Urban stormwater flooding rapid simulation method based on deep convolutional neural network

Abstract

The application discloses a city rainstorm waterlogging rapid simulation method based on a deep convolutional neural network, relates to the city waterlogging management field, and solves the problem of inaccurate judgment of the city waterlogging prevention capacity caused by the fact that the city waterlogging prevention capacity is usually obtained according to the city water supply and drainage capacity and historical peak rainfall conditions. The method comprises the following steps: setting a corresponding catchment area of a target city according to a city design map, dividing a catchment grid of the corresponding catchment area based on the catchment area division; constructing a height-area curve of the corresponding catchment area according to the catchment grid, constructing an overflow model corresponding to the catchment area based on the height-area curve, and improving the overflow model; simulating and identifying the result by using the improved overflow model, then comparing the result with a historical actual value, and judging the precision of the improved overflow model; obtaining real-time rainfall data and importing the real-time rainfall data into the improved overflow model, and obtaining the simulated flooding conditions of different catchment grids in the target city. The application realizes accurate simulation of city rainstorm waterlogging.

Description

Technical Field

[0001] This invention belongs to the field of urban flooding analysis technology, specifically a rapid simulation method for urban rainstorm flooding based on deep convolutional neural networks. Background Technology

[0002] Urban flooding refers to the phenomenon caused by short-term heavy rainfall or continuous torrential rain, which results in the inability of surface water in urban areas to be drained in time, leading to water accumulation in low-lying areas, flooding of roads, and backflow into underground spaces. Its essence is the mismatch between the capacity of the urban drainage system and the intensity of extreme rainfall, coupled with factors such as the increase of impermeable surfaces and the reduction of natural water storage space during the urbanization process, which triggers hydrological disasters.

[0003] However, at present, the assessment of urban flood control capacity is often obtained by comparing the city's water supply and drainage capacity with historical peak rainfall, rather than constructing a corresponding overflow model for the city to identify the flooded areas and duration. This leads to inaccurate assessments of urban flood control capacity and consequently, property damage.

[0004] To address this, the present invention proposes a rapid simulation method for urban rainstorm flooding based on deep convolutional neural networks. Summary of the Invention

[0005] The purpose of this invention is to propose a rapid simulation method for urban rainstorm flooding based on deep convolutional neural networks, so as to solve the problems mentioned in the background art.

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

[0007] A rapid simulation method for urban stormwater flooding based on deep convolutional neural networks is as follows:

[0008] Step S1: Collect the urban design map corresponding to the target city;

[0009] Step S2: Based on the urban design map, set the catchment area corresponding to the target city, and divide the catchment area into catchment grids based on the catchment area;

[0010] Step S3: Construct the height-area curve of the corresponding catchment area based on the catchment grid, and construct and improve the overflow model of the catchment area based on the height-area curve;

[0011] Step S4: Simulate and identify the results by improving the overflow model, and then compare the results with historical actual values ​​to determine the accuracy of the improved overflow model;

[0012] Step S5: Obtain real-time rainfall data and import it into the improved overflow model to obtain the simulated flooding situation of different catchment grids in the target city.

[0013] Further, step S2 includes the following sub-steps:

[0014] Step S201: Obtain the urban design map corresponding to the target city, and obtain the water supply pipeline layout, urban building layout and urban slope corresponding to the target city.

[0015] Step S202: Divide the urban design map corresponding to the target city into multiple water catchment areas;

[0016] Step S203: For any water catchment area, divide the corresponding water catchment area into multiple water catchment grids with a fixed grid side length, and randomly set a fixed number of grid detection points in each water catchment grid.

[0017] Furthermore, step S2 also includes the following sub-steps:

[0018] Step S204: Measure the elevation corresponding to the grid detection point, add up the elevations corresponding to different grid detection points in the same catchment grid, and take the average as the grid height SGGi of the corresponding catchment grid; where i is the number of different catchment grids in the corresponding catchment area, i=1,2,...,z, and z is a positive integer;

[0019] Step S205: The grid area of ​​different water catchment grids in the same water catchment area is denoted as SMJi. The grid area of ​​the boundary water catchment grid located in the water catchment area is measured. The grid area of ​​the non-boundary water catchment grid in the water catchment area is obtained by multiplying the grid side length by the grid side length. Among them, the boundary water catchment grid is the water catchment grid that the boundary of the water catchment area passes through.

[0020] Furthermore, step S3 includes the following sub-steps:

[0021] Step S301: Obtain the grid height and grid area of ​​the corresponding water catchment grid for different water catchment areas;

[0022] Step S302: For any catchment area, obtain the elevation of the wellhead corresponding to the inspection well in the catchment area;

[0023] Step S303: Construct a height-area curve for the corresponding catchment area based on the elevation of the wellhead of the inspection well and the grid height and grid area of ​​multiple water collection grids;

[0024] Step S304: The topography of the catchment area is generalized into a generalized pool, the bottom of which is the topography of the corresponding catchment area. An overflow model of the corresponding catchment area is constructed based on the SWMM model.

[0025] Furthermore, the overflow model satisfies the following conditions:

[0026] Condition 1: After overflowing from the manhole of the inspection well, the water flows directly into the generalized water pool of the corresponding catchment area and flows on the bottom of the generalized water pool.

[0027] Condition 2: After the water flows into the generalized pool, when the total volume of the water reaches a preset threshold, it will flow between the generalized pools corresponding to adjacent catchment areas.

[0028] Furthermore, step S3 also includes the following sub-steps:

[0029] Step S305, set parameters based on condition 1: denote the upstream node as a manhole and the downstream node as a generalized water tank. Water flows from the upstream node to the downstream node. Connecting pipes are set between the upstream and downstream nodes to control the flow of water. The bottom and top elevations of the manholes in the downstream nodes are the same as the actual bottom and top elevations of the manholes. The lowest point of the water collection grid is the same as the lowest point of the generalized water tank.

[0030] Step S306: Simulate water being poured into the connecting pipe until the water level rises to the wellhead. Record the direction of water flow. If the water flows towards the connecting pipe, no operation is performed. If the water flows towards the wellhead and overflows, reset the depth of the inspection well. The specific calculation process for the well depth in the model is as follows:

[0031] In the model, well depth = actual well depth + height of connecting pipe cross-section;

[0032] Actual well depth = Top elevation of inspection well - Bottom elevation of inspection well

[0033] Furthermore, step S3 also includes the following sub-steps:

[0034] Step S307, set parameters based on condition 2: select any catchment area and determine the adjacent catchment areas of the corresponding catchment area; wherein, the process of determining the adjacent catchment area is: if it shares the same boundary with the corresponding catchment area, then it is recorded as an adjacent catchment area;

[0035] Step S308: Connect the generalized water pool corresponding to the catchment area to the generalized water pool of the adjacent catchment area. When different generalized water pools are connected, the generalized water pool with a higher elevation corresponding to the bottom of the generalized water pool is recorded as the upstream node, and the generalized water pool with a higher elevation corresponding to the bottom of the generalized water pool is recorded as the downstream node. The upstream node and the downstream node are connected by a connecting pipe.

[0036] Furthermore, step S3 also includes the following sub-steps:

[0037] Step S309: Identify the boundary water catchment grids covered by the corresponding water catchment area and the adjacent water catchment area, identify the grid height corresponding to the boundary water catchment grids, and select the minimum value of the grid height as the preset connectivity height.

[0038] Set the upstream and downstream offsets of the connecting pipes, specifically as follows:

[0039] Upstream offset of pipeline = preset connection height minus the elevation of the bottom of the generalized pool at the upstream node;

[0040] Downstream offset of pipeline = preset connection height minus the elevation of the bottom of the generalized pool at the downstream node;

[0041] Step S310: If conditions 1 and 2 are satisfied simultaneously, the improved overflow model corresponding to the catchment area can be obtained; wherein, the upper layer of the improved overflow model is composed of interconnected generalized pools; the water flows along the terrain at the bottom of the generalized pools; the lower layer of the improved overflow model is composed of the pipe network system in the target city. If the water level does not exceed the overload depth, the water will only flow inside the overflow system and will not overflow.

[0042] Overload depth = maximum grid height of the corresponding water catchment area plus preset height minus the elevation of the manhole cover.

[0043] Further, step S4 includes the following sub-steps:

[0044] Step S41: Obtain the historical rainfall data of multiple target cities in historical years and import them into the improved overflow model;

[0045] Step S42: Read the simulated flood depth YMSi of the corresponding water catchment grid for all water catchment areas in the improved overflow model; wherein, if the water catchment grid is not flooded, the value of the simulated flood depth of the corresponding water catchment grid is zero.

[0046] Step S43: Read the actual inundation data corresponding to historical rainfall to obtain the historical inundation depth LMSi for each catchment grid; calculate the simulated inundation deviation value MPC for the corresponding catchment area using the following formula:

[0047] The larger the simulated flooding deviation value, the lower the accuracy of the surface-improved overflow model.

[0048] Step S44: Add the simulated inundation deviation values ​​of different catchment areas corresponding to the same group of historical rainfall to obtain the simulated inundation deviation value of the target city in the corresponding group of historical rainfall.

[0049] Furthermore, step S4 also includes the following sub-steps:

[0050] Step S45: Similarly, calculate the simulated inundation deviation value of the city for the target city based on the historical rainfall data of multiple groups. If the simulated inundation deviation value is greater than or equal to the simulated inundation deviation threshold, the historical rainfall simulation of the corresponding group is deemed unqualified. If the simulated inundation deviation value is less than the simulated inundation deviation threshold, the historical rainfall simulation of the corresponding group is deemed qualified.

[0051] Step S46: Count the number of times the historical rainfall simulation failed to meet the standards;

[0052] Step S47: If the number of times the historical rainfall simulation fails is less than or equal to the preset threshold, proceed directly to the next step.

[0053] If the number of times the historical rainfall simulation fails exceeds a preset threshold, the improved overflow model will be regenerated until the number of times the historical rainfall simulation fails is less than the preset threshold.

[0054] In summary, due to the adoption of the above technical solution, the beneficial effects of the present invention are:

[0055] 1. This invention sets the catchment area corresponding to the target city based on the urban design map, can divide the catchment area into a catchment grid based on the catchment area, and then construct the height-area curve of the corresponding catchment area based on the catchment grid. Finally, the overflow model corresponding to the catchment area is constructed and improved through the height-area curve.

[0056] 2. This invention improves the overflow model to simulate and identify the results, and then compares the results with historical actual values ​​to determine the accuracy of the improved overflow model. At the same time, it also obtains real-time rainfall data and imports it into the improved overflow model to know the simulated flooding situation of different water catchment grids in the target city, thereby achieving accurate simulation of urban rainstorm flooding. Attached Figure Description

[0057] To facilitate understanding by those skilled in the art, the present invention will be further described below with reference to the accompanying drawings.

[0058] Figure 1 This is a flowchart illustrating the overall method of the present invention;

[0059] Figure 2 This is a schematic diagram of the catchment area in this invention;

[0060] Figure 3 This is a schematic diagram of the height-area curve of the catchment area in this invention;

[0061] Figure 4 This is a schematic diagram of the overflow model corresponding to condition 1 in this invention;

[0062] Figure 5 This is a schematic diagram of the overflow model corresponding to condition 2 in this invention;

[0063] Figure 6 This is a schematic diagram of the improved overflow model in this invention;

[0064] Figure 7 This is a schematic diagram of the computer device in this invention. Detailed Implementation

[0065] The technical solution of the present invention will be clearly and completely described below with reference to the embodiments. 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.

[0066] Example 1, please refer to Figures 1-6 As shown, the technical solution provided by the present invention is: a rapid simulation method for urban rainstorm flooding based on deep convolutional neural networks, which divides the target city into multiple catchment areas, constructs an overflow model for each catchment area, and uses the overflow model to determine the urban rainstorm flooding situation of the target city and take appropriate measures.

[0067] In this invention, the rapid simulation method for urban rainstorm flooding is specifically as follows:

[0068] Step S1: Collect the urban design map corresponding to the target city; wherein, the urban design map includes the water supply pipeline layout, urban building layout and urban slope corresponding to the target city.

[0069] Step S2: Based on the urban design map, set the catchment area corresponding to the target city, and divide the catchment area into catchment grids based on the catchment area;

[0070] In this invention, step S2 includes the following sub-steps:

[0071] Step S201: Obtain the urban design map corresponding to the target city, and obtain the water supply pipeline layout, urban building layout and urban slope corresponding to the target city.

[0072] Step S202: According to the definition in Baidu Encyclopedia, combined with the layout of water supply pipelines, urban building layout and urban slope, the urban design map of the target city is divided into multiple water catchment areas; each water catchment area has one and only one outlet, and the rainwater in the water catchment area is eventually discharged into the outlet in the form of runoff.

[0073] Step S203, as follows Figure 2 As shown, for any catchment area, the corresponding catchment area is divided into multiple catchment grids with a fixed grid side length, and a fixed number of grid detection points are randomly set in each catchment grid.

[0074] Step S204: Measure the elevation corresponding to the grid detection point, add up the elevations corresponding to different grid detection points in the same catchment grid, and take the average as the grid height SGGi of the corresponding catchment grid; where i is the number of different catchment grids in the corresponding catchment area, i=1,2,...,z, and z is a positive integer;

[0075] Step S205: The grid area of ​​different water catchment grids in the same water catchment area is denoted as SMJi. The grid area of ​​the boundary water catchment grid located in the water catchment area is measured. The grid area of ​​the non-boundary water catchment grid in the water catchment area is obtained by multiplying the grid side length by the grid side length. Among them, the boundary water catchment grid is the water catchment grid that the boundary of the water catchment area passes through.

[0076] It should be noted that the boundary drainage grid is only partially located within the drainage area and is mostly an irregular geometric shape. Therefore, this invention chooses to obtain the area of ​​the Binajie drainage grid within the drainage area by direct measurement.

[0077] Step S3: Construct the height-area curve of the corresponding catchment area based on the catchment grid, and construct and improve the overflow model of the catchment area based on the height-area curve;

[0078] In this invention, step S3 includes the following sub-steps:

[0079] Step S301: Obtain the grid height and grid area of ​​the corresponding water catchment grid for different water catchment areas;

[0080] Step S302: For any catchment area, obtain the elevation of the wellhead corresponding to the inspection well in the catchment area;

[0081] Step S303, as Figure 3 As shown, a height-area curve for the corresponding catchment area is constructed based on the elevation of the manhole and the grid height and area of ​​multiple water collection grids.

[0082] Step S304: The topography of the catchment area is generalized into a generalized pool, the bottom of which represents the topography of the corresponding catchment area. An overflow model for the corresponding catchment area is constructed based on the SWMM model. The overflow model satisfies the following conditions:

[0083] Condition 1: After overflowing from the manhole of the inspection well, the water flows directly into the generalized water pool of the corresponding catchment area and flows on the bottom of the generalized water pool.

[0084] Condition 2: After the water flows into the generalized pool, when the total volume of the water reaches a preset threshold, it will flow between the generalized pools corresponding to adjacent catchment areas.

[0085] For step S305, please refer to [link / reference]. Figure 4 As shown, based on condition 1, the parameters are set as follows: the upstream node is denoted as a manhole, the downstream node is denoted as a generalized water tank, the water flows from the upstream node to the downstream node, and a connecting pipe is set between the upstream and downstream nodes to control the flow of the water; wherein, the bottom elevation and top elevation of the manhole in the downstream node are the same as the bottom elevation and top elevation of the actual manhole; the lowest point of the water collection grid corresponding to the grid height is the same as the lowest point of the generalized water tank.

[0086] It should be noted that the connecting pipe only serves a connection function, so when setting it up, a shorter length (such as 0.1 meters) and a smaller cross-section (such as a 1m×1m closed cross-section) can be selected.

[0087] Step S306: Simulate water being poured into the connecting pipe until the water level rises to the wellhead. Record the direction of water flow. If the water flows towards the connecting pipe, no operation is performed. If the water flows towards the wellhead and overflows, reset the depth of the inspection well. The specific calculation process for the well depth in the model is as follows:

[0088] In the model, well depth = actual well depth + height of connecting pipe cross-section;

[0089] Actual well depth = Top elevation of inspection well - Bottom elevation of inspection well;

[0090] It should be explained that, after the settings of steps S305-S306, the overflow model can realize that the water overflows from the wellhead and directly enters the generalized water tank through the connecting pipe.

[0091] For step S307, please refer to... Figure 5 As shown, based on condition 2, the parameters are set as follows: select any catchment area and determine the adjacent catchment areas of the corresponding catchment area; the process of determining the adjacent catchment area is as follows: if it shares the same boundary with the corresponding catchment area, it is recorded as the adjacent catchment area.

[0092] Step S308: Connect the generalized water pool in the corresponding catchment area to the generalized water pool in the adjacent catchment area. When different generalized water pools are connected, the generalized water pool with a higher elevation at the bottom of the generalized water pool is recorded as the upstream node, and the generalized water pool with a higher elevation at the bottom of the generalized water pool is recorded as the downstream node. The upstream node and the downstream node are connected by a connecting pipe.

[0093] like Figure 5 The upstream node is generalized pool 1, and the downstream node is generalized pool 2;

[0094] Step S309: Identify the boundary water catchment grids covered by the corresponding water catchment area and the adjacent water catchment area, identify the grid height corresponding to the boundary water catchment grids, and select the minimum value of the grid height as the preset connectivity height.

[0095] Set the upstream and downstream offsets of the connecting pipes, specifically:

[0096] Upstream offset of pipeline = preset connection height minus the elevation of the bottom of the generalized pool at the upstream node;

[0097] Downstream offset of pipeline = preset connection height minus the elevation of the bottom of the generalized pool at the downstream node;

[0098] It should be noted that this setup allows the connecting pipes to remain horizontal, meaning that when the water level in either of the two generalized pools rises to the preset connection height, the water can flow to the other generalized pool; and the connecting pipes in this step are also only used for connection purposes, with the same specific parameters as above.

[0099] Step S310: By simultaneously satisfying conditions 1 and 2, the improved overflow model corresponding to the catchment area can be obtained. The improved overflow model is as follows: Figure 6 As shown; the upper layer of the improved overflow model consists of interconnected generalized pools, the bottom of which reflects the surface topography; the water flows along the topography at the bottom of the generalized pools; the lower layer of the improved overflow model consists of the pipe network system in the target city. If the water level does not exceed the overload depth, the water will only flow within the overflow system and will not overflow.

[0100] Overload depth = maximum grid height of the corresponding water catchment area plus preset height minus the elevation of the manhole cover; the preset height is preferably five meters.

[0101] Step S4: Simulate and identify the results by improving the overflow model, and then compare the results with historical actual values ​​to determine the accuracy of the improved overflow model;

[0102] Step S41: Obtain the historical rainfall data of multiple target cities in historical years and import them into the improved overflow model;

[0103] Step S42: Read the simulated flood depth YMSi of the corresponding catchment grid for all catchment areas in the improved overflow model; it should be noted that if the catchment grid is not flooded, the simulated flood depth value of the corresponding catchment grid is zero.

[0104] Step S43: Read the actual inundation data corresponding to historical rainfall to obtain the historical inundation depth LMSi for each catchment grid; calculate the simulated inundation deviation value MPC for the corresponding catchment area using the following formula:

[0105] The larger the simulated flooding deviation value, the lower the accuracy of the surface-improved overflow model.

[0106] Step S44: Add the simulated inundation deviation values ​​of different catchment areas corresponding to the same group of historical rainfall to obtain the simulated inundation deviation value of the target city in the corresponding group of historical rainfall.

[0107] Step S45: Similarly, calculate the simulated inundation deviation value of the city for the target city based on the historical rainfall data of multiple groups. If the simulated inundation deviation value is greater than or equal to the simulated inundation deviation threshold, the historical rainfall simulation of the corresponding group is deemed unqualified. If the simulated inundation deviation value is less than the simulated inundation deviation threshold, the historical rainfall simulation of the corresponding group is deemed qualified.

[0108] Step S46: Count the number of times the historical rainfall simulation failed to meet the standards;

[0109] Step S47: If the number of times the historical rainfall simulation fails is less than or equal to the preset threshold, proceed directly to the next step; if the number of times the historical rainfall simulation fails exceeds the preset threshold, regenerate the improved overflow model until the number of times the historical rainfall simulation fails is less than the preset threshold.

[0110] Step S5: Obtain real-time rainfall data and import it into the improved overflow model to understand the simulated flooding situation of different catchment grids in the target city and take corresponding measures.

[0111] Specifically, warning thresholds are set based on historical flooding data. If the simulated flooding of the catchment grid exceeds the warning threshold, the drainage system in the target city is optimized; if the simulated flooding of the catchment grid does not exceed the warning threshold, no action is taken.

[0112] In this application, if a corresponding calculation formula appears, the above calculation formula is a dimensionless calculation. The weighting coefficient, proportional coefficient and other coefficients in the formula are set to quantify each parameter to obtain a result value. The size of the weighting coefficient and proportional coefficient is only required to not affect the proportional relationship between the parameter and the result value.

[0113] Example 2, Figure 7This is a schematic diagram of a computer device, which may include a processor, a communication interface, memory, and a communication bus. The processor, communication interface, and memory communicate with each other via the communication bus. The processor can call logical instructions in the memory to execute a rapid simulation method for urban stormwater flooding based on a deep convolutional neural network. This method includes: acquiring urban design maps of the target city; defining the catchment area of ​​the target city based on the urban design maps; dividing the catchment area into corresponding catchment grids; constructing height-area curves for the corresponding catchment areas based on the catchment grids; constructing and improving overflow models for the catchment areas based on the height-area curves; simulating and identifying the results using the improved overflow model; comparing the results with historical actual values ​​to determine the accuracy of the improved overflow model; acquiring real-time rainfall data and importing it into the improved overflow model to determine the simulated flooding situation of different catchment grids in the target city.

[0114] Furthermore, when the logical instructions in the aforementioned memory can be implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0115] Example 3: This application also provides a computer program product, which includes a computer program stored on a computer-readable storage medium. The computer program includes program instructions. When the program instructions are executed by the computer, the computer can execute the rapid simulation method for urban stormwater flooding based on deep convolutional neural networks provided by the above methods. The method includes: collecting urban design maps corresponding to the target city; setting the catchment area corresponding to the target city according to the urban design map; dividing the catchment area into catchment grids based on the catchment area; constructing the height-area curve of the catchment area based on the catchment grid; constructing and improving the overflow model corresponding to the catchment area based on the height-area curve; simulating and identifying the results through the improved overflow model; comparing the results with historical actual values ​​to determine the accuracy of the improved overflow model; acquiring real-time rainfall data and importing it into the improved overflow model to know the simulated flooding situation of different catchment grids in the target city.

[0116] Example 4: This application also provides a computer-readable storage medium storing a computer program thereon. When executed by a processor, the computer program implements the above-described methods for rapid simulation of urban stormwater flooding based on deep convolutional neural networks. The method includes: acquiring urban design maps corresponding to the target city; setting the catchment area corresponding to the target city based on the urban design map; dividing the catchment area into catchment grids based on the catchment area; constructing a height-area curve for the catchment area based on the catchment grid; constructing and improving an overflow model for the catchment area based on the height-area curve; simulating and identifying the results through the improved overflow model; comparing the results with historical actual values ​​to determine the accuracy of the improved overflow model; acquiring real-time rainfall data and importing it into the improved overflow model to determine the simulated flooding situation of different catchment grids in the target city.

[0117] The device embodiments described above are merely illustrative. 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 modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0118] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0119] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application.

Claims

1. A rapid simulation method for urban rainstorm flooding based on deep convolutional neural networks, characterized in that, The method is as follows: Step S1: Collect the urban design map corresponding to the target city; Step S2: Based on the urban design map, set the catchment area corresponding to the target city, and divide the catchment area into catchment grids based on the catchment area; Step S3: Construct the height-area curve of the corresponding catchment area based on the catchment grid, and construct and improve the overflow model of the catchment area based on the height-area curve; Step S3 includes the following sub-steps: Step S301: Obtain the grid height and grid area of ​​the corresponding water catchment grid for different water catchment areas; Step S302: For any catchment area, obtain the elevation of the wellhead corresponding to the inspection well in the catchment area; Step S303: Construct a height-area curve for the corresponding catchment area based on the elevation of the wellhead and the grid height and area of ​​multiple water collection grids; Step S304: The topography of the catchment area is generalized into a generalized pool, and the bottom surface of the generalized pool is the topography of the corresponding catchment area. An overflow model of the corresponding catchment area is constructed based on the SWMM model. The overflow model satisfies the following conditions: Condition 1: After overflowing from the manhole of the inspection well, the water flows directly into the generalized water pool of the corresponding catchment area and flows on the bottom of the generalized water pool. Condition 2: After the water flows into the generalized pool, when the total volume of the water reaches a preset threshold, it will flow between the generalized pools corresponding to adjacent catchment areas. Step S305, set parameters based on condition 1: denote the upstream node as a manhole and the downstream node as a generalized water tank. Water flows from the upstream node to the downstream node. Connecting pipes are set between the upstream and downstream nodes to control the flow of water. The bottom and top elevations of the manholes in the downstream nodes are the same as the actual bottom and top elevations of the manholes. The lowest point of the water collection grid is the same as the lowest point of the generalized water tank. Step S306: Simulate water being poured into the connecting pipe until the water level rises to the wellhead. Record the direction of water flow. If the water flows towards the connecting pipe, no operation is performed. If the water flows towards the wellhead and overflows, reset the depth of the inspection well. The specific calculation process for the well depth in the model is as follows: In the model, well depth = actual well depth + height of connecting pipe cross-section; Actual well depth = Top elevation of inspection well - Bottom elevation of inspection well; Step S4: Simulate and identify the results by improving the overflow model, and then compare the results with historical actual values ​​to determine the accuracy of the improved overflow model; Step S5: Obtain real-time rainfall data and import it into the improved overflow model to obtain the simulated flooding situation of different catchment grids in the target city.

2. The rapid simulation method for urban rainstorm flooding based on deep convolutional neural networks according to claim 1, characterized in that, Step S2 includes the following sub-steps: Step S201: Obtain the urban design map corresponding to the target city, and obtain the water supply pipeline layout, urban building layout and urban slope of the target city. Step S202: Divide the urban design map corresponding to the target city into multiple watershed areas; Step S203: For any water catchment area, divide the corresponding water catchment area into multiple water catchment grids with a fixed grid side length, and randomly set a fixed number of grid detection points in each water catchment grid.

3. The rapid simulation method for urban rainstorm flooding based on deep convolutional neural networks according to claim 2, characterized in that, Step S2 further includes the following sub-steps: Step S204: Measure the elevation corresponding to the grid detection point, add up the elevations corresponding to different grid detection points in the same catchment grid, and take the average as the grid height SGGi of the corresponding catchment grid; where i is the number of different catchment grids in the corresponding catchment area, i=1,2,...,z, and z is a positive integer; Step S205: The grid area of ​​different water catchment grids in the same water catchment area is denoted as SMJi. The grid area of ​​the boundary water catchment grid located in the water catchment area is measured. The grid area of ​​the non-boundary water catchment grid in the water catchment area is obtained by multiplying the grid side length by the grid side length. Among them, the boundary water catchment grid is the water catchment grid that the boundary of the water catchment area passes through.

4. The rapid simulation method for urban rainstorm flooding based on deep convolutional neural networks according to claim 1, characterized in that, Step S3 further includes the following sub-steps: Step S307, set parameters based on condition 2: select any catchment area and determine the adjacent catchment areas of the corresponding catchment area; wherein, the process of determining the adjacent catchment area is: if it shares the same boundary with the corresponding catchment area, then it is recorded as an adjacent catchment area; Step S308: Connect the generalized water pool corresponding to the catchment area to the generalized water pool of the adjacent catchment area. When different generalized water pools are connected, the generalized water pool with a higher elevation corresponding to the bottom of the generalized water pool is recorded as the upstream node, and the generalized water pool with a higher elevation corresponding to the bottom of the generalized water pool is recorded as the downstream node. The upstream node and the downstream node are connected by a connecting pipe.

5. The rapid simulation method for urban rainstorm flooding based on deep convolutional neural networks according to claim 3, characterized in that, Step S3 further includes the following sub-steps: Step S309: Identify the boundary water catchment grids covered by the corresponding water catchment area and the adjacent water catchment area, identify the grid height corresponding to the boundary water catchment grids, and select the minimum value of the grid height as the preset connectivity height. Set the upstream and downstream offsets of the connecting pipes, specifically as follows: Upstream offset of pipeline = preset connection height minus the elevation of the bottom of the generalized pool at the upstream node; Downstream offset of pipeline = preset connection height minus the elevation of the bottom of the generalized pool at the downstream node; Step S310: If conditions 1 and 2 are satisfied simultaneously, the improved overflow model corresponding to the catchment area can be obtained; wherein, the upper layer of the improved overflow model is composed of interconnected generalized pools; the water flows along the terrain at the bottom of the generalized pools; the lower layer of the improved overflow model is composed of the pipe network system in the target city. If the water level does not exceed the overload depth, the water will only flow inside the overflow system and will not overflow. Overload depth = maximum grid height of the corresponding water catchment area plus preset height minus the elevation of the manhole cover.

6. The rapid simulation method for urban rainstorm flooding based on deep convolutional neural networks according to claim 5, characterized in that, Step S4 includes the following sub-steps: Step S41: Obtain the historical rainfall data of multiple target cities in historical years and import them into the improved overflow model; Step S42: Read the simulated flood depth YMSi of the corresponding water catchment grid for all water catchment areas in the improved overflow model; wherein, if the water catchment grid is not flooded, the value of the simulated flood depth of the corresponding water catchment grid is zero. Step S43: Read the actual inundation data corresponding to historical rainfall to obtain the historical inundation depth LMSi for each catchment grid; calculate the simulated inundation deviation value MPC for the corresponding catchment area using the following formula: The larger the simulated flooding deviation value, the lower the accuracy of the surface-improved overflow model. Step S44: Add the simulated inundation deviation values ​​of different catchment areas corresponding to the same group of historical rainfall to obtain the simulated inundation deviation value of the target city in the corresponding group of historical rainfall.

7. The rapid simulation method for urban rainstorm flooding based on deep convolutional neural networks according to claim 6, characterized in that, Step S4 further includes the following sub-steps: Step S45: Similarly, calculate the simulated inundation deviation value of the city for the target city based on the historical rainfall data of multiple groups. If the simulated inundation deviation value is greater than or equal to the simulated inundation deviation threshold, the historical rainfall simulation of the corresponding group is deemed unqualified. If the simulated inundation deviation value is less than the simulated inundation deviation threshold, the historical rainfall simulation of the corresponding group is deemed qualified. Step S46: Count the number of times the historical rainfall simulation failed to meet the standards; Step S47: If the number of times the historical rainfall simulation fails is less than or equal to the preset threshold, proceed directly to the next step. If the number of times the historical rainfall simulation fails exceeds a preset threshold, the improved overflow model will be regenerated until the number of times the historical rainfall simulation fails is less than the preset threshold.

Images