Urban waterlogging prediction method and device, computer device, and storage medium

Abstract

The present application relates to the technical field of waterlogging prediction, and particularly relates to a city waterlogging prediction method and device, computer equipment and a storage medium. Hydrological connectivity correlation data and a functional vulnerability index parameter set are obtained. A hydrological connectivity index is calculated according to the hydrological connectivity correlation data, and a hydrological connectivity index is obtained. A comprehensive vulnerability index is calculated according to the functional vulnerability index parameter set, and a comprehensive vulnerability index is obtained. A coupling factor is constructed according to the hydrological connectivity index and the comprehensive vulnerability index, and a physical dynamic coupling factor is obtained. The hydrological connectivity correlation data, the hydrological connectivity index, the comprehensive vulnerability index and the physical dynamic coupling factor are input into a preset waterlogging prediction model for waterlogging prediction, and waterlogging prediction data is obtained. A waterlogging risk area is identified according to the waterlogging prediction data of each grid, and a waterlogging risk area is obtained.

Description

Technical Field

[0001] This invention relates to the field of urban flooding prediction technology, and in particular to a method, apparatus, computer equipment, and storage medium for predicting urban flooding. Background Technology

[0002] Urban flooding refers to water accumulation caused by the overload of urban drainage systems under extreme precipitation events. It is characterized by its suddenness, spatial heterogeneity, and widespread nature. Against the backdrop of intensifying climate change, how to conduct efficient, accurate, and low-cost urban flooding risk assessments to formulate effective and spatially differentiated flooding mitigation strategies has become an important direction and research challenge in current urban flooding management. Summary of the Invention

[0003] Based on this, the purpose of this invention is to provide an urban flooding prediction method, device, computer equipment, and storage medium. It constructs a hydrological connectivity index and a comprehensive vulnerability index by combining hydrological connectivity correlation data and a functional vulnerability index parameter set. It identifies flooding risk areas based on physical constraint deep learning and multi-source coupling to generate a flooding prediction model, thereby obtaining flooding risk areas and providing scientific guidance for disaster prevention and future urban planning.

[0004] In a first aspect, embodiments of this application provide a method for predicting urban flooding, comprising the following steps:

[0005] Obtain hydrological connectivity correlation data and a set of functional vulnerability index parameters for several grids in the area to be predicted;

[0006] Hydrological connectivity data of each grid is used to perform upstream and downstream hydrological analysis to obtain the upstream contribution and downstream impedance terms of each grid. Hydrological connectivity index is calculated based on the upstream contribution and downstream impedance terms of each grid to obtain the hydrological connectivity index of each grid.

[0007] The comprehensive vulnerability index of each grid is calculated based on the set of functional vulnerability index parameters of each grid.

[0008] The coupling factors are constructed based on the hydrological connectivity index and the comprehensive vulnerability index of each grid to obtain the physical-dynamic coupling factor of each grid.

[0009] The hydrological connectivity correlation data, hydrological connectivity index, comprehensive vulnerability index and physical-dynamic coupling factor of each grid are input into the preset waterlogging prediction model to make waterlogging predictions and obtain waterlogging prediction data for each grid.

[0010] Based on the flooding prediction data, flooding probability prediction data, and flooding duration prediction data in the flooding prediction data of each grid, the flooding risk area is identified, and the flooding risk area of ​​the area to be predicted is obtained.

[0011] Secondly, embodiments of this application provide an urban flooding prediction device, comprising:

[0012] The data acquisition module is used to acquire hydrological connectivity correlation data and functional vulnerability index parameter set of several grids collected in the area to be predicted within a preset time period.

[0013] The first index calculation module is used to perform upstream and downstream hydrological analysis based on the hydrological connectivity correlation data of each grid, and obtain the upstream contribution and downstream impedance terms of each grid; and to calculate the hydrological connectivity index of each grid based on the upstream contribution and downstream impedance terms of each grid.

[0014] The second index calculation module is used to calculate the comprehensive vulnerability index based on the functional vulnerability index parameter set of each grid, and obtain the comprehensive vulnerability index of each grid.

[0015] The coupling factor construction module is used to construct coupling factors based on the hydrological connectivity index and the comprehensive vulnerability index of each grid, and obtain the physical-dynamic coupling factor of each grid.

[0016] The waterlogging prediction module is used to input the hydrological connectivity correlation data, hydrological connectivity index, comprehensive vulnerability index and physical-dynamic coupling factor of each grid into the preset waterlogging prediction model to make waterlogging predictions and obtain waterlogging prediction data for each grid.

[0017] The area identification module is used to identify waterlogging risk areas based on the flooding prediction data, waterlogging probability prediction data, and flooding duration prediction data in the waterlogging prediction data of each grid, and to obtain the waterlogging risk areas of the area to be predicted.

[0018] Thirdly, embodiments of this application provide a computer device, including: a processor, a memory, and a computer program stored in the memory and executable on the processor; when the computer program is executed by the processor, it implements the steps of the urban flooding prediction method as described in the first aspect.

[0019] Fourthly, embodiments of this application provide a storage medium storing a computer program that, when executed by a processor, implements the steps of the urban flooding prediction method as described in the first aspect.

[0020] In this application embodiment, an urban flooding prediction method, device, computer equipment, and storage medium are provided. The method constructs a hydrological connectivity index and a comprehensive vulnerability index by combining hydrological connectivity correlation data and a functional vulnerability index parameter set. The method identifies flooding risk areas based on physical constraint deep learning and multi-source coupling to obtain flooding risk areas, providing scientific guidance for disaster prevention and future urban planning.

[0021] To better understand and implement this invention, the following detailed description is provided in conjunction with the accompanying drawings. Attached Figure Description

[0022] Figure 1 A flowchart illustrating an urban flooding prediction method provided in one embodiment of this application;

[0023] Figure 2 This is a flowchart illustrating step S2 of an urban flooding prediction method provided in one embodiment of this application.

[0024] Figure 3 This is a flowchart illustrating step S23 of an urban flooding prediction method provided in one embodiment of this application.

[0025] Figure 4 This is a flowchart illustrating step S24 of an urban flooding prediction method provided in one embodiment of this application.

[0026] Figure 5 This is a flowchart illustrating step S3 of an urban flooding prediction method provided in one embodiment of this application.

[0027] Figure 6 This is a flowchart illustrating step S5 of an urban flooding prediction method provided in one embodiment of this application.

[0028] Figure 7 A flowchart illustrating step S7 of an urban flooding prediction method provided in another embodiment of this application;

[0029] Figure 8 This is a flowchart illustrating step S6 of an urban flooding prediction method provided in one embodiment of this application.

[0030] Figure 9 This is a schematic diagram of the structure of an urban flooding prediction device provided in one embodiment of this application;

[0031] Figure 10 This is a schematic diagram of the structure of a computer device provided in one embodiment of this application. Detailed Implementation

[0032] Exemplary embodiments will now be described in detail, examples of which are illustrated in the accompanying drawings. When the following description relates to the drawings, unless otherwise indicated, the same numbers in different drawings denote the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this application. Rather, they are merely examples of apparatuses and methods consistent with some aspects of this application as detailed in the appended claims.

[0033] The terminology used in this application is for the purpose of describing particular embodiments only and is not intended to be limiting of the application. The singular forms “a,” “the,” and “the” used in this application and the appended claims are also intended to include the plural forms unless the context clearly indicates otherwise. It should also be understood that the term “and / or” as used herein refers to and includes any or all possible combinations of one or more of the associated listed items.

[0034] It should be understood that although the terms first, second, third, etc., may be used in this application to describe various information, such information should not be limited to these terms. These terms are only used to distinguish information of the same type from one another. For example, without departing from the scope of this application, first information may also be referred to as second information, and similarly, second information may also be referred to as first information. Depending on the context, the word "if" as used herein may be interpreted as "when," "when," or "in response to determination."

[0035] Please see Figure 1 , Figure 1 The flowchart illustrates an urban flooding prediction method according to an embodiment of this application. The method includes the following steps:

[0036] S1: Obtain hydrological connectivity correlation data and functional vulnerability index parameter set for several grids in the area to be predicted.

[0037] The entity executing the urban flooding prediction method of this application is the prediction device (hereinafter referred to as the prediction device). In an optional embodiment, the prediction device may be a computer device, a server, or a server cluster composed of multiple computer devices.

[0038] In this embodiment, the prediction device obtains hydrological connectivity correlation data and a set of functional vulnerability index parameters for several grids collected in the area to be predicted at the current time. The hydrological connectivity correlation data includes surface layer data and underground pipe network layer data. The surface layer data is used to indicate the spatial movement of surface runoff in the grids; the underground pipe network layer data is used to indicate the spatial movement of water transported by underground pipelines from the grids to the target sink. Specifically, the surface layer data includes rainfall-related data, slope data, retention factor, area data, and elevation data; the rainfall-related data includes effective areal rainfall data and rainfall intensity data. The underground pipe network layer data includes target sink bottleneck capacity data and target sink water distribution weights.

[0039] The functional vulnerability index parameter set includes vulnerability index parameter data across several dimensions, including both positive and negative vulnerability index parameters. Specifically, the dimensions include physical exposure vulnerability, infrastructure vulnerability, economic vulnerability, population vulnerability, and environmental vulnerability. The physical exposure vulnerability index parameters include the proportion of impervious surface area, building height, building volume ratio, and building density; the infrastructure vulnerability index parameters include road density, distance from roads, drainage system density, number of hospitals, number of schools, and number of emergency service facilities; the economic vulnerability index parameters include commercial activity density and GDP; the population vulnerability index parameters include population density, proportion of elderly population, and proportion of child population; and the environmental vulnerability index parameters include the proportion of green space, proportion of water bodies, soil moisture, and distance from rivers. Positive vulnerability indicators include the proportion of impervious area, building volume ratio, building density, road density, number of hospitals, number of schools, number of emergency service facilities, commercial activity density, GDP, population density, proportion of elderly population, proportion of child population, soil moisture, and distance from rivers. Negative vulnerability indicators include building height, distance from roads, drainage system density, proportion of green space, and proportion of water bodies.

[0040] S2: Perform upstream and downstream hydrological analysis based on the hydrological connectivity correlation data of each grid to obtain the upstream contribution and downstream impedance terms of each grid; calculate the hydrological connectivity index based on the upstream contribution and downstream impedance terms of each grid to obtain the hydrological connectivity index of each grid.

[0041] In this embodiment, the prediction device performs upstream and downstream hydrological analysis based on the hydrological connectivity correlation data of each grid to obtain the upstream contribution term and downstream impedance term of each grid.

[0042] The prediction device calculates the hydrological connectivity index for each grid cell based on the upstream contribution and downstream impedance terms. The hydrological connectivity index is:

[0043]

[0044] In the formula, For the first x The hydrological connectivity index of each grid cell For the first x The upstream contribution of each grid cell, For the first x Downstream impedance term of each grid.

[0045] Please see Figure 2 , Figure 2 The flowchart of S2 in the urban flooding prediction method provided in one embodiment of this application is as follows: S21 to S24 are detailed below:

[0046] S21: Calculate the permeable and impermeable runoff based on the effective areal rainfall data of each grid, and obtain the permeable and impermeable runoff data of each grid; construct runoff weights based on the permeable and impermeable runoff data and rainfall intensity data of each grid, and obtain the runoff weights of each grid.

[0047] In this embodiment, the prediction device calculates the permeable and impermeable runoff based on the effective areal rainfall data of each grid, obtaining the permeable runoff data and impermeable runoff data for each grid. The permeable runoff data is as follows:

[0048]

[0049] In the formula, For the first x Data on the permeable portion of each grid cell. To find the maximum value function, For the first x Effective ground rainfall data for each grid cell. For the first x Infiltration rate of each grid cell.

[0050] The flow data for the impermeable portion are as follows:

[0051]

[0052] In the formula, For the first x Data on flow generation from the impermeable portion of each grid cell. The impermeable flow coefficient, For impermeable micro-depressions to store water consumption, For time step.

[0053] The prediction device constructs runoff weights for each grid cell based on the runoff data from the permeable portion, the runoff data from the impermeable portion, and the rainfall intensity data. The runoff weights are as follows:

[0054]

[0055] In the formula, For the first x The flow weight of each grid cell, For impermeability, I For rainfall intensity data, It is a small positive number, with a value range of 10. -4 -10 -3 Between these, avoid the influence of zero slope / dry day values.

[0056] S22: Determine the upstream and downstream grids of each grid based on the slope data of each grid, and construct several outflow edges and inflow edges of each grid; calculate the path cost of the outflow edges and inflow edges based on the slope data, retardation factor and flow generation weight of each grid, and obtain the path cost data of each outflow edge and each inflow edge of each grid.

[0057] In this embodiment, the prediction device determines the upstream and downstream grids of each grid based on the slope data of each grid, and constructs several outflow edges and inflow edges for each grid. The outflow edges are used to indicate the directed edges from the current grid to the downstream neighboring grid; the inflow edges are used to indicate the directed edges from the upstream neighboring grid to the current grid.

[0058] The prediction device calculates the path cost of the outflow and inflow edges based on the slope data, retardation factor, and flow generation weight of each grid cell, obtaining the path cost data of each outflow edge and each inflow edge of each grid cell. The path cost data is as follows:

[0059]

[0060] In the formula, For the first x Path cost data for each grid cell. The direction angle of the directed edge to which the grid points. For the first x Slope data for each grid cell. For the first x Each grid along the direction The retardation factor for unit compensation movement, wherein the retardation factor is composed of both vegetation and buildings, specifically obtained by multiplying the building retardation factor and the vegetation retardation factor.

[0061] S23: Calculate the upstream contribution of each grid cell based on the surface layer data of each grid cell and the path cost data of each inflow edge of each grid cell.

[0062] In this embodiment, the prediction device calculates the upstream contribution of each grid cell based on the surface layer data of each grid cell and the path cost data of each inflow edge of each grid cell.

[0063] Please see Figure 3 , Figure 3 The flowchart of step S23 in the urban flooding prediction method provided in one embodiment of this application is as follows: Steps S231 to S233 are detailed below:

[0064] S231: Calculate the multidirectional diversion coefficient of each inflow edge based on the slope data, elevation data, and path cost data of each inflow edge of each grid, and obtain the multidirectional diversion coefficient of each inflow edge of each grid.

[0065] In this embodiment, the prediction device calculates the multidirectional diversion coefficient of each inflow edge based on the slope data, elevation data, and path cost data of each inflow edge for each grid cell, thereby obtaining the multidirectional diversion coefficient of each inflow edge for each grid cell. The multidirectional diversion coefficient of each inflow edge is:

[0066]

[0067] In the formula, This is the inflow edge identifier, indicating the first... j The upstream neighbor grid points to the first x The inflow edge of each grid cell, For the first j The upstream neighbor grid points to the first x The multidirectional flow splitting coefficient of the inflow edge of each grid cell. For the first j The upstream neighbor grid points to the first x The unnormalized split weights of the inflow edges of each grid cell. For the first x The set of upstream neighboring grid cells of a given grid cell. For the first j The upstream neighbor grid points to the first x The direction angle of the inflow edge of each grid cell. For the first j The upstream neighbor grid points to the first xPath cost data for the inflow edges of each grid cell. For the first x Elevation data for each grid cell. For the first j Elevation data of upstream neighboring grids, The grid side length q The preset low-resistance preference index, r This is the preset slope weight index.

[0068] S232: Calculate the local upstream source term based on the slope data, area data, and runoff weight of each grid cell to obtain the local upstream source term of each grid cell.

[0069] In this embodiment, the prediction device calculates the local upstream source term based on the slope data, area data, and runoff weight of each grid cell to obtain the local upstream source term for each grid cell. The local upstream source term is:

[0070]

[0071] In the formula, For the first x Local upstream source terms for each grid cell, For the first x Area data for each grid cell.

[0072] S233: Obtain the upstream cumulative supply of each grid cell; calculate the upstream contribution term based on the local upstream source term, upstream cumulative supply of each grid cell, and the multidirectional splitting coefficient of each inflow edge of each grid cell, and obtain the upstream contribution term of each grid cell.

[0073] In this embodiment, the prediction device obtains the upstream cumulative supply of each grid cell; the prediction device calculates the upstream contribution term based on the local upstream source term, the upstream cumulative supply of each grid cell, and the multidirectional splitting coefficient of each inflow edge of each grid cell, thereby obtaining the upstream contribution term of each grid cell, wherein the upstream contribution term is:

[0074]

[0075] In the formula, For the first x The upstream contribution of each grid cell, For the first x The first grid j The cumulative upstream supply of each upstream neighbor grid. This is the impedance attenuation coefficient.

[0076] S24: Based on the underground pipe network layer data of each grid, determine the water conveyance path from each grid to each target sink; the water conveyance path includes several intermediate grids; calculate the downstream impedance term based on the underground pipe network layer data of each grid and the path cost data of each outflow edge of each intermediate grid, and obtain the downstream impedance term of each grid.

[0077] In this embodiment, the prediction device determines the water transport path from each grid to each target sink based on the underground pipe network layer data of each grid, wherein the water transport path includes several intermediate grids.

[0078] The prediction device calculates the downstream impedance term based on the underground pipe network layer data of each grid and the path cost data of each outflow edge of each intermediate grid, and obtains the downstream impedance term of each grid.

[0079] The underground pipe network layer data includes target sink bottleneck capacity data and target sink water distribution weights; please refer to... Figure 4 , Figure 4 The flowchart of S24 in the urban flooding prediction method provided in one embodiment of this application is shown below, including steps S241 to S242, as follows:

[0080] S241: Calculate the absorption efficiency of each target sink based on the bottleneck capacity data of each grid, the water distribution weight of each target sink, and the permeable runoff data of each grid to obtain the absorption efficiency of each target sink; calculate the destination weight of each target sink based on the absorption efficiency of each target sink to obtain the destination weight of each target sink.

[0081] In this embodiment, the prediction device calculates the absorption efficiency of the target sink based on the bottleneck capacity data of the target sink, the water distribution weight of the target sink, and the permeable flow data of each grid, and obtains the absorption efficiency of each target sink, as follows:

[0082] The prediction device obtains the total inflow load of each target sink based on the target sink water distribution weight, the permeable portion runoff data of each grid, and a preset total inflow load calculation algorithm. The total inflow load calculation algorithm is as follows:

[0083]

[0084] In the formula, For the first c Total inflow load at each target sink, For the first x The first grid c The target sink water allocation weight for each target sink point A grid summation operator.

[0085] The predictive device obtains the absorption efficiency of each target sink based on the bottleneck capacity of each target sink in the target sink bottleneck capacity data, the total inflow load of each target sink, and a preset absorption efficiency calculation algorithm. The absorption efficiency calculation algorithm is as follows:

[0086]

[0087] In the formula, For the first c The capacity ratio of each target sink point No. c The bottleneck capacity of a target sink represents the target sink's... c Along a predetermined water conveyance path, for the set { The bottleneck capability of a series system obtained by taking the minimum value. For the first v Water transport capacity of the pipeline section For the first c Absorption efficiency of each target sink. This is the sink sensitivity coefficient.

[0088] The prediction device calculates the destination weight of each target sink based on the absorption efficiency of each target sink and a preset destination weight calculation algorithm. The destination weight calculation algorithm is as follows:

[0089]

[0090] In the formula, For the first c The destination weight of each target sink point This represents the destination sensitivity coefficient.

[0091] S242: Calculate the minimum path cost based on the path cost data of each outflow edge of each intermediate grid in the water conveyance path of each grid to obtain the minimum path cost data of each grid; calculate the effective path cost based on the obtained minimum path cost data of each grid and the destination weight of the target sink to obtain the effective path cost data of each grid, and take the minimum value as the downstream impedance term to obtain the downstream impedance term of each grid.

[0092] In this embodiment, the prediction device calculates the minimum path cost based on the path cost data of each outflow edge of each intermediate grid in the water conveyance path of each grid, thereby obtaining the minimum path cost data of each grid; based on the obtained minimum path cost data of each grid and the destination weight of the target sink, it calculates the effective path cost data of each grid, and takes the minimum value as the downstream impedance term to obtain the downstream impedance term of each grid, wherein the downstream impedance term is:

[0093]

[0094] In the formula, For the first x The downstream impedance term of each grid, C The number of target sinks. For effective path cost data, This is the minimum path cost data. The outgoing edge identifier indicates the first edge. x The grid points to the first n The outflow edge of a downstream neighboring grid For the first x The grid points to the first n The direction angle of the outflow edge of each downstream neighboring grid. For the first x From one grid to the target sink c The data set involves the path cost of outflow edges. For the first x The grid points to the first n Path cost data for the outflow edges of each downstream neighbor grid. for The grid side length of the outflow edge.

[0095] S3: Calculate the comprehensive vulnerability index based on the functional vulnerability index parameter set of each grid to obtain the comprehensive vulnerability index of each grid.

[0096] In this embodiment, the prediction device calculates a comprehensive vulnerability index based on the functional vulnerability index parameter set of each grid, thereby obtaining the comprehensive vulnerability index of each grid.

[0097] Please see Figure 5 , Figure 5 The flowchart of S3 in the urban flooding prediction method provided in one embodiment of this application is as follows: S31 to S36 are detailed below.

[0098] S31: Standardize the vulnerability index parameters in each dimension of each grid to obtain the standardized positive and negative vulnerability index parameters in each dimension of each grid.

[0099] In this embodiment, the prediction device standardizes each vulnerability index parameter in the vulnerability index parameter data of each dimension of each grid to obtain the standardized positive vulnerability index parameters and negative vulnerability index parameters of each dimension of each grid.

[0100] S32: Using the information entropy calculation method, the information entropy weights are calculated based on the positive and negative vulnerability index parameters of each dimension of each grid after standardization, to obtain the information entropy weights of each positive and negative vulnerability index parameters of each grid.

[0101] In this embodiment, the prediction device employs an information entropy calculation method. Based on a preset information entropy weight calculation algorithm, it calculates information entropy weights for each positive and negative vulnerability index parameter across each dimension of each grid after standardization. This yields the information entropy weights for each positive and negative vulnerability index parameter across each dimension of each grid. The information entropy weight calculation algorithm is as follows:

[0102]

[0103] In the formula, For the first l The information entropy weights of each vulnerability indicator parameter r The number of vulnerability index parameters for each dimension. For the first l The information entropy value of each vulnerability indicator parameter. For the first x After the first grid normalization process, the first l The values ​​of the vulnerability index parameters, o This represents the number of grid cells.

[0104] S33: Based on the positive vulnerability index parameters, negative vulnerability index parameters, and corresponding information entropy weights of each dimension of each grid after standardization, perform weighted processing to obtain the weighted values ​​of the positive vulnerability index parameters and negative vulnerability index parameters of each grid in each dimension.

[0105] In this embodiment, the prediction device performs weighted processing based on the positive vulnerability index parameters and negative vulnerability index parameters of each dimension of each grid after standardization, as well as the corresponding information entropy weights, to obtain the weighted values ​​of the positive vulnerability index parameters and negative vulnerability index parameters of each grid in each dimension.

[0106] In an optional embodiment, the prediction device introduces scenario factors to adjust the weights. Specifically, the prediction device sets several rainstorm scenarios. For each rainstorm scenario, a set of weight adjustment coefficients is defined. These coefficients are obtained from historical data analysis, i.e., determined by analyzing the correlation between various indicators and urban flooding under different rainstorm scenarios. The prediction device performs weighted processing based on the weight adjustment coefficients corresponding to each rainstorm scenario, the standardized positive vulnerability index parameters and negative vulnerability index parameters of each grid in each dimension, and the corresponding information entropy weights, to obtain the weighted values ​​of the positive vulnerability index parameters and negative vulnerability index parameters of each grid in each dimension under each rainstorm scenario, as described below:

[0107]

[0108] In the formula, For the first g The first rainstorm scenario l The weighted values ​​of each vulnerability indicator parameter, For the first g The first rainstorm scenario l Adjustment coefficients for each vulnerability indicator parameter.

[0109] S34: Calculate the sub-indices for each dimension of each grid by using the positive vulnerability index parameters, negative vulnerability index parameters, and the weighted values ​​of the positive and negative vulnerability index parameters. Combine the sub-indices of each grid in the same dimension to construct a sub-indice matrix for each dimension. Standardize the sub-indices of each grid in the sub-indice matrix for each dimension to obtain the standardized sub-indices for each dimension of each grid.

[0110] In this embodiment, the prediction device calculates sub-indices for each dimension of each grid based on the positive vulnerability index parameters, negative vulnerability index parameters, and the weighted values ​​of the positive and negative vulnerability index parameters for each dimension of each grid, thereby obtaining the sub-indices for each dimension of each grid.

[0111] Specifically, the prediction device multiplies the standardized values ​​of the positive vulnerability index parameters and negative vulnerability index parameters of each dimension of each grid with the corresponding weighted values ​​to construct a weighted standardized matrix of the positive vulnerability index parameters and negative vulnerability index parameters of each grid in each dimension.

[0112] The prediction device calculates sub-indices for each grid cell based on a pre-defined sub-index calculation algorithm, using a weighted normalized matrix of positive and negative vulnerability index parameters for each dimension of the grid. The sub-index calculation algorithm is as follows:

[0113]

[0114] In the formula, For the first x The first grid P Sub-indices of each dimension, For the first x The negative relative proximity of each grid cell For the first x The relative proximity of each grid cell For the first x The first grid l A weighted standardized matrix of vulnerability index parameters. For the first x The first grid l The maximum value in the weighted standardized matrix of the vulnerability index parameters. For the first x The first grid l The minimum value in the weighted standardized matrix of the vulnerability index parameters.

[0115] S35: Using the information entropy calculation method, the information entropy weights of each dimension of each grid are calculated based on the sub-indices of each dimension after standardization, and the information entropy weights of each dimension of each grid are obtained; the weighted values ​​of the sub-indices of each dimension of each grid are obtained by weighting based on the sub-indices of each dimension of each grid after standardization and the corresponding information entropy weights.

[0116] In this embodiment, the prediction device employs an information entropy calculation method. It calculates information entropy weights based on the sub-indices of each dimension of each grid after standardization, obtaining the information entropy weights of the sub-indices of each dimension of each grid. Then, it performs weighted processing based on the sub-indices of each dimension of each grid after standardization and the corresponding information entropy weights, obtaining the weighted values ​​of the sub-indices of each dimension of each grid. Specific embodiments can be found in steps S32 and S33, which will not be repeated here.

[0117] S36: Calculate the sub-indices based on the weighted values ​​of the sub-indices of each dimension of each grid, and use the obtained sub-indices as the comprehensive vulnerability index to obtain the comprehensive vulnerability index of each grid.

[0118] In this embodiment, the prediction device calculates the sub-index based on the weighted value of the sub-index of each dimension of each grid, and uses the obtained sub-index as the comprehensive vulnerability index to obtain the comprehensive vulnerability index of each grid. For specific implementation, please refer to step S34, which will not be repeated here.

[0119] S4: Based on the hydrological connectivity index and the comprehensive vulnerability index of each grid, a coupling factor is constructed to obtain the physical-dynamic coupling factor of each grid.

[0120] In this embodiment, the prediction device constructs coupling factors based on the hydrological connectivity index and the comprehensive vulnerability index of each grid, thereby obtaining the physical-dynamic coupling factor of each grid. The physical-dynamic coupling factor is:

[0121]

[0122] In the formula, For the first t The physical dynamic coupling factor at a given moment. FVI To form a comprehensive vulnerability index, The dynamic rate of change of the hydrological connectivity index. These are the coupling weight coefficients. This is a correction function for rainfall intensity and time.

[0123] S5: Input the hydrological connectivity correlation data, hydrological connectivity index, comprehensive vulnerability index and physical-dynamic coupling factor of each grid into the preset waterlogging prediction model to make waterlogging predictions and obtain waterlogging prediction data for each grid.

[0124] In this embodiment, the prediction device inputs the hydrological connectivity correlation data, hydrological connectivity index, comprehensive vulnerability index and physical-dynamic coupling factor of each grid into a preset waterlogging prediction model to perform waterlogging prediction and obtain waterlogging prediction data for each grid. The waterlogging prediction model includes a forward propagation module, a feature processing module and a multi-task prediction module.

[0125] Please see Figure 6 , Figure 6 The flowchart of step S5 in the urban flooding prediction method provided in one embodiment of this application is as follows: Steps S51 to S55 are detailed below:

[0126] S51: Based on the hydrological connectivity correlation data, hydrological connectivity index, comprehensive vulnerability index and physical-dynamic coupling factor, static features and dynamic features are divided into static features and dynamic features.

[0127] In this embodiment, the prediction device divides static and dynamic features based on the hydrological connectivity correlation data, hydrological connectivity index, comprehensive vulnerability index, and physical-dynamic coupling factor, and constructs static and dynamic features.

[0128] S52: Input the static features and dynamic features into the forward propagation module for standardization processing to obtain the standardized static features and dynamic features.

[0129] In this embodiment, the prediction device inputs the static features and dynamic features into the forward propagation module for standardization processing to obtain standardized static features and dynamic features.

[0130] Specifically, for the static features, the prediction device performs standardization processing on the static features according to a preset static feature standardization algorithm to obtain standardized static features, wherein the static feature standardization algorithm is as follows:

[0131]

[0132] In the formula, For the first c The static characteristics of each channel after standardization. For the first c The static characteristics of each channel For the first c The average of each channel, For the first c The standard deviation of each channel It is the numerical stability constant. For the first c Position in the static features of each channel The value, H For height parameters, W This is the width parameter.

[0133] For the dynamic features, the prediction device performs standardization processing on the dynamic features according to a preset dynamic feature standardization algorithm to obtain standardized dynamic features, wherein the dynamic feature standardization algorithm is:

[0134]

[0135] In the formula, For the standardized dynamic features, As a dynamic feature, For the rainstorm scenario parameter vector, The function for calculating the attenuation factor. The symbol for element-wise multiplication.

[0136] S53: Input the standardized static features and dynamic features into the feature processing module to extract spatiotemporal information, and obtain static spatial information features and dynamic temporal information features; use an attention mechanism to fuse the static spatial information features and dynamic temporal information features to obtain spatiotemporal fused information features.

[0137] In this embodiment, the prediction device inputs the standardized static features and dynamic features into the feature processing module for spatiotemporal information extraction to obtain static spatial information features and dynamic temporal information features, as described below:

[0138]

[0139]

[0140] In the formula, It is a static spatial information feature. For CNN encoder functions, These are the static features after standardization. As a feature of dynamic time information, For ConvLSTM encoder functions, The hidden state at time step t, For time step t The state of cells, This is the hidden state from the previous time step. This represents the cell state at the previous time step. For time step t The dynamic characteristics after standardization.

[0141] The prediction device employs an attention mechanism to fuse the static spatial information features and dynamic temporal information features to obtain spatiotemporal fusion information features.

[0142] Specifically, the prediction device uses the static spatial information features as a query matrix and the dynamic temporal information features as a key matrix and a value matrix. Based on a preset attention calculation algorithm, it fuses the static spatial information features and the dynamic temporal information features to obtain spatiotemporal fused information features. The attention calculation algorithm is as follows:

[0143]

[0144] In the formula, For spatiotemporal fusion information features, For normalized exponential functions, Q For querying the matrix, K The key matrix, V For value matrices, For dimension parameters, T This is the transpose symbol.

[0145] S54: Input the spatiotemporal fusion information features into the multi-task prediction module to perform initial inundation depth prediction and obtain initial inundation depth prediction data; extract flow velocity information and mass conservation residual information based on the inundation depth prediction data to obtain flow velocity information data and mass conservation residual information data; extract features based on the flow velocity information data and mass conservation residual information data to obtain flow velocity feature representation and mass conservation residual feature representation.

[0146] In this embodiment, the prediction device inputs the spatiotemporal fusion information features into the multi-task prediction module to perform initial inundation depth prediction, thereby obtaining initial inundation depth prediction data.

[0147]

[0148] In the formula, This is the initial flood depth prediction data. For activation function, As the first weight, For the first bias, The maximum possible water depth.

[0149] The prediction device extracts flow velocity information and mass conservation residual information based on the flooding depth prediction data, obtaining flow velocity information data and mass conservation residual information data as follows:

[0150]

[0151] In the formula, v For flow velocity information data, n For the Manning roughness coefficient field, S For the surface slope field, For residual information data related to quality conservation, This is the flood depth prediction data from the previous moment. R Rainfall intensity, I Infiltration rate, This represents the divergence of the one-width flow vector.

[0152] The prediction device uses a feature normalization method to extract features based on the flow velocity information data and the mass conservation residual information data, respectively, to obtain flow velocity feature representation and mass conservation residual feature representation.

[0153] S55: The flow velocity feature representation and the mass conservation residual feature representation are spliced ​​together to obtain a feature splicing representation; multi-task prediction is performed based on the feature splicing representation to obtain the flooding depth prediction data, waterlogging probability prediction data and flooding duration prediction data for each grid.

[0154] In this embodiment, the prediction device concatenates the flow velocity feature representation and the mass conservation residual feature representation to obtain a feature concatenation representation; based on the feature concatenation representation, multi-task prediction is performed to obtain flood depth prediction data, waterlogging probability prediction data, and flood duration prediction data for each grid, as described below:

[0155]

[0156] In the formula, For flood depth prediction data, As the second weight, For the second bias, For feature splicing representation, This is data for predicting the probability of urban flooding. As the third weight, For the third bias, For flood duration prediction data, As the fourth weight, For the fourth bias, It is a non-linear activation function.

[0157] In an optional embodiment, the method further includes step S7: performing backpropagation optimization on the flooding prediction model based on the flooding probability prediction data, inundation depth prediction data, inundation duration prediction data, flow velocity characteristic representation, and mass conservation residual characteristic representation of each grid cell; before step S6, please refer to [link to previous section]. Figure 7 , Figure 7 A flowchart illustrating step S7 of the urban flooding prediction method provided in another embodiment of this application is shown, including steps S71 to S75, as follows:

[0158] S71: Obtain terrain gradient data and label data for each grid cell.

[0159] In this embodiment, the prediction device obtains terrain gradient data and label data for each grid, wherein the label data includes waterlogging probability label data, flooding depth label data, and flooding duration label data.

[0160] S72: Based on the flooding probability prediction data, flooding depth prediction data, flooding duration prediction data of each grid, the tag data, and the preset data loss function, obtain the flooding probability loss value, flooding depth loss value, and flooding duration loss value.

[0161] In this embodiment, the prediction device obtains flooding probability loss values, flooding depth loss values, and flooding duration loss values ​​based on the flooding probability prediction data, flooding depth prediction data, flooding duration prediction data, the tag data, and a preset data loss function for each grid. The data loss function is as follows:

[0162]

[0163] In the formula, The loss value is the depth of flooding. N The number of grid cells. Location in flood depth prediction data The value, Location in the submerged depth label data The value, This represents the probability loss value of waterlogging. Location in the probability prediction data of urban flooding The value, Location in the probability label data of urban flooding The value, The value representing the loss due to the duration of flooding. Location in the flood duration prediction data The value, Location in the flood duration label data The value, for L 1-norm loss.

[0164] S73: Obtain the flow direction rationality loss value based on the flow velocity characteristics of each grid, the terrain gradient data, and the preset flow direction rationality loss function; obtain the Manning consistency loss value based on the inundation depth prediction data of each grid, the flow velocity characteristics, and the preset Manning consistency loss function.

[0165] In this embodiment, the prediction device obtains a flow direction rationality loss value based on the flow velocity characteristics of each grid, terrain gradient data, and a preset flow direction rationality loss function, wherein the flow direction rationality loss function is:

[0166]

[0167] In the formula, To determine the reasonable loss value of the flow, Position in the flow velocity feature representation The value, Location in terrain gradient data The value of .

[0168] In this embodiment, the prediction device obtains the Manning consistency loss value based on the flooding depth prediction data of each grid, the flow velocity characteristic representation, and a preset Manning consistency loss function, wherein the Manning consistency loss function is:

[0169]

[0170] In the formula, This represents the Manning consistency loss value. Position of Manning roughness coefficient The value, Location in surface slope data The value of .

[0171] S74: Obtain the mass conservation loss value based on the mass conservation residual characteristic representation of each grid and the preset mass conservation loss function.

[0172] In this embodiment, the prediction device obtains the mass conservation loss value based on the mass conservation residual feature representation of each grid and a preset mass conservation loss function, wherein the mass conservation loss function is:

[0173]

[0174] In the formula, The loss value due to mass conservation. Position in the residual characteristic representation of mass conservation The value of .

[0175] S75: Based on the waterlogging probability loss value, inundation depth loss value, inundation duration loss value, flow direction rationality loss value, Manning consistency loss value, and mass conservation loss value of each grid, the waterlogging prediction model is optimized by backpropagation.

[0176] In this embodiment, the prediction device normalizes the waterlogging probability loss value, inundation depth loss value, inundation duration loss value, flow direction rationality loss value, Manning consistency loss value, and mass conservation loss value of each grid, and multiplies them with the corresponding weight coefficients. The results of the multiplication are accumulated to obtain the total loss value. Based on the total loss value, the waterlogging prediction model is optimized by backpropagation.

[0177] S6: Identify waterlogging risk areas based on the flooding prediction data, waterlogging probability prediction data, and flooding duration prediction data in the waterlogging prediction data of each grid, and obtain the waterlogging risk areas of the area to be predicted.

[0178] In this embodiment, the prediction device identifies flood risk areas based on the flooding prediction data of each grid, including flooding depth prediction data, flooding probability prediction data, and flooding duration prediction data, and obtains the flood risk areas of the area to be predicted.

[0179] Please see Figure 8 , Figure 8 The flowchart of S6 in the urban flooding prediction method provided in one embodiment of this application is as follows: S61~S62 are detailed below:

[0180] S61: Based on the predicted probability of waterlogging occurrence, predicted inundation depth, predicted inundation duration, and a preset comprehensive waterlogging risk index calculation algorithm for each grid, obtain the comprehensive waterlogging risk index for each grid; classify the waterlogging risk level according to the comprehensive waterlogging risk index for each grid, obtain the waterlogging risk level for each grid, and construct a waterlogging risk map.

[0181] In this embodiment, the prediction device normalizes the flood depth prediction data and flood duration prediction data of each grid cell to obtain normalized flood depth prediction data and normalized flood duration prediction data of each grid cell.

[0182] The prediction device obtains the comprehensive waterlogging risk index for each grid cell based on the predicted probability of waterlogging occurrence, the predicted inundation depth after normalization, the predicted inundation duration after normalization, and a preset comprehensive waterlogging risk index calculation algorithm. The comprehensive waterlogging risk index calculation algorithm is as follows:

[0183]

[0184] In the formula, To comprehensively assess the risk of urban flooding, , The risk weights corresponding to the probability prediction data of urban flooding occurrence. Risk weights corresponding to flood depth prediction data. Risk weights corresponding to the flood duration prediction data. , This is data for predicting the probability of urban flooding. The flood depth prediction data is after normalization. This is the predicted inundation duration data after normalization.

[0185] The prediction device uses a quantile threshold division method to classify the waterlogging risk level according to the comprehensive waterlogging risk index of each grid, thereby obtaining the waterlogging risk level of each grid and constructing a waterlogging risk map.

[0186] S62: Identify connected regions and calculate vector boundaries of the connected regions on the waterlogging risk map to obtain several connected regions and vector boundary data of the connected regions; calculate a comprehensive risk score based on the vector boundary data of each connected region and a preset comprehensive risk score calculation algorithm to obtain a comprehensive risk score for each connected region; identify waterlogging risk areas based on the comprehensive risk scores of each connected region to obtain the waterlogging risk areas of the area to be predicted.

[0187] In this embodiment, the prediction device identifies connected regions and calculates the vector boundaries of the connected regions on the waterlogging risk map to obtain several connected regions and their vector boundary data.

[0188] Specifically, the prediction device extracts high-risk areas from the waterlogging risk map and generates a high-risk binary grid map. The high-risk binary grid map is converted into a labeled grid using an 8-connected region labeling algorithm. Then, the labeled grid is used to calculate the area, average risk index, maximum inundation depth, and regional compactness features of each connected region, thereby obtaining several connected regions and vector boundary data of the connected regions.

[0189] The prediction device calculates a comprehensive risk score for each connected region based on the vector boundary data of each connected region and a preset comprehensive risk score calculation algorithm. The prediction device then identifies waterlogging risk areas based on these comprehensive risk scores to obtain the waterlogging risk areas of the region to be predicted. The comprehensive risk score calculation algorithm is as follows:

[0190]

[0191] In the formula, A comprehensive risk score for the connected regions. For the index of the connected region, As area weight, The area of ​​the connected region. For the maximum area, As the average risk index weight, This represents the average risk index of the connected regions. Weighted by the maximum flooding depth. This represents the maximum flooding depth of the connected region. For the maximum flooding depth, As a weight for regional compactness, The regional compactness of the connected regions.

[0192] Please refer to Figure 9 , Figure 9 This is a schematic diagram of the structure of an urban flooding prediction device according to an embodiment of this application. The device can be implemented entirely or partially through software, hardware, or a combination of both. The urban flooding prediction device 9 includes:

[0193] The data acquisition module 91 is used to acquire hydrological connectivity correlation data and functional vulnerability index parameter set of several grids collected in the area to be predicted within a preset time period.

[0194] The first index calculation module 92 is used to perform upstream and downstream hydrological analysis based on the hydrological connectivity correlation data of each grid, and obtain the upstream contribution term and downstream impedance term of each grid; and to calculate the hydrological connectivity index based on the upstream contribution term and downstream impedance term of each grid to obtain the hydrological connectivity index of each grid.

[0195] The second index calculation module 93 is used to calculate the comprehensive vulnerability index based on the functional vulnerability index parameter set of each grid, and obtain the comprehensive vulnerability index of each grid.

[0196] The coupling factor construction module 94 is used to construct coupling factors based on the hydrological connectivity index and the comprehensive vulnerability index of each grid, and obtain the physical-dynamic coupling factor of each grid.

[0197] The waterlogging prediction module 95 is used to input the hydrological connectivity correlation data, hydrological connectivity index, comprehensive vulnerability index and physical-dynamic coupling factor of each grid into the preset waterlogging prediction model to make waterlogging predictions and obtain waterlogging prediction data for each grid.

[0198] The area identification module 96 is used to identify the flood risk area based on the flood depth prediction data, flood occurrence probability prediction data and flood duration prediction data in the flood prediction data of each grid, and obtain the flood risk area of ​​the area to be predicted.

[0199] In this embodiment, the data acquisition module obtains hydrological connectivity correlation data and a set of functional vulnerability index parameters for several grids collected within a preset time period in the area to be predicted; the first index calculation module performs upstream and downstream hydrological analysis based on the hydrological connectivity correlation data of each grid to obtain the upstream contribution and downstream impedance terms of each grid; the hydrological connectivity index is calculated based on the upstream contribution and downstream impedance terms of each grid to obtain the hydrological connectivity index of each grid; the second index calculation module performs a comprehensive vulnerability index calculation based on the set of functional vulnerability index parameters of each grid to obtain the comprehensive vulnerability index of each grid; and the process is further coupled... The factor construction module constructs coupling factors based on the hydrological connectivity index and the comprehensive vulnerability index of each grid, obtaining the physical-dynamic coupling factors for each grid. The waterlogging prediction module inputs the hydrological connectivity correlation data, hydrological connectivity index, comprehensive vulnerability index, and physical-dynamic coupling factors of each grid into a preset waterlogging prediction model to predict waterlogging, obtaining waterlogging prediction data for each grid. The region identification module identifies waterlogging risk areas based on the inundation depth prediction data, waterlogging probability prediction data, and inundation duration prediction data from the waterlogging prediction data of each grid, obtaining the waterlogging risk areas for the region to be predicted. Combining hydrological connectivity correlation data and functional vulnerability index parameter sets to identify waterlogging risk areas provides scientific guidance for disaster prevention and future urban planning.

[0200] Please refer to Figure 10 , Figure 10 This is a schematic diagram of the structure of a computer device provided in one embodiment of this application. The computer device 10 includes: a processor 101, a memory 102, and a computer program 103 stored in the memory 102 and executable on the processor 101. The computer device can store multiple instructions, which are adapted to be loaded and executed by the processor 101. Figures 1 to 8 For the method steps and specific execution process, please refer to [link / reference]. Figures 1 to 8 Specific details will not be elaborated here.

[0201] The processor 101 may include one or more processing cores. The processor 101 connects to various parts of the server using various interfaces and lines, and executes various functions and processes data of the urban flooding prediction device 9 by running or executing instructions, programs, code sets, or instruction sets stored in the memory 102, and by calling data stored in the memory 102. Optionally, the processor 101 may be implemented using at least one hardware form of Digital Signal Processing (DSP), Field-Programmable Gate Array (FPGA), or Programmable Logic Array (PLA). The processor 101 may integrate one or a combination of several of the following: Central Processing Unit (CPU), Graphics Processing Unit (GPU), and modem. The CPU primarily handles the operating system, user interface, and applications; the GPU is responsible for rendering and drawing the content to be displayed on the touch screen; and the modem handles wireless communication. It is understood that the modem may also not be integrated into the processor 101 and may be implemented as a separate chip.

[0202] The memory 102 may include random access memory (RAM) or read-only memory. Optionally, the memory 102 may include a non-transitory computer-readable storage medium. The memory 102 can be used to store instructions, programs, code, code sets, or instruction sets. The memory 102 may include a program storage area and a data storage area, wherein the program storage area may store instructions for implementing an operating system, instructions for at least one function (such as touch instructions), instructions for implementing the various method embodiments described above, etc.; the data storage area may store data involved in the various method embodiments described above, etc. Optionally, the memory 102 may also be at least one storage device located remotely from the aforementioned processor 101.

[0203] This application embodiment also provides a storage medium that can store multiple instructions, which are adapted to be loaded and executed by a processor as described above. Figures 1 to 8 For the method steps and specific execution process, please refer to [link / reference]. Figures 1 to 8 Specific details will not be elaborated here.

[0204] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the above-described division of functional units and modules is merely an example. In practical applications, the above functions can be assigned to different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above. The functional units and modules in the embodiments can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit. Furthermore, the specific names of the functional units and modules are only for easy differentiation and are not intended to limit the scope of protection of this application. The specific working process of the units and modules in the above system can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0205] In the above embodiments, the descriptions of each embodiment have different focuses. For parts that are not described in detail or recorded in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0206] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the algorithm. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this invention.

[0207] In the embodiments provided by this invention, it should be understood that the disclosed apparatus / terminal devices and methods can be implemented in other ways. For example, the apparatus / terminal device embodiments described above are merely illustrative. For instance, the division of modules or units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between devices or units may be electrical, mechanical, or other forms.

[0208] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0209] Furthermore, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0210] If the integrated module / unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, all or part of the processes in the methods of the above embodiments of the present invention can also be implemented 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 various method embodiments described above. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms.

[0211] This invention is not limited to the above-described embodiments. If any modifications or variations to this invention do not depart from the spirit and scope of this invention, and if such modifications and variations fall within the scope of the claims and equivalent technologies of this invention, then this invention also intends to include such modifications and variations.

Claims

1. A method for predicting urban flooding, characterized in that, Includes the following steps: Obtain hydrological connectivity correlation data and a set of functional vulnerability index parameters for several grids in the area to be predicted; Hydrological connectivity data of each grid is used to perform upstream and downstream hydrological analysis to obtain the upstream contribution and downstream impedance terms of each grid. Hydrological connectivity index is calculated based on the upstream contribution and downstream impedance terms of each grid to obtain the hydrological connectivity index of each grid. The comprehensive vulnerability index of each grid is calculated based on the set of functional vulnerability index parameters of each grid. The coupling factors are constructed based on the hydrological connectivity index and the comprehensive vulnerability index of each grid to obtain the physical-dynamic coupling factor of each grid. The hydrological connectivity correlation data, hydrological connectivity index, comprehensive vulnerability index and physical-dynamic coupling factor of each grid are input into a preset waterlogging prediction model to predict waterlogging, thereby obtaining waterlogging prediction data for each grid. The waterlogging prediction model includes a forward propagation module, a feature processing module and a multi-task prediction module. The step of inputting the hydrological connectivity correlation data, hydrological connectivity index, comprehensive vulnerability index, and physical-dynamic coupling factor of each grid into a preset waterlogging prediction model to predict waterlogging and obtain waterlogging prediction data for each grid includes the following steps: Based on the aforementioned hydrological connectivity correlation data, hydrological connectivity index, comprehensive vulnerability index, and physical-dynamic coupling factor, static and dynamic features are divided into static and dynamic features, and static and dynamic features are constructed. The static and dynamic features are input into the forward propagation module for standardization processing to obtain standardized static and dynamic features. The standardized static and dynamic features are input into the feature processing module for spatiotemporal information extraction to obtain static spatial information features and dynamic temporal information features; an attention mechanism is then used to fuse the static spatial information features and dynamic temporal information features to obtain spatiotemporal fused information features. The spatiotemporal fusion information features are input into the multi-task prediction module to perform initial inundation depth prediction, thereby obtaining initial inundation depth prediction data; flow velocity information and mass conservation residual information are extracted based on the inundation depth prediction data to obtain flow velocity information data and mass conservation residual information data; feature extraction is performed based on the flow velocity information data and mass conservation residual information data respectively to obtain flow velocity feature representation and mass conservation residual feature representation. The flow velocity feature representation and the mass conservation residual feature representation are spliced ​​together to obtain the feature splicing representation; multi-task prediction is performed based on the feature splicing representation to obtain the flooding depth prediction data, waterlogging probability prediction data and flooding duration prediction data for each grid. Based on the flooding prediction data, flooding probability prediction data, and flooding duration prediction data in the flooding prediction data of each grid, the flooding risk area is identified, and the flooding risk area of ​​the area to be predicted is obtained.

2. The urban flooding prediction method according to claim 1, characterized in that: The hydrological connectivity correlation data includes surface layer data and underground pipe network layer data. The surface layer data includes rainfall-related data, slope data, blocking factor, and area data. The rainfall-related data includes effective areal rainfall data and rainfall intensity data. The surface layer data is used to indicate the spatial movement of surface runoff in the raster. The underground pipe network layer data is used to indicate the spatial movement of water transported by underground pipes from the raster to the target sink. The step of performing upstream and downstream hydrological analysis based on the hydrological connectivity correlation data of each grid to obtain the upstream contribution and downstream impedance terms of each grid includes the following steps: The permeable and impermeable portions of the runoff are calculated based on the effective areal rainfall data of each grid, and the permeable and impermeable portion runoff data of each grid are obtained. The runoff weights of each grid are constructed based on the permeable and impermeable portion runoff data and the rainfall intensity data. Based on the slope data of each grid, the upstream and downstream grids of each grid are determined, and several outflow edges and inflow edges of each grid are constructed. Based on the slope data, retardation factor, and flow generation weight of each grid, the path cost of the outflow edges and inflow edges is calculated to obtain the path cost data of each outflow edge and each inflow edge of each grid. The outflow edge is used to indicate the directed edge from the current grid to the downstream neighboring grid; the inflow edge is used to indicate the directed edge from the upstream neighboring grid to the current grid. The upstream contribution is calculated based on the surface layer data of each grid and the path cost data of each inflow edge of each grid to obtain the upstream contribution of each grid. Based on the underground pipe network layer data of each grid, the water conveyance path from each grid to each target sink is determined; the water conveyance path includes several intermediate grids; the downstream impedance term is calculated based on the underground pipe network layer data of each grid and the path cost data of each outflow edge of each intermediate grid to obtain the downstream impedance term of each grid.

3. The urban flooding prediction method according to claim 2, characterized in that, The process of calculating the upstream contribution of each grid cell based on the surface layer data of each grid cell and the path cost data of each inflow edge of each grid cell includes the following steps: Based on the slope data, elevation data and path cost data of each inflow edge of each grid, the multi-directional diversion coefficient of each inflow edge is calculated to obtain the multi-directional diversion coefficient of each inflow edge of each grid. The local upstream source term is calculated based on the slope data, area data, and runoff weight of each grid cell to obtain the local upstream source term of each grid cell. Obtain the upstream cumulative supply of each grid cell; calculate the upstream contribution term based on the local upstream source term, upstream cumulative supply of each grid cell, and the multidirectional splitting coefficient of each inflow edge of each grid cell, and obtain the upstream contribution term of each grid cell.

4. The urban flooding prediction method according to claim 2, characterized in that: The underground pipeline network layer data includes target sink bottleneck capacity data and target sink water distribution weight; The process of calculating the downstream impedance term based on the underground pipeline network layer data of each grid and the path cost data of each outflow edge of each intermediate grid to obtain the downstream impedance term of each grid includes the following steps: The absorption efficiency of each target sink is calculated based on the bottleneck capacity data of each grid, the water distribution weight of the target sink, and the permeable part runoff data of each grid. The destination weight of each target sink is calculated based on the absorption efficiency of each target sink. The minimum path cost is calculated based on the path cost data of each outflow edge of each intermediate grid in the water conveyance path of each grid, and the minimum path cost data of each grid is obtained. Based on the obtained minimum path cost data of each grid and the destination weight of the target sink, the effective path cost is calculated to obtain the effective path cost data of each grid. The minimum value is taken as the downstream impedance term to obtain the downstream impedance term of each grid.

5. The urban flooding prediction method according to claim 1, characterized in that: The functional vulnerability index parameter set includes vulnerability index parameter data in several dimensions, including several positive vulnerability index parameters and negative vulnerability index parameters. The step of calculating the comprehensive vulnerability index based on the functional vulnerability index parameter set of each grid to obtain the comprehensive vulnerability index of each grid includes the following steps: The vulnerability index parameters of each dimension of each grid are standardized to obtain the positive and negative vulnerability index parameters of each dimension of each grid after standardization. The information entropy calculation method is adopted. Based on the positive vulnerability index parameters and negative vulnerability index parameters of each dimension of each grid after standardization, the information entropy weights are calculated to obtain the information entropy weights of each positive vulnerability index parameter and the information entropy weights of each negative vulnerability index parameter of each grid. The positive vulnerability index parameters and negative vulnerability index parameters of each dimension of each grid after standardization, along with the corresponding information entropy weights, are weighted to obtain the weighted values ​​of the positive vulnerability index parameters and negative vulnerability index parameters of each grid in each dimension. Sub-indices are calculated for each dimension of each grid based on the positive vulnerability index parameters, negative vulnerability index parameters, and the weighted values ​​of the positive and negative vulnerability index parameters. Sub-indices for each dimension of each grid are then combined to construct a sub-indice matrix for each dimension. The sub-indices for each grid in the sub-indice matrix for each dimension are then standardized to obtain the standardized sub-indices for each dimension of each grid. The information entropy calculation method is adopted. Based on the sub-indices of each dimension of each grid after standardization, the information entropy weights are calculated to obtain the information entropy weights of each dimension of each grid. Based on the sub-indices of each dimension of each grid after standardization and the corresponding information entropy weights, the weighted values ​​of the sub-indices of each dimension of each grid are obtained. The sub-indices are calculated by weighting the sub-indices of each dimension of each grid, and the obtained sub-indices are used as the comprehensive vulnerability index to obtain the comprehensive vulnerability index of each grid.

6. The urban flooding prediction method according to claim 1, characterized in that, It also includes the step of: performing backpropagation optimization on the waterlogging prediction model based on the waterlogging occurrence probability prediction data, inundation depth prediction data, inundation duration prediction data, flow velocity characteristic representation, and mass conservation residual characteristic representation of each grid. The backpropagation optimization of the flooding prediction model based on the flooding occurrence probability prediction data, inundation depth prediction data, inundation duration prediction data, and mass conservation residual feature representation of each grid includes the following steps: Obtain terrain gradient data and label data for each grid, wherein the label data includes waterlogging probability label data, inundation depth label data, and inundation duration label data; Based on the flooding probability prediction data, flooding depth prediction data, flooding duration prediction data of each grid, the tag data, and the preset data loss function, the flooding probability loss value, flooding depth loss value, and flooding duration loss value are obtained. The flow direction rationality loss value is obtained based on the flow velocity characteristics of each grid, the terrain gradient data, and the preset flow direction rationality loss function; the Manning consistency loss value is obtained based on the inundation depth prediction data of each grid, the flow velocity characteristics, and the preset Manning consistency loss function. Based on the mass conservation residual characteristic representation of each grid and the preset mass conservation loss function, the mass conservation loss value is obtained; The flood prediction model is optimized by backpropagation based on the flood occurrence probability loss value, inundation depth loss value, inundation duration loss value, flow direction rationality loss value, Manning consistency loss value, and mass conservation loss value of each grid.

7. The urban flooding prediction method according to claim 1, characterized in that, The process of identifying flood risk areas based on flooding prediction data, flooding probability prediction data, and flooding duration prediction data from each grid, to obtain the flood risk areas of the region to be predicted, includes the following steps: Based on the predicted probability of waterlogging, predicted inundation depth, predicted inundation duration, and a preset comprehensive waterlogging risk index calculation algorithm for each grid, a comprehensive waterlogging risk index for each grid is obtained; based on the comprehensive waterlogging risk index of each grid, waterlogging risk levels are classified to obtain the waterlogging risk level of each grid, and a waterlogging risk map is constructed. The waterlogging risk map is used to identify connected regions and calculate the vector boundaries of these regions to obtain several connected regions and their vector boundary data. Based on the vector boundary data of each connected region and a preset comprehensive risk scoring algorithm, a comprehensive risk score is calculated to obtain the comprehensive risk score for each connected region. Based on the comprehensive risk score of each connected region, the waterlogging risk area is identified, and the waterlogging risk area of ​​the region to be predicted is obtained.

8. An urban flooding prediction device, characterized in that, include: The data acquisition module is used to obtain hydrological connectivity correlation data and a set of functional vulnerability index parameters for several grids in the area to be predicted. The first index calculation module is used to perform upstream and downstream hydrological analysis based on the hydrological connectivity correlation data of each grid, and obtain the upstream contribution and downstream impedance terms of each grid; and to calculate the hydrological connectivity index of each grid based on the upstream contribution and downstream impedance terms of each grid. The second index calculation module is used to calculate the comprehensive vulnerability index based on the functional vulnerability index parameter set of each grid, and obtain the comprehensive vulnerability index of each grid. The coupling factor construction module is used to construct coupling factors based on the hydrological connectivity index and the comprehensive vulnerability index of each grid, and obtain the physical-dynamic coupling factor of each grid. The waterlogging prediction module is used to input the hydrological connectivity correlation data, hydrological connectivity index, comprehensive vulnerability index and physical-dynamic coupling factor of each grid into a preset waterlogging prediction model to predict waterlogging and obtain waterlogging prediction data for each grid. The waterlogging prediction model includes a forward propagation module, a feature processing module and a multi-task prediction module. The step of inputting the hydrological connectivity correlation data, hydrological connectivity index, comprehensive vulnerability index, and physical-dynamic coupling factor of each grid into a preset waterlogging prediction model to predict waterlogging and obtain waterlogging prediction data for each grid includes the following steps: Based on the aforementioned hydrological connectivity correlation data, hydrological connectivity index, comprehensive vulnerability index, and physical-dynamic coupling factor, static and dynamic features are divided into static and dynamic features, and static and dynamic features are constructed. The static and dynamic features are input into the forward propagation module for standardization processing to obtain standardized static and dynamic features. The standardized static and dynamic features are input into the feature processing module for spatiotemporal information extraction to obtain static spatial information features and dynamic temporal information features; an attention mechanism is then used to fuse the static spatial information features and dynamic temporal information features to obtain spatiotemporal fused information features. The spatiotemporal fusion information features are input into the multi-task prediction module to perform initial inundation depth prediction, thereby obtaining initial inundation depth prediction data; flow velocity information and mass conservation residual information are extracted based on the inundation depth prediction data to obtain flow velocity information data and mass conservation residual information data; feature extraction is performed based on the flow velocity information data and mass conservation residual information data respectively to obtain flow velocity feature representation and mass conservation residual feature representation. The flow velocity feature representation and the mass conservation residual feature representation are spliced ​​together to obtain the feature splicing representation; multi-task prediction is performed based on the feature splicing representation to obtain the flooding depth prediction data, waterlogging probability prediction data and flooding duration prediction data for each grid. The area identification module is used to identify waterlogging risk areas based on the flooding prediction data, waterlogging probability prediction data, and flooding duration prediction data in the waterlogging prediction data of each grid, and to obtain the waterlogging risk areas of the area to be predicted.

9. A computer device, characterized in that, It includes a processor, a memory, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of the urban flooding prediction method as described in any one of claims 1 to 7.

Images