Rainwater pipe network hydraulic bearing capacity assessment method, device and equipment and storage medium

Abstract

The invention provides a rainwater pipe network hydraulic bearing capacity assessment method, device and equipment and a storage medium, and relates to the technical field of rainwater pipe networks. The method comprises the following steps: acquiring rainwater pipe network information of a to-be-evaluated area, and converting the rainwater pipe network information into graph data of a rainwater pipe network; the rainwater pipe network information comprises pipe section characteristics of all target pipe sections and a connection relation of all the target pipe sections, the graph data of the rainwater pipe network comprises a pipe network node attribute characteristic matrix and a pipe network connection matrix, and the pipe network node attribute characteristic matrix is obtained based on the pipe section characteristics of all the target pipe sections; the pipe network connection matrix is determined based on the connection relation of all the target pipe sections; inputting the graph data of the rainwater pipe network into a pre-trained rainwater pipe network bearing capacity evaluation model, and outputting the maximum fullness of all target pipe sections in the to-be-evaluated area; the rainwater pipe network bearing capacity evaluation model is constructed based on a graph neural network model. The method can improve the evaluation speed.

Description

Technical Field

[0001] The present invention relates to the technical field of rainwater pipe networks, and in particular to a method, device, equipment and storage medium for evaluating the hydraulic bearing capacity of a rainwater pipe network. Background Art

[0002] As a key component of urban infrastructure, the rainwater pipe network is responsible for collecting, transporting and discharging urban rainwater, which is directly related to the city's water resource utilization and flood control and drainage safety. Against the backdrop of the continuous expansion of urban scale, continuous population growth and increasingly frequent extreme climate events, the effective operation of the rainwater pipe network is the basic guarantee for the normal operation of the city and an indispensable part of the process of urban resilience construction.

[0003] The current assessment of the hydraulic bearing capacity of rainwater pipe networks mainly relies on mechanism models. The mechanism model is based on the principles of hydraulics and constructs complex mathematical equations to simulate the movement of water in the rainwater pipe network, thereby assessing the drainage capacity of the pipe network.

[0004] However, the calculation process of the current evaluation method is extremely complicated and involves large-scale numerical calculations, which leads to low evaluation efficiency and makes it difficult to meet the needs of rapid decision-making. Especially when facing large-scale rainwater pipe network systems or situations where frequent evaluation updates are required, the current mechanism model is even more inadequate. Summary of the invention

[0005] The embodiments of the present invention provide a method, device, equipment and storage medium for evaluating the hydraulic bearing capacity of a rainwater pipe network, so as to solve the problem that the current method for evaluating the hydraulic bearing capacity of a rainwater pipe network requires a long evaluation time.

[0006] In a first aspect, an embodiment of the present invention provides a method for evaluating the hydraulic bearing capacity of a rainwater pipe network, comprising:

[0007] The rainwater pipe network information of the area to be evaluated is obtained, and the rainwater pipe network information is converted into the graph data of the rainwater pipe network; wherein the rainwater pipe network information includes the pipe segment characteristics of all target pipe segments and the connection relationship of all target pipe segments, and the graph data of the rainwater pipe network includes the pipe network node attribute feature matrix and the pipe network connection matrix, the pipe network node attribute feature matrix is ​​obtained by splicing the pipe segment characteristics of all target pipe segments, and the pipe network connection matrix is ​​determined based on the connection relationship of all target pipe segments; the target pipe segment is any pipe segment in the rainwater pipe network;

[0008] The graph data of the rainwater pipe network is input into a pre-trained rainwater pipe network carrying capacity assessment model, and the maximum fullness of all target pipe sections in the area to be assessed is output; wherein the rainwater pipe network carrying capacity assessment model is constructed based on a graph neural network model.

[0009] In a possible implementation, the pipe segment characteristics of the target pipe segment include pipe segment attributes, pipe segment upstream inspection well attributes, pipe segment downstream inspection well attributes, and rainfall parameters;

[0010] The connection relationship of all target pipe sections is determined based on whether they have a public inspection well.

[0011] In a possible implementation, the pipe segment attributes include the pipe segment length, the pipe segment cross-sectional area, the starting elevation, the ending elevation, the starting depth, and the ending depth;

[0012] The properties of the upstream inspection well of the pipe section include the ground elevation of the upstream inspection well of the pipe section, the buried depth of the upstream inspection well of the pipe section, the catchment area of ​​the upstream inspection well of the pipe section, the average slope of the catchment area of ​​the upstream inspection well of the pipe section, and the impermeability of the catchment area of ​​the upstream inspection well of the pipe section;

[0013] The properties of the downstream inspection well of the pipe section include the ground elevation of the downstream inspection well of the pipe section, the buried depth of the downstream inspection well of the pipe section, the catchment area of ​​the downstream inspection well of the pipe section, the average slope of the catchment area of ​​the downstream inspection well of the pipe section, and the impermeability of the catchment area of ​​the downstream inspection well of the pipe section;

[0014] Rainfall parameters include rainstorm return period.

[0015] In one possible implementation, a sample data set for training a stormwater pipe network bearing capacity assessment model is constructed based on the stormwater pipe network in a target area. The sample data set includes graph data of the stormwater pipe network in the target area and the maximum fullness of each pipe section in the stormwater pipe network in the target area. The graph data of the stormwater pipe network in the target area is constructed based on the stormwater pipe network information in the target area.

[0016] In one possible implementation, the maximum filling degree of each pipe segment in the stormwater pipe network within the target area is determined based on inputting the design storm into a pre-built stormwater flood management model;

[0017] The stormwater pipe network bearing capacity assessment model includes a sequentially connected input layer, a first hidden layer, a second hidden layer, a third hidden layer and an output layer. The first hidden layer, the second hidden layer and the third hidden layer all include graph convolution layers. The first hidden layer and the second hidden layer both include sequentially connected graph convolution layers, batch normalization layers, activation function layers and random inactivation layers. The third hidden layer includes sequentially connected graph convolution layers and fully connected layers.

[0018] In one possible implementation, the stormwater flood management model is constructed based on the structural information of the stormwater pipe network and the underlying surface parameters in the target area;

[0019] The structural information of the rainwater pipe network includes pipe length, pipe diameter, pipe shape, pipe top and bottom elevation, pipe top and bottom burial depth, inspection well burial depth and ground elevation;

[0020] The underlying surface parameters include the average slope, imperviousness and catchment area of ​​each catchment area.

[0021] In a second aspect, an embodiment of the present invention provides a rainwater pipe network hydraulic bearing capacity assessment device, comprising:

[0022] A conversion module is used to obtain the rainwater pipe network information of the area to be evaluated, and convert the rainwater pipe network information into the graph data of the rainwater pipe network; wherein the rainwater pipe network information includes the pipe segment characteristics of all target pipe segments and the connection relationship of all target pipe segments, and the graph data of the rainwater pipe network includes the pipe network node attribute feature matrix and the pipe network connection matrix, the pipe network node attribute feature matrix is ​​obtained by splicing the pipe segment characteristics of all target pipe segments, and the pipe network connection matrix is ​​determined based on the connection relationship of all target pipe segments; the target pipe segment is any pipe segment in the rainwater pipe network;

[0023] The evaluation module is used to input the graph data of the rainwater pipe network into a pre-trained rainwater pipe network bearing capacity evaluation model, and output the maximum fullness of all target pipe sections in the area to be evaluated; wherein the rainwater pipe network bearing capacity evaluation model is constructed based on a graph neural network model.

[0024] In a third aspect, an embodiment of the present invention provides an electronic device, including a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the method in the first aspect or any possible implementation of the first aspect is implemented.

[0025] In a fourth aspect, an embodiment of the present invention provides a computer-readable storage medium, which stores a computer program. When the computer program is executed by a processor, it implements the method in the first aspect or any possible implementation of the first aspect.

[0026] In a fifth aspect, an embodiment of the present invention provides a computer program product, including a computer program, which, when executed by a processor, implements the method in the first aspect or any possible implementation manner of the first aspect.

[0027] In an embodiment of the present invention, by converting the acquired rainwater pipe network information into the graph data of the rainwater pipe network, and inputting the obtained graph data of the rainwater pipe network into the pre-trained rainwater pipe network bearing capacity assessment model, the maximum fullness of all target pipe sections in the area to be assessed can be output. The rainwater pipe network bearing capacity assessment model in the present invention is constructed based on the graph neural network model. By converting the rainwater pipe network information into the graph data of the rainwater pipe network, the topological structure and attribute information of the rainwater pipe network can be effectively mapped to the low-dimensional vector space. Thereby, the rich topological relationship data of the drainage pipe network can be fully utilized to dig out the deep-level features hidden in the pipe network structure, and these features play a very important role in the evaluation accuracy of the pipe network's bearing capacity. Compared with the existing mechanism model, the assessment method provided by the present invention is more flexible in data processing and can integrate multi-source data more quickly. In addition, once the rainwater pipe network bearing capacity assessment model provided by the present invention is trained, the calculation speed is extremely fast when conducting the assessment, and it can provide strong support for the planning, management and emergency decision-making of the urban drainage pipe network in a short time, effectively making up for the shortcomings of the existing assessment method. Moreover, the rainwater pipe network bearing capacity assessment model provided by the present invention can be widely used once training is completed. When conducting hydraulic bearing capacity assessment, it is only necessary to obtain the rainwater pipe network information without rebuilding a new model, thereby further speeding up the assessment. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] Figure 1 is a structural schematic diagram of a rainwater pipe network bearing capacity assessment model provided by an embodiment of the present invention;

[0029] Figure 2 is a flow chart for implementing a method for evaluating the hydraulic bearing capacity of a rainwater pipe network provided by an embodiment of the present invention;

[0030] Figure 3 is a block diagram of a method for assessing the hydraulic bearing capacity of a rainwater pipe network provided by an embodiment of the present invention;

[0031] Figure 4 It is a schematic diagram of a rainwater pipe network in a certain area of ​​Shenzhen City provided by an embodiment of the present invention;

[0032] Figure 5 It is a schematic diagram of the SWMM model construction provided by an embodiment of the present invention;

[0033] Figure 6 It is a schematic diagram of a design rainstorm in Shenzhen provided by an embodiment of the present invention;

[0034] Figure 7 is a schematic diagram of pipe segment features and a connection matrix provided by an embodiment of the present invention;

[0035] Figure 8 is a schematic diagram of model training provided by an embodiment of the present invention;

[0036] Fig. 9 is a schematic diagram of a shallow neural network constructed according to an embodiment of the present invention;

[0037] Fig.10 It is a schematic diagram of the hydraulic bearing capacity assessment of a rainwater pipe network in a certain area of ​​Shenzhen City provided by an embodiment of the present invention;

[0038] Fig.11 It is a structural schematic diagram of a rainwater pipe network hydraulic bearing capacity assessment device provided by an embodiment of the present invention;

[0039] Fig.12 is a schematic diagram of an electronic device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0040] The embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0041] With the continuous acceleration of urbanization, the issue of urban rainwater discharge has received increasing attention. As an important infrastructure for urban rainwater discharge, the hydraulic carrying capacity of the rainwater pipe network is directly related to the drainage efficiency, waterlogging prevention and control capabilities of the city during rainfall, and the health of the urban water ecology.

[0042] The hydraulic bearing capacity of a rainwater pipe network refers to the maximum rainwater flow that the pipe network can withstand under specific pipe system structure, pipe diameter, slope, roughness and other conditions. It is a key indicator for measuring the drainage capacity of a rainwater pipe network. As introduced in the background technology, the current assessment of the hydraulic bearing capacity of a rainwater pipe network mainly relies on the mechanism model for assessment. However, the construction of the mechanism model requires a large number of detailed underlying surface parameters, and these underlying surface parameters are difficult and costly to obtain. In addition, the calculation process of the mechanism model is extremely complex and involves large-scale numerical calculations, which leads to low assessment efficiency and makes it difficult to meet the needs of rapid decision-making.

[0043] In order to solve the problems of the prior art, the embodiments of the present invention provide a method, device, equipment and storage medium for evaluating the hydraulic bearing capacity of a rainwater pipe network. In order to clearly introduce the method for evaluating the hydraulic bearing capacity of a rainwater pipe network provided by the present invention, it is necessary to first introduce the rainwater pipe network bearing capacity evaluation model in the present invention.

[0044] The following first introduces the rainwater pipe network bearing capacity assessment model provided by the embodiment of the present invention.

[0045] In some embodiments, the stormwater pipe network bearing capacity assessment model is constructed based on a graph neural network model.

[0046] In this embodiment, the constructed rainwater pipe network bearing capacity assessment model includes an input layer, a first hidden layer, a second hidden layer, a third hidden layer and an output layer connected in sequence. The first hidden layer, the second hidden layer and the third hidden layer all include graph convolution layers.

[0047] Specifically, Figure 1 As shown in the figure, the first hidden layer and the second hidden layer both include a graph convolution layer (GCN layer), a batch normalization layer (BatchNorm layer), an activation function layer (ReLU layer) and a random dropout layer (DropOut layer) connected in sequence, and the third hidden layer includes a graph convolution layer (GCN layer) and a fully connected layer (Linear layer) connected in sequence. Among them, the main structure of the first hidden layer and the second hidden layer are both graph convolution layers (GCN layers), and the first two convolution layers use BatchNorm layers for normalization, ReLU layers for activation, and DropOut layers to prevent overfitting. The data finally output by the output layer is the maximum fullness of all target pipe segments.

[0048] Exemplarily, the number of neurons in the first hidden layer, the second hidden layer, and the third hidden layer are set to 32, 32, and 16, respectively, and the neuron discarding rate of the Dropout layer is set to 0.5.

[0049] After the architecture of the stormwater pipe network carrying capacity assessment model is completed, it is also necessary to prepare the sample data set required for model training.

[0050] In some embodiments, the sample data set for training the rainwater pipe network bearing capacity assessment model is constructed based on the rainwater pipe network in the target area. The sample data set includes the graph data of the rainwater pipe network in the target area and the maximum fullness of each pipe section in the rainwater pipe network in the target area. The graph data of the rainwater pipe network in the target area is constructed based on the rainwater pipe network information of the target area.

[0051] In this embodiment, the rainwater pipe network in the target area may be a rainwater pipe network in any area and is not limited here.

[0052] In some embodiments, the maximum fullness of each pipe segment in the stormwater network within the target area is determined based on inputting the design storm into a pre-built stormwater flood management model.

[0053] In this embodiment, a sample data set may be constructed by selecting rainwater pipe networks (about 1,000 pipe networks) in a small target area within the area to be evaluated or in the surrounding adjacent areas.

[0054] First, a mechanism model based on the Storm Water Management Model (SWMM) was constructed based on the structural information and underlying surface parameters of the rainwater pipe network in the target area, and the outlet flow data under multiple rainfall conditions were collected at the outlet to calibrate and verify the constructed SWMM model. Among them, the structural information of the rainwater pipe network in the target area includes pipe length, pipe diameter, pipe shape, pipe top and bottom elevation, pipe top and bottom burial depth, inspection well burial depth and ground elevation. The underlying surface parameters include the average slope, impermeability and catchment area of ​​each catchment area.

[0055] Secondly, design storms and perform batch simulations. Design storms with different storm return periods are designed based on the storm intensity formula in the target area and the Chicago rain pattern. Use the design storm as rainfall input, run the SWMM model, and obtain the maximum fill of each pipe segment from the .rpt file of the run results.

[0056] Next, the rainwater pipe network information of the target area is represented in a graph embedding manner to obtain the graph data of the rainwater pipe network of the target area, with the pipe network as the node of the graph structure, and the public inspection wells between the pipe networks as the links of the graph structure. Among them, the rainwater pipe network information includes the pipe segment characteristics of each pipe segment and the connection relationship between each pipe segment. The graph data of the rainwater pipe network includes the pipe network node attribute feature matrix and the pipe network connection matrix. The pipe network node attribute feature matrix is ​​obtained by splicing the pipe segment characteristics of each pipe segment, and the pipe network connection matrix is ​​determined based on the connection relationship between each pipe segment.

[0057] Finally, the sample data set is divided into training set, test set and validation set. The graph data of the rainwater pipe network in the target area after graph embedding in the training set is input into the architecture of the pre-built rainwater pipe network carrying capacity assessment model, and the maximum fullness of each pipe section in the rainwater pipe network in the target area is used as the output of the regression prediction model. MSE is used as the loss function and Adam is used as the optimizer to continuously optimize the network parameters through back propagation. The loss values ​​of the training set and the validation set are monitored during the training process. When the loss value converges, the learning is stopped to obtain the model, and the error of the test set is evaluated. By tuning the hyperparameters in the model (learning rate, dropout ratio, weight decay, etc.), the optimal model is selected as the final rainwater pipe network carrying capacity assessment model.

[0058] In some embodiments, when determining the map data of the stormwater pipe network of the target area, the pipe segment characteristics of the pipe segment include pipe segment attributes, pipe segment upstream inspection well attributes, pipe segment downstream inspection well attributes and rainfall parameters.

[0059] In this embodiment, the attributes of each pipe section include the length of the pipe section, the cross-sectional area of ​​the pipe section, the starting elevation, the final elevation, the starting depth and the ending depth. The attributes of the upstream inspection well of the pipe section include the ground elevation of the upstream inspection well of the pipe section, the buried depth of the upstream inspection well of the pipe section, the catchment area of ​​the upstream inspection well of the pipe section, the average slope of the catchment area of ​​the upstream inspection well of the pipe section and the impermeability of the catchment area of ​​the upstream inspection well of the pipe section. The attributes of the downstream inspection well of the pipe section include the ground elevation of the downstream inspection well of the pipe section, the buried depth of the downstream inspection well of the pipe section, the catchment area of ​​the downstream inspection well of the pipe section, the average slope of the catchment area of ​​the downstream inspection well of the pipe section and the impermeability of the catchment area of ​​the downstream inspection well of the pipe section. The rainfall parameters include the recurrence period of heavy rain.

[0060] In some embodiments, after the pipe segment features of all the pipe segments are spliced ​​together, the numerical features need to be standardized.

[0061] In some embodiments, the graph data of the rainwater pipe network in the target area can be expressed as G=(T,H), where T is the attribute feature matrix of each pipe network node, which is obtained by splicing the pipe segment features of all pipe segments. H is the pipe network connection matrix used to represent the connection relationship between pipe networks in the rainwater pipe network of the target area.

[0062] In this embodiment, T can be expressed as: H can be expressed as:

[0063]

[0064] Among them, y i,j is the jth feature of the i-th pipe network in the target area rainwater pipe network, h is the total number of pipe segment features of each pipe segment, m is the total number of all target pipe segments in the rainwater pipe network in the area to be evaluated, and n 1,a is the starting point number of the first target pipe network, n 1,b The end point number of the first target pipe network.

[0065] After the rainwater pipe network carrying capacity assessment model is trained, the trained rainwater pipe network carrying capacity assessment model can be used to assess the carrying capacity of the rainwater pipe network.

[0066] The following is an introduction to the method for evaluating the hydraulic bearing capacity of a rainwater pipe network provided in an embodiment of the present invention.

[0067] See also Figure 2 , which shows a flow chart of the implementation of the method for evaluating the hydraulic bearing capacity of a rainwater pipe network provided by an embodiment of the present invention, and is described in detail as follows:

[0068] S210: Obtain rainwater pipe network information of the area to be evaluated, and convert the rainwater pipe network information into map data of the rainwater pipe network.

[0069] The rainwater pipe network information of the area to be evaluated includes the pipe segment characteristics of all target pipe segments and the connection relationship of all target pipe segments. The graph data of the rainwater pipe network includes the pipe network node attribute feature matrix and the pipe network connection matrix. In order to represent the pipe network information in a graph embedding manner, the pipe network is used as the node of the graph structure, and the public inspection wells between the pipe networks are used as the connection of the graph structure.

[0070] Among them, the pipe network node attribute feature matrix is ​​obtained by splicing the pipe segment features of all target pipe segments, and the pipe network connection matrix is ​​determined based on the connection relationship of all target pipe segments. It should also be noted that since the rainwater pipe network in the area to be evaluated includes many pipe segments, the target pipe segment here is any pipe segment in the rainwater pipe network.

[0071] In some embodiments, the pipeline segment characteristics of the target pipeline segment include pipeline segment attributes, pipeline segment upstream inspection well attributes, pipeline segment downstream inspection well attributes, and rainfall parameters.

[0072] In this embodiment, the pipe segment attributes include the pipe segment length, the pipe segment cross-sectional area, the starting elevation, the final elevation, the starting buried depth, and the ending buried depth. The upstream inspection well attributes of the pipe segment include the ground elevation of the upstream inspection well of the pipe segment, the buried depth of the upstream inspection well of the pipe segment, the catchment area of ​​the upstream inspection well of the pipe segment, the average slope of the catchment area of ​​the upstream inspection well of the pipe segment, and the water-impermeability of the catchment area of ​​the upstream inspection well of the pipe segment. The downstream inspection well attributes of the pipe segment include the ground elevation of the downstream inspection well of the pipe segment, the buried depth of the downstream inspection well of the pipe segment, the catchment area of ​​the downstream inspection well of the pipe segment, the average slope of the catchment area of ​​the downstream inspection well of the pipe segment, and the water-impermeability of the catchment area of ​​the downstream inspection well of the pipe segment. The rainfall parameters include the rainstorm recurrence period.

[0073] In some embodiments, after the pipe segment features of all target pipe segments are spliced, the numerical features need to be standardized.

[0074] In some embodiments, the graph data of the rainwater pipe network can be represented as G=(F,M), where F is the pipe network node attribute feature matrix, which is obtained by splicing the pipe segment features of all target pipe segments. M is the pipe network connection matrix used to represent the connection relationship between pipe networks in the rainwater pipe network of the area to be evaluated.

[0075] In this embodiment, F can be expressed as: M can be expressed as:

[0076]

[0077] Among them, x i,j is the jth feature of the i-th target pipe network, t is the total number of pipe segment features of the target pipe segment, r is the total number of all target pipe segments in the rainwater pipe network in the area to be evaluated, and n 1,s is the starting point number of the first target pipe network, n1,e The end point number of the first target pipe network.

[0078] By adopting the graph embedding method, the pipe segment features of all target pipe segments and the connection relationship of all target pipe segments are mapped to the low-dimensional vector space and converted into the graph data of the rainwater pipe network. This can make full use of the rich topological relationship data of the rainwater pipe network and mine the deep-level features hidden in the pipe network structure, which often have an important impact on the carrying capacity of the pipe network. Therefore, the carrying capacity can be evaluated more accurately.

[0079] S220, inputting the graph data of the rainwater pipe network into a pre-trained rainwater pipe network bearing capacity assessment model, and outputting the maximum filling degree of all target pipe sections in the area to be assessed.

[0080] The stormwater pipe network bearing capacity assessment model is built based on the graph neural network model.

[0081] The evaluation method provided by the present invention can obtain the maximum fullness of all target pipe sections in the area to be evaluated by inputting the graph data of the rainwater pipe network into a pre-trained rainwater pipe network bearing capacity evaluation model, so that the maximum rainstorm recurrence period of all target pipe sections can be determined according to the maximum fullness of all target pipe sections, thereby completing the hydraulic bearing capacity evaluation of the rainwater pipe network in the area to be evaluated.

[0082] The existing mechanism model for evaluating hydraulic carrying capacity takes several weeks to build, and it takes about 10 minutes to conduct hydraulic carrying capacity assessment. Since the data in different regions are different, the mechanism model needs to be rebuilt for the evaluation of different regions. The constructed mechanism model does not have wide applicability.

[0083] However, the rainwater pipe network bearing capacity assessment model provided by the present invention is constructed based on the graph convolution layer, and the model training takes only 1 hour. When conducting a large-scale rainwater pipe network hydraulic bearing capacity assessment, it only takes about 3 seconds, and the efficiency is significantly improved. In addition, the rainwater pipe network bearing capacity assessment model constructed by the present invention can be applied to the hydraulic bearing capacity assessment of any region. It only needs to obtain the rainwater pipe network information of the assessment area, and convert the rainwater pipe network information into the graph data of the rainwater pipe network. Finally, the graph data of the rainwater pipe network is input into the pre-trained rainwater pipe network bearing capacity assessment model, and the maximum fullness of all target pipe sections in the area to be assessed can be obtained without rebuilding a new model.

[0084] In an embodiment of the present invention, by converting the acquired rainwater pipe network information into the graph data of the rainwater pipe network, and inputting the obtained graph data of the rainwater pipe network into the pre-trained rainwater pipe network bearing capacity assessment model, the maximum fullness of all target pipe sections in the area to be assessed can be output. The rainwater pipe network bearing capacity assessment model in the present invention is constructed based on the graph neural network model. By converting the rainwater pipe network information into the graph data of the rainwater pipe network, the topological structure and attribute information of the rainwater pipe network can be effectively mapped to the low-dimensional vector space. Thereby, the rich topological relationship data of the drainage pipe network can be fully utilized to dig out the deep-level features hidden in the pipe network structure, and these features play a very important role in the evaluation accuracy of the pipe network's bearing capacity. Compared with the existing mechanism model, the assessment method provided by the present invention is more flexible in data processing and can integrate multi-source data more quickly. In addition, once the rainwater pipe network bearing capacity assessment model provided by the present invention is trained, the calculation speed is extremely fast when conducting the assessment, and it can provide strong support for the planning, management and emergency decision-making of the urban drainage pipe network in a short time, effectively making up for the shortcomings of the existing assessment method. Moreover, the rainwater pipe network bearing capacity assessment model provided by the present invention can be widely used once training is completed. When conducting hydraulic bearing capacity assessment, it is only necessary to obtain the rainwater pipe network information without rebuilding a new model, thereby further speeding up the assessment.

[0085] For ease of understanding, Figure 3 As shown, a specific embodiment will be used as an example to illustrate the invention, including two major steps: constructing a rainwater pipe network bearing capacity assessment model and using the model for assessment.

[0086] S310. Build a rainwater pipe network bearing capacity assessment model.

[0087] The initial stormwater pipe network carrying capacity assessment model includes a five-layer graph neural network, including a sequentially connected input layer, a first hidden layer, a second hidden layer, a third hidden layer, and an output layer. The first hidden layer and the second hidden layer both include a sequentially connected graph convolution layer (GCN layer), a batch normalization layer (BatchNorm layer), an activation function layer (ReLU layer), and a random dropout layer (DropOut layer), and the third hidden layer includes a sequentially connected graph convolution layer (GCN layer) and a fully connected layer (Linear layer). Among them, the main structure of the first hidden layer and the second hidden layer are both graph convolution layers (GCN layers), and the first two convolution layers use the BatchNorm layer for normalization, the ReLU layer for activation, and the DropOut layer to prevent overfitting. The data output by the output layer is the maximum fullness of all target pipe sections. The number of neurons in the first hidden layer, the second hidden layer, and the third hidden layer are set to 32, 32, and 16, respectively, and the neuron dropout rate of the Dropout layer is set to 0.5.

[0088] S3110. Build a SWMM model.

[0089] See also Figure 4 and Figure 5 , a rainwater pipe network in a small area of ​​Shenzhen (including 2015 pipe sections) was selected, and the CAD data of the rainwater pipe network structure information including pipe length, pipe diameter, pipe shape, pipe top and pipe bottom elevation, pipe top and pipe bottom burial depth, inspection well burial depth and ground elevation, as well as the raster data of the underlying surface property parameters including land use type, DEM, etc. were obtained and imported into ArcGIS software. Using ArcGIS software, the watershed subdivision (i.e., Subcatchment layer) was divided based on the Thiessen polygon method and manual adjustment method with the inspection well as the inspection well as the center; at the same time, the average slope, impermeability, watershed area and other underlying surface parameters of each watershed subdivision were calculated based on DEM and land use type data. Python programming was used to read the processed pipeline layer (CONDUITS), inspection well layer (NODES) and watershed subdivision layer (SUBCATCHMENTS) vector files, and sorted into inp files according to the format of SWMM model input files. The outlet flow data and corresponding rainfall data under ten rainfall conditions were collected at the outlet of the area. The main water volume parameters in the SWMM model, including the Manning's roughness coefficient of the impervious area, the Manning's roughness coefficient of the permeable area, the water storage depth of the depression in the impervious area, the water storage depth of the depression in the permeable area, the percentage of impervious area without depression storage, the maximum infiltration rate, the minimum infiltration rate, the attenuation rate constant, the drainage time, and the Manning's roughness coefficient of the pipe network were calibrated and verified. The Nash coefficient predicted by the model reached about 0.70.

[0090] S3120, design for heavy rain.

[0091] According to the rainstorm intensity formula The rainstorm intensity under different rainstorm return periods and rainfall durations was calculated, and then the Chicago rainfall pattern was used to calculate the rainstorm intensity. The rainstorm intensity is expanded over time to obtain the design rainstorm under different rainstorm return periods and rainfall durations. The design rainstorm is used as rainfall input, the SWMM model is run, and the maximum filling degree of each pipe segment is obtained from the running results. The maximum filling degree of each pipe segment can be used as the prediction result in the training sample set of the subsequent stormwater pipe network bearing capacity assessment model.

[0092] Among them, I is the rainstorm intensity, in mm / min; P is the rainstorm recurrence period, in yr; A, B, C, n are local parameters, dimensionless (the local parameters of Shenzhen are 8.701, 11.13, 0.594 and 0.555 respectively); t is the rainfall duration, in min. t is the rainstorm intensity at time t, in mm / min; t pis the time when the peak rainfall occurs, in minutes; t d is the total duration of rainfall, in minutes; r is the comprehensive rain peak position coefficient, dimensionless; a is A(1+ClgP).

[0093] Through the above calculation, we can get Figure 6 The design rainstorm in Shenzhen City shown here lasts for 2 hours.

[0094] S3130. Determine the map data of the rainwater pipe network in this small area of ​​Shenzhen City.

[0095] Get the rainwater pipe network information of 2015 pipe sections, including the pipe section characteristics of 2015 pipe sections and the connection relationship of 2015 pipe sections. Figure 7 As shown in the figure, the pipe segment characteristics include pipe segment attributes, pipe segment upstream inspection well attributes, pipe segment downstream inspection well attributes and rainfall parameters. The pipe segment attributes include pipe segment length, pipe segment cross-sectional area, starting bottom elevation, final bottom elevation, starting point buried depth and final point buried depth. The pipe segment upstream inspection well attributes include the ground elevation of the pipe segment upstream inspection well, the buried depth of the pipe segment upstream inspection well, the catchment area of ​​the pipe segment upstream inspection well, the average slope of the catchment area of ​​the pipe segment upstream inspection well and the impermeability of the catchment area of ​​the pipe segment upstream inspection well. The pipe segment downstream inspection well attributes include the ground elevation of the pipe segment downstream inspection well, the buried depth of the pipe segment downstream inspection well, the catchment area of ​​the pipe segment downstream inspection well, the average slope of the catchment area of ​​the pipe segment downstream inspection well and the impermeability of the catchment area of ​​the pipe segment downstream inspection well. The rainfall parameters include the rainstorm recurrence period. Therefore, the pipe segment characteristics include 17 features in total. The connection relationship of 2015 pipe segments is determined based on whether there is a public inspection well.

[0096] The 17 characteristics of 2015 pipe segments were based on The splicing is performed to obtain the pipe network node attribute feature matrix and the pipe network connection matrix, thereby obtaining the graph data G = (F, M) of the rainwater pipe network.

[0097] The pipe network connection matrix is ​​composed of the connection relationship between pipe networks. The two rows of the pipe network connection matrix represent a pipe section and another pipe section connected to it, that is, the upstream pipe section number and the downstream pipe section number of the public inspection well.

[0098] It should be noted that after splicing, the numerical features need to be standardized.

[0099] S3140. Train the constructed stormwater pipe network bearing capacity assessment model.

[0100] The graph data of the rainwater pipe network represented by the graph embedding method, namely the pipe network node attribute feature matrix G and the pipe network connection matrix M, as well as the maximum fullness of each pipe segment, are used to form a sample data set to train the rainwater pipe network bearing capacity assessment model. Among them, the input dimension of the pipe network node attribute feature matrix G is (2015 pipe segments, 17 features), the input dimension of the pipe network connection matrix M is (2, the connection relationship of all target pipe segments), and the output of the rainwater pipe network bearing capacity assessment model is (2015 pipe segments, 1).

[0101] The sample data set is randomly divided into training set, validation set and test set according to the 8:1:1 principle, and the mean square error loss (MSELoss) is used as the loss function, such as Figure 8 As shown in the figure, Adam is used as the optimizer, the learning rate is set to 0.001, the weight decay is set to 0.0005, and the network parameters are continuously optimized through back propagation. The loss values ​​of the training set and the validation set are monitored during the training process. When the loss values ​​converge, the learning is stopped and the model is obtained. The mean square error MSE and the coefficient of determination R of the test set are calculated. 2 After evaluation, the MSE value of the trained rainwater pipe network bearing capacity assessment model was 0.0235, R 2 is 0.75.

[0102] In order to demonstrate the evaluation effect of the rainwater pipe network bearing capacity evaluation model provided by the present invention, the following Fig. 9 The shallow neural network shown is compared with the rainwater pipe network bearing capacity assessment model of this application. The shallow neural network replaces the graph convolution layer in this application with the linear layer Linear. The shallow neural network is trained with the same sample data set to obtain a trained shallow neural network. The evaluation results of the trained rainwater pipe network bearing capacity assessment model and the shallow neural network are compared to obtain the evaluation results in Table 1.

[0103] Table 1 Comparison of evaluation results of different machine learning algorithms

[0104] algorithm MSE <![CDATA[R 2 ]]> Shallow Neural Networks 0.0756 0.23 Stormwater pipe network bearing capacity assessment model 0.0235 0.75

[0105] From the data in the above table, we can see that the MSE value of the trained stormwater pipe network bearing capacity assessment model is about 69% lower than that of the shallow neural network, and R 2 Improved by about 2.26 times.

[0106] The rainwater pipe network bearing capacity assessment model provided by the present invention is used to assess and predict the graph data of the rainwater pipe network in the small area of ​​Shenzhen City, and the following is obtained: Fig.10 The prediction results are shown.

[0107] S320: Use the trained stormwater pipe network bearing capacity assessment model to assess the hydraulic bearing capacity of the large-scale stormwater pipe network to be assessed.

[0108] The rainwater pipe network information of the large-scale rainwater pipe network to be evaluated is obtained, and the rainwater pipe network information is converted into the graph data of the rainwater pipe network. The conversion method of the rainwater pipe network information and the graph data has been discussed previously and will not be repeated here. The graph data of the rainwater pipe network is input into the pre-trained rainwater pipe network bearing capacity assessment model, and the maximum fullness of all target pipe sections in the area to be evaluated is output.

[0109] Based on the obtained maximum filling degrees of all target pipe sections in the area to be evaluated, the hydraulic bearing capacity assessment of the rainwater pipe network in the area to be evaluated can be completed.

[0110] It should be understood that the order of execution of the steps in the above embodiment does not necessarily mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiment of the present invention.

[0111] The following is an embodiment of the device of the present invention. For details not described in detail therein, reference may be made to the corresponding method embodiment described above.

[0112] Fig.11 The structural schematic diagram of the rainwater pipe network hydraulic bearing capacity assessment device provided by an embodiment of the present invention is shown. For the convenience of explanation, only the parts related to the embodiment of the present invention are shown, which are described in detail as follows:

[0113] like Fig.11 As shown, the rainwater pipe network hydraulic bearing capacity assessment device 1100 includes:

[0114] The conversion module 1110 is used to obtain the rainwater pipe network information of the area to be evaluated, and convert the rainwater pipe network information into the graph data of the rainwater pipe network; wherein the rainwater pipe network information includes the pipe segment characteristics of all target pipe segments and the connection relationship of all target pipe segments, and the graph data of the rainwater pipe network includes the pipe network node attribute feature matrix and the pipe network connection matrix, the pipe network node attribute feature matrix is ​​obtained by splicing the pipe segment characteristics of all target pipe segments, and the pipe network connection matrix is ​​determined based on the connection relationship of all target pipe segments; the target pipe segment is any pipe segment in the rainwater pipe network;

[0115] The evaluation module 1120 is used to input the graph data of the rainwater pipe network into a pre-trained rainwater pipe network bearing capacity evaluation model, and output the maximum fullness of all target pipe sections in the area to be evaluated; wherein the rainwater pipe network bearing capacity evaluation model is constructed based on a graph neural network model.

[0116] In a possible implementation, the pipe segment characteristics of the target pipe segment include pipe segment attributes, pipe segment upstream inspection well attributes, pipe segment downstream inspection well attributes, and rainfall parameters;

[0117] The connection relationship of all target pipe sections is determined based on whether they have a public inspection well.

[0118] In a possible implementation, the pipe segment attributes include the pipe segment length, the pipe segment cross-sectional area, the starting elevation, the ending elevation, the starting depth, and the ending depth;

[0119] The properties of the upstream inspection well of the pipe section include the ground elevation of the upstream inspection well of the pipe section, the buried depth of the upstream inspection well of the pipe section, the catchment area of ​​the upstream inspection well of the pipe section, the average slope of the catchment area of ​​the upstream inspection well of the pipe section, and the impermeability of the catchment area of ​​the upstream inspection well of the pipe section;

[0120] The properties of the downstream inspection well of the pipe section include the ground elevation of the downstream inspection well of the pipe section, the buried depth of the downstream inspection well of the pipe section, the catchment area of ​​the downstream inspection well of the pipe section, the average slope of the catchment area of ​​the downstream inspection well of the pipe section, and the impermeability of the catchment area of ​​the downstream inspection well of the pipe section;

[0121] Rainfall parameters include rainstorm return period.

[0122] In one possible implementation, a sample data set for training a stormwater pipe network bearing capacity assessment model is constructed based on the stormwater pipe network in a target area. The sample data set includes graph data of the stormwater pipe network in the target area and the maximum fullness of each pipe section in the stormwater pipe network in the target area. The graph data of the stormwater pipe network in the target area is constructed based on the stormwater pipe network information in the target area.

[0123] In one possible implementation, the maximum filling degree of each pipe segment in the stormwater pipe network within the target area is determined based on inputting the design storm into a pre-built stormwater flood management model;

[0124] The stormwater pipe network bearing capacity assessment model includes a sequentially connected input layer, a first hidden layer, a second hidden layer, a third hidden layer and an output layer. The first hidden layer, the second hidden layer and the third hidden layer all include graph convolution layers. The first hidden layer and the second hidden layer both include sequentially connected graph convolution layers, batch normalization layers, activation function layers and random inactivation layers. The third hidden layer includes sequentially connected graph convolution layers and fully connected layers.

[0125] In one possible implementation, the stormwater flood management model is constructed based on the structural information of the stormwater pipe network and the underlying surface parameters in the target area;

[0126] The structural information of the rainwater pipe network includes pipe length, pipe diameter, pipe shape, pipe top and bottom elevation, pipe top and bottom burial depth, inspection well burial depth and ground elevation;

[0127] The underlying surface parameters include the average slope, imperviousness and catchment area of ​​each catchment area.

[0128] Fig.12 Schematic diagram of an electronic device provided by an embodiment of the present invention. Fig.12 As shown, the electronic device 12 of this embodiment includes: a processor 120 and a memory 121. The memory 121 stores a computer program 122. When the processor 120 executes the computer program 122, the steps in the above-mentioned method embodiments are implemented. Alternatively, when the processor 120 executes the computer program 122, the functions of each module / unit in the above-mentioned device embodiments are implemented.

[0129] Exemplarily, the computer program 122 may be divided into one or more modules / units, which are stored in the memory 121 and executed by the processor 120 to implement the present invention. The one or more modules / units may be a series of computer program instruction segments capable of implementing specific functions, which are used to describe the execution process of the computer program 122 in the electronic device 12.

[0130] The electronic device 12 may include, but is not limited to, a processor 120 and a memory 121. Those skilled in the art will appreciate that Fig.12 It is only an example of the electronic device 12 and does not constitute a limitation of the electronic device 12. It may include more or fewer components than shown in the figure, or a combination of certain components, or different components. For example, the electronic device 12 may also include input and output devices, network access devices, buses, etc.

[0131] The processor 120 may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or the processor may be any conventional processor, etc.

[0132] The memory 121 may be an internal storage unit of the electronic device 12, such as a hard disk or memory of the electronic device 12. The memory 121 may also be an external storage device of the electronic device 12, such as a plug-in hard disk, a smart media card (SMC), a secure digital (SD) card, a flash card, etc. equipped on the electronic device 12. Further, the memory 121 may also include both an internal storage unit of the electronic device 12 and an external storage device. The memory 121 is used to store the computer program 122 and other programs and data required by the electronic device 12. The memory 121 may also be used to temporarily store data that has been output or is to be output.

[0133] For the convenience and simplicity of description, only the division of the above functional modules / units is used as an example for illustration. In actual applications, the above functions can be assigned to different functional modules / units as needed. The above modules / units can be implemented in the form of hardware, software, or a combination of hardware and software.

[0134] The embodiment of the present invention further provides a computer-readable storage medium, which stores a computer program. When the computer program is executed by a processor, the methods in the above method embodiments are implemented.

[0135] The embodiment of the present invention further provides a computer program product, including a computer program. When the computer program is executed by a processor, the methods in the above method embodiments are implemented.

[0136] The computer program includes computer program code, which may be in source code form, object code form, executable file or some intermediate form, etc. Computer readable media may include: any entity or device capable of carrying computer program code, recording medium, USB flash drive, mobile hard disk, magnetic disk, optical disk, computer memory, read-only memory (ROM), random access memory (RAM), electric carrier signal, telecommunication signal and software distribution medium, etc.

[0137] In the above embodiments, the description of each embodiment has its own emphasis. For parts that are not described or recorded in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments. If there is no special explanation and logical conflict, the terms and / or descriptions between different embodiments are consistent and can be referenced to each other. The technical features in different embodiments can be combined to form a new embodiment according to their internal logical relationship.

[0138] The embodiments described above are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. Such modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be included in the protection scope of the present invention.

Claims

1. A method for evaluating the hydraulic bearing capacity of a rainwater pipe network, characterized in that: include: Acquire rainwater pipe network information of the area to be evaluated, and convert the rainwater pipe network information into graph data of the rainwater pipe network; wherein the rainwater pipe network information includes pipe segment features of all target pipe segments and connection relationships of all target pipe segments, and the graph data of the rainwater pipe network includes a pipe network node attribute feature matrix and a pipe network connection matrix, wherein the pipe network node attribute feature matrix is ​​obtained by splicing the pipe segment features of all target pipe segments, and the pipe network connection matrix is ​​determined based on the connection relationship of all target pipe segments; the target pipe segment is any pipe segment in the rainwater pipe network; The graph data of the rainwater pipe network is input into a pre-trained rainwater pipe network bearing capacity assessment model, and the maximum fullness of all target pipe sections in the area to be assessed is output; wherein the rainwater pipe network bearing capacity assessment model is constructed based on a graph neural network model.

2. The method for evaluating the hydraulic bearing capacity of a rainwater pipe network according to claim 1, characterized in that: The pipe segment characteristics of the target pipe segment include pipe segment attributes, pipe segment upstream inspection well attributes, pipe segment downstream inspection well attributes and rainfall parameters; The connection relationship of all target pipe sections is determined based on whether there is a public inspection well.

3. The method for assessing the hydraulic bearing capacity of a rainwater pipe network according to claim 2, characterized in that: The pipe segment attributes include pipe segment length, pipe segment cross-sectional area, starting elevation, ending elevation, starting depth and ending depth; The properties of the upstream inspection well of the pipe section include the ground elevation of the upstream inspection well of the pipe section, the buried depth of the upstream inspection well of the pipe section, the catchment area of ​​the upstream inspection well of the pipe section, the average slope of the catchment area of ​​the upstream inspection well of the pipe section, and the impermeability of the catchment area of ​​the upstream inspection well of the pipe section; The attributes of the inspection well downstream of the pipe section include the ground elevation of the inspection well downstream of the pipe section, the buried depth of the inspection well downstream of the pipe section, the catchment area of ​​the inspection well downstream of the pipe section, the average slope of the catchment area of ​​the inspection well downstream of the pipe section, and the impermeability of the catchment area of ​​the inspection well downstream of the pipe section; The rainfall parameters include the rainstorm return period.

4. The method for evaluating the hydraulic bearing capacity of a rainwater pipe network according to claim 1, characterized in that: The sample data set used for training the rainwater pipe network bearing capacity assessment model is constructed based on the rainwater pipe network in the target area. The sample data set includes the graph data of the rainwater pipe network in the target area and the maximum fullness of each pipe section in the rainwater pipe network in the target area. The graph data of the rainwater pipe network in the target area is constructed based on the rainwater pipe network information in the target area.

5. The method for assessing the hydraulic bearing capacity of a rainwater pipe network according to claim 4, characterized in that: The maximum filling degree of each pipe segment in the stormwater pipe network within the target area is determined based on inputting the design storm into a pre-built stormwater flood management model; The stormwater pipe network bearing capacity assessment model includes a sequentially connected input layer, a first hidden layer, a second hidden layer, a third hidden layer and an output layer, the first hidden layer, the second hidden layer and the third hidden layer all include graph convolution layers, the first hidden layer and the second hidden layer all include sequentially connected graph convolution layers, batch normalization layers, activation function layers and random inactivation layers, and the third hidden layer includes sequentially connected graph convolution layers and fully connected layers.

6. The method for assessing the hydraulic bearing capacity of a rainwater pipe network according to claim 5, characterized in that: The storm flood management model is constructed based on the structural information of the storm water pipe network and the underlying surface parameters of the target area; The structural information of the rainwater pipe network includes pipe length, pipe diameter, pipe shape, pipe top and pipe bottom elevation, pipe top and pipe bottom burial depth, inspection well burial depth and ground elevation; The underlying surface parameters include the average slope, impermeability and catchment area of ​​each catchment area.

7. A rainwater pipe network hydraulic bearing capacity assessment device, characterized in that: include: A conversion module, used to obtain rainwater pipe network information of the area to be evaluated, and convert the rainwater pipe network information into graph data of the rainwater pipe network; wherein the rainwater pipe network information includes the pipe segment characteristics of all target pipe segments and the connection relationship of all target pipe segments, and the graph data of the rainwater pipe network includes a pipe network node attribute feature matrix and a pipe network connection matrix, wherein the pipe network node attribute feature matrix is ​​obtained by splicing the pipe segment characteristics of all target pipe segments, and the pipe network connection matrix is ​​determined based on the connection relationship of all target pipe segments; the target pipe segment is any pipe segment in the rainwater pipe network; An evaluation module is used to input the graph data of the rainwater pipe network into a pre-trained rainwater pipe network bearing capacity evaluation model, and output the maximum fullness of all target pipe sections in the area to be evaluated; wherein the rainwater pipe network bearing capacity evaluation model is constructed based on a graph neural network model.

8. An electronic device, characterized in that: The method comprises a memory and a processor, wherein the memory stores a computer program, and the processor implements the method according to any one of claims 1 to 6 when executing the computer program.

9. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the method according to any one of claims 1 to 6 is implemented.

10. A computer program product, characterized in that The method comprises a computer program, which implements the method according to any one of claims 1 to 6 when being executed by a processor.

Images