the i-th grid cell at time t. Let be the water flow velocity in the i-th grid cell at time t. This is a probability indicator function.

[0074] In one optional implementation, the standardization of the independent and dependent variables, and the data augmentation using a phased resampling strategy to obtain an augmented sample dataset, include: The independent variables are standardized using Z-score. The dependent variable is subjected to Yeo-Johnson transformation to make it approximately normally distributed, and the transformed value is scaled to a preset range using Min-Max. The sample dataset was divided into three initial single-stage sample datasets according to the stages of flooding. Calculate the Pearson correlation coefficient between the independent and dependent variables in each initial single-stage sample dataset, and classify the independent variables into weakly correlated features and strongly correlated features based on the correlation strength. Each initial single-stage sample dataset is resampled with replacement, and adaptive Gaussian noise is injected according to the strength of the features during the resampling process to obtain an enhanced single-stage sample dataset. The three enhanced single-stage sample datasets are then merged to obtain the enhanced sample dataset.

[0075] Specifically, the independent variable is standardized using Z-score to eliminate dimensions. The specific formula is as follows: ; in, It is the standardized value. It is the original independent variable The Each sample value It is the independent variable The mean, It is the independent variable The standard deviation.

[0076] Due to the randomness and complexity of flood disasters, the original dependent variable data often exhibits a significant skewed distribution. Bayesian mixed-effects models require a certain degree of normality in the input data; skewed data can lead to biased parameter estimation and numerical instability, affecting the model's evaluation accuracy. Therefore, the Yeo-Johnson transform is needed to correct the skewed distribution of the dependent variable, making it approximately conform to a normal distribution. The specific formula for the Yeo-Johnson transform is as follows: ; in, It is the value after the Yeo-Johnson transform. These are the original dependent variable values. To change parameters and control the degree of stretching or compression.

[0077] The values ​​after the Yeo-Johnson transform are then scaled to a preset range using Min-Max to improve the numerical stability of Bayesian sampling. The specific formula is as follows: ; in, It is a value that has been normalized and stretched to [0, 20]. and These are the minimum and maximum values ​​after Yeo-Johnson correction, respectively.

[0078] Subsequently, the standardized sample dataset was divided according to the flood stage, resulting in three initial single-stage sample datasets, corresponding to the initial period, the flood peak period, and the receding period, respectively. The correlation between the independent variables and the dependent variable differed in different flood stages. For example, during the flood peak period, the average levee crest elevation, the density of municipal drainage pipe network, and the water accumulation in the built-up area were strongly correlated, while during the initial period, the green space ratio, the proportion of sunken green space, and the water accumulation in the built-up area were strongly correlated. If the same enhancement strategy is applied to all independent variables, it will lead to excessive noise interference in the strongly correlated features or insufficient enhancement effect in the weakly correlated features. Therefore, the Pearson correlation coefficient between the independent variables and the dependent variable in each initial single-stage sample dataset was calculated, and the independent variables were classified into weakly correlated features and strongly correlated features according to the correlation strength.

[0079] Combining phased features and correlation features, a phased Bootstrap resampling strategy is adopted. During the resampling process, adaptive Gaussian noise is injected according to the feature strength, as shown in the following formula: ; in, This represents the value of the j-th independent variable corresponding to the i-th dependent variable in the resampled sample after standardization. The value after injecting noise. This is the injected Gaussian noise term.

[0080] The noise term follows a normal distribution with a mean of 0, as follows: ; in, Let $j$ be the standard deviation of the sample in this resampling batch, used to adaptively adjust the noise intensity according to the dispersion of the samples in this batch. The noise ratio is as follows: ; in, The proportion of noise with weak correlation characteristics. The proportion of noise representing a strong correlation feature. For a set of weakly correlated variables, For a set of strongly correlated variables, the definitions of the two are as follows: ; ; in, For the j-th independent variable The Pearson correlation coefficient with the dependent variable y, In this embodiment, the correlation strength threshold is used. .

[0081] Specifically, when the absolute value of the Pearson correlation coefficient between the j-th independent variable and the dependent variable is ≤0.2, the independent variable exhibits a weak correlation and is assigned to the weakly correlated variable set W. When the absolute value of the Pearson correlation coefficient between the j-th independent variable and the dependent variable is >0.2, the independent variable exhibits a strong correlation and is assigned to the strongly correlated variable set S. S={j}\W indicates that the strongly correlated variable set is the remaining part after removing the weakly correlated variable set from all independent variable sets, ensuring that all independent variables can be assigned to their corresponding sets without omission.

[0082] This embodiment provides high-quality, highly stable sample data support for the subsequent construction of Bayesian mixture effects models and parameter inference by standardizing the samples and performing phased data augmentation.

[0083] In one optional implementation, the prior distribution for setting the parameters of the Bayesian mixture effect model includes: A Laplace prior distribution is applied to the fixed effects coefficients of the Bayesian mixed effects model, a normal prior is set for the intercept, and a semi-normal prior is set for the standard deviation of the random effects. The fixed effects are the deterministic contributions of the spatial attributes of blue-green-gray infrastructure to flood resilience, while the stochastic effects are the environmental uncertainties caused by differences in hydrological and climatic scenarios.

[0084] Specifically, applying a Laplace (double exponential) prior to the fixed effects coefficients produces an effect similar to L1 regularization, promoting sparsity of insignificant parameters to prevent overfitting. The formula for calculating the fixed effects prior is as follows: ; in, Let be the basic fixed effects coefficient of the j-th independent variable (j=1,…,12, i.e., 12 spatial attribute indicators of blue, green, and gray infrastructure) in the k-th infrastructure group (e.g., blue infrastructure group, green infrastructure group, gray infrastructure group, coupling layer group), used to characterize the basic deterministic impact of this independent variable on flood resilience. Let be the stage-fixed effect coefficient of the j-th independent variable in the k-th infrastructure group at the p-th flood stage (p∈{init,peak,rec}, i.e., the initial period, the peak period, and the receding period). This coefficient is used to characterize the difference in the deterministic impact of this independent variable on flood resilience at different flood stages. For position parameters, This is the scale parameter.

[0085] Secondly, a normal prior is set for the intercept, and a semi-normal prior is set for the standard deviation of the random effects. The formula for calculating the intercept prior is as follows: ; in, Let be the intercept parameter corresponding to the k-th group of infrastructure. Let be the mean of the normal prior distribution, and take a value of 0. The variance is 0.3.

[0086] The formula for calculating the prior standard deviation of random effects is as follows: ; in, Let be the standard deviation of the random effects of the k-th infrastructure group under a certain hydro-climate scenario, and be the scale parameter of the semi-normal prior distribution, with a value of 0.002.

[0087] This embodiment applies a Laplace prior distribution to the fixed effects coefficients, effectively solving the overfitting problem during model training and ensuring that the model can accurately capture the deterministic contribution of the spatial attributes of blue, green, and gray infrastructure to flood resilience. A normal prior distribution is set for the intercept parameter to ensure accurate estimation of the baseline level of flood resilience. A semi-normal prior distribution is set for the standard deviation of random effects to accurately quantify the environmental uncertainty interference caused by differences in hydrological and climatic scenarios, ensuring that the model can clearly distinguish between the deterministic contribution of infrastructure and the random interference of the environment, thereby improving the comprehensiveness and rationality of the model assessment.

[0088] In one optional implementation, the posterior distribution of parameters is obtained using a sampling algorithm based on the enhanced sample dataset. Based on the posterior distribution, a phased dynamic assessment result of the flood resilience of blue-green-grey infrastructure is output, including: Based on the enhanced sample dataset, the Markov chain Monte Carlo sampling algorithm is used to sample the model parameters, and the posterior distribution of all parameters is obtained by fitting the sampling results. Extract the posterior mean and the 95% highest density interval of the parameters from the posterior distribution; Based on the 95% highest density range, the significance of the independent variables and independent variable groups on flood resilience was determined. Based on the posterior mean of the parameters, calculate the stage contribution of each independent variable and each group of independent variables to flood resilience at different flood stages; Based on the stage contribution value and significance determination results, output the phased dynamic assessment results of the flood resilience of blue, green and gray infrastructure.

[0089] Specifically, the NUTS (No-U-Turn Sampler) algorithm is used for MCMC (Markov Chain Monte Carlo) sampling, sampling is performed for both individual fixed-effects variables and group fixed-effects variables. The sampling formula for an individual fixed-effects variable is as follows: ; in, Indicates the enhanced sample dataset Under the given conditions, the posterior probability distribution of the model parameters.

[0090] The core function of this formula is to calculate the posterior distribution of all model parameters corresponding to a single fixed-effect independent variable based on sample data and prior distribution, thus providing a basis for subsequent impact analysis of individual indicators.

[0091] The sampling formula for the group fixed-effect independent variable is as follows: ; in, The basic fixed effects coefficient is the fixed effects coefficient of the g-th group of infrastructure in the k-th group (i.e., the blue-green-grey infrastructure grouping), which is used to characterize the basic deterministic contribution of this group to flood resilience.

[0092] The core function of this formula is to calculate the posterior distribution of all model parameters corresponding to the group fixed effect independent variables based on sample data and prior distribution, providing a basis for subsequent impact analysis of infrastructure grouping.

[0093] Secondly, the posterior mean and the 95% highest density interval (HDI) are extracted from the posterior distribution. The posterior mean is the optimal estimate of the model parameters. It is extracted for individual fixed-effects independent variables and group fixed-effects independent variables, respectively. The specific formulas are as follows: ; ; in, Let j be the posterior mean of the single fixed-effect independent variable j. Let g be the posterior mean of the group fixed-effect independent variable g.

[0094] The 95% HDI is the narrowest interval in the posterior distribution with the highest probability density and containing 95% of the samples. Its core function is to characterize the uncertainty of parameter estimation. It is also used to determine the significance of the impact of various indicators and groups of blue, green and gray infrastructure on flood resilience. The specific judgment rule is as follows: if the 95% HDI does not cross zero (i.e., all values ​​in the interval are greater than 0 or all values ​​are less than 0), then the spatial attribute (single indicator or group) is determined to have a significant driving effect on flood resilience in this flood stage. If the 95% HDI crosses zero (i.e., the interval contains both positive and negative numbers), then the spatial attribute is determined to have no significant impact on flood resilience in this flood stage.

[0095] Next, based on the posterior mean of the parameters, the phase contribution (PhC) of each independent variable and its group at different flood stages is calculated. The phase contribution is used to quantify the specific impact of a single blue-green-grey infrastructure indicator or group on flood resilience at different flood stages (initial stage, peak stage, and receding stage). Its core is to combine the posterior mean of the basic fixed effects and the posterior mean of the phase fixed effects to comprehensively reflect the comprehensive contribution of the indicator or group at a specific stage. This can clearly show the differences in the role of infrastructure at different stages and provide core data support for phased dynamic assessment.

[0096] Finally, based on the stage contribution values ​​and significance determination results, the phased dynamic assessment results of the flood resilience of blue, green and gray infrastructure are output. The specific output includes: significant impact indicators and groupings for each flood stage (initial stage, peak stage, and receding stage), stage contribution values ​​(quantified values) of each significant indicator and group, influence direction (positive / negative), and influence intensity (absolute value of contribution). At the same time, combined with the characteristics of flood dynamic indicators (water accumulation in built-up areas, water storage of green infrastructure, etc.), the core role of infrastructure in different stages is analyzed to form a complete phased dynamic assessment report.

[0097] This embodiment, through the above-mentioned technical solution of posterior distribution acquisition and phased dynamic evaluation, can accurately identify the core infrastructure and key impact indicators at different flood stages, clarify the optimization direction of various types of infrastructure, and realize the refined and dynamic evaluation of the flood resilience of blue, green and gray infrastructure.

[0098] In one alternative implementation, the stage contribution value is the sum of the average base contribution over the entire flood cycle and the additional differentiated contribution in the corresponding flood stage.

[0099] Specifically, based on the significance and sign of the posterior distribution, the 12 fixed-effect independent variables are determined. (Single blue-green-gray infrastructure indicator) and 4 grouped fixed-effects independent variables The stage contribution of (BI, GI, GyI, CD) to flood resilience is specifically divided into the stage contribution values ​​of individual fixed-effects independent variables and group fixed-effects independent variables. The formula for calculating the stage contribution value of individual fixed-effects independent variables is as follows: ; in, For the j-th single fixed-effect independent variable in the k-th group The stage contribution value of the p-th flood stage to the k-th flood resilience index.

[0100] The formula for calculating the stage contribution of the group fixed effects independent variable is as follows: ; Wherein, is the fixed effect independent variable of the g-th group in the k-th group. (Blue-green-grey infrastructure grouping) The stage contribution of the p-th flood stage to the k-th flood resilience index.

[0101] It should be noted that the calculation of the stage contribution value needs to be combined with the significance judgment result. If the 95% HDI of a certain independent variable or group does not cross 0, it is determined that the characteristic has a significant effect on the flood resilience index in that stage, and its stage contribution value has practical significance and can be used for subsequent evaluation and analysis. If the 95% HDI crosses 0, it means that the effect is not significant, and its stage contribution value is only used as a reference and can be removed in subsequent evaluations to avoid interfering with the accuracy of the evaluation results.

[0102] This embodiment accurately quantifies the specific impact of individual infrastructure indicators and groups on flood resilience at different flood stages by calculating the stage contribution value, clearly demonstrating the differences in the role of infrastructure at different stages, and breaking through the limitation of traditional assessment methods that cannot achieve phased dynamic assessment.

[0103] Secondly, embodiments of the present invention provide an infrastructure flood resilience assessment device, see [link to relevant documentation]. Figure 2 This is a schematic diagram of one embodiment of an infrastructure flood resilience assessment device provided by the present invention.

[0104] like Figure 2 As shown, the device includes: The multi-scenario flood simulation module 21 is used to simulate the entire process of urban flooding in the target area under multiple scenarios using a flood simulation coupling model. The feature extraction module 22 is used to extract spatial attribute indicators of blue-green-gray infrastructure as independent variables from the simulation process, extract flood dynamic indicators of different flood stages as dependent variables, form a sample dataset, and divide the independent variables into multiple independent variable groups according to function; wherein, the flood stages include the initial period, the flood peak period, and the receding period; The data processing and enhancement module 23 is used to standardize the independent variable and the dependent variable respectively, and to perform data enhancement using a phased resampling strategy to obtain an enhanced sample dataset. The model building and evaluation output module 24 is used to build a Bayesian mixture effect model, set a prior distribution for the parameters of the Bayesian mixture effect model, obtain the posterior distribution of the parameters based on the enhanced sample dataset using a sampling algorithm, and output the phased dynamic evaluation results of the flood resilience of blue-green-grey infrastructure according to the posterior distribution.

[0105] In one optional implementation, the step of using a flood simulation coupling model to simulate the entire process of urban flooding in the target area under multiple scenarios includes: Obtain historical hydrological and meteorological data of flood events in the target area, as well as predicted vertical data of sea level rise, and construct various hydrological and climatic flood scenarios based on the above data; For each flood scenario, corresponding boundary conditions are set, including water level, rainfall intensity, typhoon intensity, sea level rise prediction data, and land subsidence prediction data. The MIKE FLOOD coupled model was used to simulate the entire process of urban flooding; the coupled model includes the MIKE 11 one-dimensional river hydrodynamic module, the MIKE 21 two-dimensional surface runoff module, and the MIKE URBAN municipal pipe network drainage module.

[0106] In one optional implementation, the spatial attribute indicators of the blue-green-gray infrastructure include at least one of the following: water surface ratio, river network density, and water connectivity of blue infrastructure; green space ratio, sunken green space ratio, and green space connectivity of green infrastructure; average embankment crest elevation, municipal drainage network density, and average drainage pipe diameter of gray infrastructure; ecological embankment ratio, drainage outlet and overflow weir density, and average surface permeability of the degree of facility coupling; and the flood dynamic indicators include at least one of the following: water accumulation in built-up areas, water storage capacity of green infrastructure, water storage capacity of pipe networks, and inundated area where the product of water depth and water flow velocity is ≥0.4 m² / s.

[0107] In one optional implementation, the standardization of the independent and dependent variables, and the data augmentation using a phased resampling strategy to obtain an augmented sample dataset, include: The independent variables are standardized using Z-score. The dependent variable is subjected to Yeo-Johnson transformation to make it approximately normally distributed, and the transformed value is scaled to a preset range using Min-Max. The sample dataset was divided into three initial single-stage sample datasets according to the stages of flooding. Calculate the Pearson correlation coefficient between the independent and dependent variables in each initial single-stage sample dataset, and classify the independent variables into weakly correlated features and strongly correlated features based on the correlation strength. Each initial single-stage sample dataset is resampled with replacement, and adaptive Gaussian noise is injected according to the strength of the features during the resampling process to obtain an enhanced single-stage sample dataset. The three enhanced single-stage sample datasets are then merged to obtain the enhanced sample dataset.

[0108] In one optional implementation, the prior distribution for setting the parameters of the Bayesian mixture effect model includes: A Laplace prior distribution is applied to the fixed effects coefficients of the Bayesian mixed effects model, a normal prior is set for the intercept, and a semi-normal prior is set for the standard deviation of the random effects. The fixed effects are the deterministic contributions of the spatial attributes of blue-green-gray infrastructure to flood resilience, while the stochastic effects are the environmental uncertainties caused by differences in hydrological and climatic scenarios.

[0109] In one optional implementation, the posterior distribution of parameters is obtained using a sampling algorithm based on the enhanced sample dataset. Based on the posterior distribution, a phased dynamic assessment result of the flood resilience of blue-green-grey infrastructure is output, including: Based on the enhanced sample dataset, the Markov chain Monte Carlo sampling algorithm is used to sample the model parameters, and the posterior distribution of all parameters is obtained by fitting the sampling results. Extract the posterior mean and the 95% highest density interval of the parameters from the posterior distribution; Based on the 95% highest density range, the significance of the independent variables and independent variable groups on flood resilience was determined. Based on the posterior mean of the parameters, calculate the stage contribution of each independent variable and each group of independent variables to flood resilience at different flood stages; Based on the stage contribution value and significance determination results, output the phased dynamic assessment results of the flood resilience of blue, green and gray infrastructure.

[0110] In one alternative implementation, the stage contribution value is the sum of the average base contribution over the entire flood cycle and the additional differentiated contribution in the corresponding flood stage.

[0111] It should be noted that the infrastructure flood resilience assessment device provided in this embodiment of the invention is used to execute all the process steps of the infrastructure flood resilience assessment method in the above embodiment. The working principles and beneficial effects of the two are one-to-one, so they will not be described again.

[0112] Thirdly, embodiments of the present invention provide an electronic device, see [link to previous document]. Figure 3 The diagram shown is a structural schematic of an electronic device provided in an embodiment of the present invention.

[0113] like Figure 3 As shown, the device includes: Memory 31 is used to store computer programs; Processor 32 is used to execute the computer program; When the processor 32 executes the computer program, it implements the infrastructure flood resilience assessment method as described in any of the above embodiments.

[0114] For example, the computer program may be divided into one or more modules / units, which are stored in the memory 31 and executed by the processor 32 to complete the present invention. The one or more modules / units may be a series of computer program instruction segments capable of performing a specific function, which describe the execution process of the computer program in the electronic device.

[0115] The processor 32 can be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor can be a microprocessor or any conventional processor.

[0116] The memory 31 can be used to store the computer programs and / or modules. The processor 32 implements various functions of the electronic device by running or executing the computer programs and / or modules stored in the memory 31 and calling the data stored in the memory 31. The memory 31 may mainly include a program storage area and a data storage area. The program storage area may store the operating system, at least one application program required for a function (such as sound playback function, image playback function, etc.), etc.; the data storage area may store data created according to the use of the mobile phone (such as audio data, phonebook, etc.). In addition, the memory 31 may include high-speed random access memory, and may also include non-volatile memory, such as hard disk, memory, plug-in hard disk, smart media card (SMC), secure digital card (SD) card, flash card, at least one disk storage device, flash memory device, or other volatile solid-state storage device.

[0117] It should be noted that the aforementioned electronic devices include, but are not limited to, processors and memory, as will be understood by those skilled in the art. Figure 3 The structural diagram is merely an example of the electronic device described above and does not constitute a limitation on the electronic device. It may include more components than shown in the diagram, or combine certain components, or use different components.

[0118] Fourthly, embodiments of the present invention also provide a computer-readable storage medium storing a computer program that, when executed, implements the infrastructure flood resilience assessment method described in any of the above embodiments.

[0119] It should be understood that the present invention can implement all or part of the processes in the above-described infrastructure flood resilience assessment method, or it can be accomplished by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the above-described infrastructure flood resilience assessment method. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms. The computer-readable medium can include: any entity or device capable of carrying the computer program code, a recording medium, a USB flash drive, a portable hard drive, a magnetic disk, an optical disk, a computer memory, a read-only memory (ROM), a random access memory (RAM), an electrical carrier signal, a telecommunication signal, and a software distribution medium, etc.

[0120] The above description is merely a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. It should be noted that, for those skilled in the art, several equivalent obvious modifications and / or equivalent substitutions can be made without departing from the technical principles of the present invention, and these obvious modifications and / or equivalent substitutions should also be considered within the scope of protection of the present invention.

Claims

1. A method for assessing the flood resilience of infrastructure, characterized in that, include: A flood simulation coupled model was used to simulate the entire process of urban flooding in the target area under multiple scenarios. Spatial attribute indicators of blue, green, and gray infrastructure were extracted as independent variables during the simulation process, and dynamic flood indicators of different flood stages were extracted as dependent variables to form a sample dataset. The independent variables were then divided into multiple groups according to their functions. The flood stages included the initial stage, the peak flood stage, and the receding flood stage. The independent and dependent variables are standardized respectively, and a phased resampling strategy is used for data augmentation to obtain an augmented sample dataset. A Bayesian mixture effect model is constructed, and a prior distribution is set for the parameters of the Bayesian mixture effect model. Based on the enhanced sample dataset, a sampling algorithm is used to obtain the posterior distribution of the parameters. According to the posterior distribution, the phased dynamic evaluation results of the flood resilience of blue-green-grey infrastructure are output.

2. The infrastructure flood resilience assessment method as described in claim 1, characterized in that, The method of using a flood simulation coupling model to simulate the entire process of urban flooding in the target area under multiple scenarios includes: Obtain historical hydrological and meteorological data of flood events in the target area, as well as predicted vertical data of sea level rise, and construct various hydrological and climatic flood scenarios based on the above data; For each flood scenario, corresponding boundary conditions are set, including water level, rainfall intensity, typhoon intensity, sea level rise prediction data, and land subsidence prediction data. The MIKE FLOOD coupled model was used to simulate the entire process of urban flooding; the coupled model includes the MIKE 11 one-dimensional river hydrodynamic module, the MIKE 21 two-dimensional surface runoff module, and the MIKE URBAN municipal pipe network drainage module.

3. The infrastructure flood resilience assessment method as described in claim 1, characterized in that, The spatial attribute indicators of the blue-green-gray infrastructure include at least one of the following: water surface ratio, river network density, and water connectivity of blue infrastructure; green space ratio, sunken green space ratio, and green space connectivity of green infrastructure; average embankment crest elevation, municipal drainage network density, and average drainage pipe diameter of gray infrastructure; ecological embankment ratio, drainage outlet and overflow weir density, and average surface permeability of the degree of facility coupling. The flood dynamic indicators include at least one of the following: water accumulation in built-up areas, water storage capacity of green infrastructure, water storage capacity of pipe networks, and inundated area where the product of water depth and flow velocity is ≥0.4 m² / s.

4. The infrastructure flood resilience assessment method as described in claim 1, characterized in that, The process involves standardizing the independent and dependent variables, and then performing data augmentation using a phased resampling strategy to obtain an augmented sample dataset, including: The independent variables are standardized using Z-score. The dependent variable is subjected to Yeo-Johnson transformation to make it approximately normally distributed, and the transformed value is scaled to a preset range using Min-Max. The sample dataset was divided into three initial single-stage sample datasets according to the stages of flooding. Calculate the Pearson correlation coefficient between the independent and dependent variables in each initial single-stage sample dataset, and classify the independent variables into weakly correlated features and strongly correlated features based on the correlation strength. Each initial single-stage sample dataset is resampled with replacement, and adaptive Gaussian noise is injected according to the strength of the features during the resampling process to obtain an enhanced single-stage sample dataset. The three enhanced single-stage sample datasets are then merged to obtain the enhanced sample dataset.

5. The infrastructure flood resilience assessment method as described in claim 1, characterized in that, The prior distribution for setting the parameters of the Bayesian mixture effect model includes: A Laplace prior distribution is applied to the fixed effects coefficients of the Bayesian mixed effects model, a normal prior is set for the intercept, and a semi-normal prior is set for the standard deviation of the random effects. The fixed effects are the deterministic contributions of the spatial attributes of blue-green-gray infrastructure to flood resilience, while the stochastic effects are the environmental uncertainties caused by differences in hydrological and climatic scenarios.

6. The infrastructure flood resilience assessment method as described in claim 1, characterized in that, The enhanced sample dataset is used to obtain the posterior distribution of parameters using a sampling algorithm. Based on the posterior distribution, a phased dynamic assessment result of the flood resilience of blue-green-grey infrastructure is output, including: Based on the enhanced sample dataset, the Markov chain Monte Carlo sampling algorithm is used to sample the model parameters, and the posterior distribution of all parameters is obtained by fitting the sampling results. Extract the posterior mean and the 95% highest density interval of the parameters from the posterior distribution; Based on the 95% highest density range, the significance of the independent variables and independent variable groups on flood resilience was determined. Based on the posterior mean of the parameters, calculate the stage contribution of each independent variable and each group of independent variables to flood resilience at different flood stages; Based on the stage contribution value and significance determination results, the phased dynamic assessment results of the flood resilience of blue, green and gray infrastructure are output.

7. The infrastructure flood resilience assessment method as described in claim 6, characterized in that, The stage contribution value is the sum of the average base contribution over the entire flood cycle and the additional differentiated contribution in the corresponding flood stage.

8. An infrastructure flood resilience assessment device, characterized in that, include: The multi-scenario flood simulation module is used to simulate the entire process of urban flooding in a target area under multiple scenarios using a flood simulation coupling model. The feature extraction module is used to extract spatial attribute indicators of blue-green-gray infrastructure as independent variables and extract flood dynamic indicators of different flood stages as dependent variables from the simulation process to form a sample dataset. The independent variables are divided into multiple independent variable groups according to their functions. The flood stages include the initial stage, the flood peak stage, and the receding stage. The data processing and enhancement module is used to standardize the independent variable and the dependent variable respectively, and to perform data enhancement using a phased resampling strategy to obtain the enhanced sample dataset. The model building and evaluation output module is used to build a Bayesian mixture effect model, set a prior distribution for the parameters of the Bayesian mixture effect model, obtain the posterior distribution of the parameters based on the enhanced sample dataset using a sampling algorithm, and output the phased dynamic evaluation results of the flood resilience of blue-green-grey infrastructure according to the posterior distribution.

9. An electronic device, characterized in that, include: Memory, used to store computer programs; A processor for executing the computer program; The processor executes the computer program to implement the infrastructure flood resilience assessment method as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed, implements the infrastructure flood resilience assessment method as described in any one of claims 1 to 7.

Images