Urban inland inundation model parameter rapid determination method and device based on deep learning

Abstract

The invention provides a deep learning-based urban inland inundation model parameter rapid determination method and device, and the method comprises the steps: constructing a rainstorm inland inundation model of each research region through a processor, and carrying out the calibration and verification of each rainstorm inland inundation model; then, constructing a deep learning model, and training the deep learning model through the model input data and the model parameters of the rainstorm waterlogging model; and inputting the input data of the target research area into the trained deep learning model to obtain model parameters of the rainstorm waterlogging model of the target research area. It can be seen that in the method, the deep learning model is applied to determination of the parameters of the urban rainstorm waterlogging model with a clear physical mechanism, the model parameters of the urban rainstorm waterlogging model can be rapidly and accurately determined, and therefore the accuracy and timeliness of early warning of the urban rainstorm waterlogging are improved.

Description

Technical Field

[0001] The present application relates to the field of machine learning technology, and in particular to a method and device for quickly determining parameters of an urban waterlogging model based on deep learning. Background Art

[0002] In terms of urban rainstorm waterlogging simulation, the physical mechanism model based on the traditional hydrological and hydrodynamic model has been widely used in waterlogging risk assessment. The physical mechanism model has a clear physical mechanism and can provide accurate simulation results. However, the model parameters in the physical mechanism model rely on manual debugging methods. Manual debugging of model parameters is time-consuming and labor-intensive, and is highly dependent on the experience and technology of the debugging personnel. The model simulation accuracy is affected, resulting in low timeliness and accuracy of model simulation, which cannot meet the timeliness requirements of urban rainstorm waterlogging forecasting and warning. How to overcome the time-consuming problem of physical mechanism model parameter calibration and verification, so as to improve the timeliness of its simulation and forecasting and warning, is a key issue that needs to be solved. Summary of the invention

[0003] The present application provides a method and device for quickly determining parameters of an urban waterlogging model based on deep learning. The deep learning model is applied to the determination of parameters of an urban rainstorm waterlogging model with a clear physical mechanism, so that the model parameters of the urban rainstorm waterlogging model can be quickly and accurately determined, thereby improving the accuracy and timeliness of urban rainstorm waterlogging warnings.

[0004] In a first aspect, the present application provides a method for quickly determining parameters of an urban waterlogging model based on deep learning, the method comprising: obtaining underlying surface data, pipe network data, rainfall data and measured hydrological data of multiple study areas; constructing a rainstorm waterlogging model for each study area based on the underlying surface data, the pipe network data and the rainfall data; calibrating and verifying the corresponding rainstorm waterlogging model based on the rainfall data and measured hydrological data of each study area to obtain a model parameter database, wherein the model parameter database includes optimized model parameters of each rainstorm waterlogging model; constructing a deep learning model, and constructing an input data database based on the underlying surface data, pipe network data and rainfall data of each study area; training the deep learning model using the input data database as an independent variable and the model parameter database as a dependent variable; inputting the underlying surface data, pipe network data and rainfall data of the target study area into the trained deep learning model to obtain model parameters of the rainstorm waterlogging model of the target study area.

[0005] In a second aspect, the present application provides a device for quickly determining parameters of an urban waterlogging model based on deep learning, comprising: an acquisition unit for acquiring underlying surface data, pipe network data, rainfall data and measured hydrological data of multiple study areas; a processing unit for constructing a rainstorm waterlogging model for each study area based on the underlying surface data, the pipe network data and the rainfall data; and, calibrating and verifying the corresponding rainstorm waterlogging model based on the rainfall data and measured hydrological data of each study area to obtain a model parameter database, wherein the model parameter database includes optimized model parameters of each rainstorm waterlogging model; and, constructing a deep learning model, constructing an input data database based on the underlying surface data, pipe network data and rainfall data of each study area; and, training the deep learning model using the input data database as an independent variable and the model parameter database as a dependent variable; and, inputting the underlying surface data, pipe network data and rainfall data of the target study area into the trained deep learning model to obtain model parameters of the rainstorm waterlogging model of the target study area.

[0006] In a third aspect, the present application provides an electronic device, including a processor and a memory, wherein the memory stores a computer program, and when the processor calls the computer program in the memory, it executes the method as described in any one of the first aspects.

[0007] It can be seen that in the embodiment of the present application, the processor obtains the underlying surface data, pipeline network data, rainfall data and measured hydrological data of multiple study areas, and constructs a rainstorm waterlogging model for each study area based on the underlying surface data, pipeline network data and rainfall data; calibrates and verifies the corresponding rainstorm waterlogging model based on the rainfall data and measured hydrological data of each study area to obtain a model parameter database, which includes the optimized model parameters of each rainstorm waterlogging model; constructs a deep learning model, and constructs an input data database based on the underlying surface data, pipeline network data and rainfall data of each study area; trains the deep learning model with the input data database as the independent variable and the model parameter database as the dependent variable; inputs the underlying surface data, pipeline network data and rainfall data of the target study area into the trained deep learning model to obtain the model parameters of the rainstorm waterlogging model of the target study area. It can be seen that in this application, the deep learning model is applied to the parameter determination of the urban rainstorm waterlogging model with a clear physical mechanism. Compared with the traditional manual debugging method, the model parameters of the urban rainstorm waterlogging model can be quickly and accurately determined, thereby improving the accuracy and timeliness of urban rainstorm waterlogging warnings. BRIEF DESCRIPTION OF THE DRAWINGS

[0008] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0009] Figure 1 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application;

[0010] Figure 2 A flowchart of a method for quickly determining parameters of an urban waterlogging model based on deep learning provided in an embodiment of the present application;

[0011] Figure 3 A schematic diagram showing a comparison between measured hydrological data and simulated hydrological data of a first rainstorm waterlogging model in a first study area provided in an embodiment of the present application;

[0012] Figure 4 A block diagram of functional units of a device for quickly determining parameters of an urban waterlogging model based on deep learning provided in an embodiment of the present application. DETAILED DESCRIPTION

[0013] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0014] The terms "first", "second", etc. in the specification and claims of this application and the above-mentioned drawings are used to distinguish different objects, rather than to describe a specific order. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions. For example, a process, method, system, product or device that includes a series of steps or units is not limited to the listed steps or units, but optionally includes steps or units that are not listed, or optionally includes other steps or units inherent to these processes, methods, products or devices.

[0015] Reference to "embodiments" herein means that a particular feature, structure, or characteristic described in conjunction with the embodiments may be included in at least one embodiment of the present application. The appearance of the phrase in various locations in the specification does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment that is mutually exclusive with other embodiments. It is explicitly and implicitly understood by those skilled in the art that the embodiments described herein may be combined with other embodiments.

[0016] In the embodiments of the present application, "and / or" describes the association relationship of the associated objects, indicating that three relationships may exist. For example, A and / or B can represent the following three situations: A exists alone; A and B exist at the same time; B exists alone. Among them, A and B can be singular or plural.

[0017] In the embodiment of the present application, the symbol " / " can indicate that the objects associated with each other are in an "or" relationship. In addition, the symbol " / " can also indicate a division sign, that is, performing a division operation. For example, A / B can indicate A divided by B.

[0018] In the embodiments of the present application, "at least one item" or similar expressions refer to any combination of these items, including any combination of single items or plural items, and refer to one or more, and multiple refers to two or more. For example, at least one item of a, b, or c can represent the following seven situations: a, b, c, a and b, a and c, b and c, a, b, and c. Among them, each of a, b, and c can be an element or a set containing one or more elements.

[0019] In the embodiments of the present application, "equal to" can be used in conjunction with greater than, and is applicable to the technical solution adopted when greater than, and can also be used in conjunction with less than, and is applicable to the technical solution adopted when less than. When equal to is used in conjunction with greater than, it is not used in conjunction with less than; when equal to is used in conjunction with less than, it is not used in conjunction with greater than.

[0020] In terms of urban rainstorm waterlogging simulation, it mainly includes physical mechanism models based on hydrodynamics and artificial intelligence technologies (such as machine learning and deep learning). The physical mechanism model has a clear physical mechanism and can provide accurate simulation results, but the calibration and verification of model parameters is time-consuming and cannot meet the timeliness requirements of urban rainstorm waterlogging prediction and warning; artificial intelligence technology has powerful data processing capabilities and high computing efficiency, but does not have a physical mechanism.

[0021] One solution is to combine artificial intelligence technology with physical mechanism models, and apply artificial intelligence technology to the determination of model parameters of the physical mechanism model of urban rainstorm waterlogging, so as to achieve rapid and accurate simulation of urban rainstorm waterlogging.

[0022] Based on the above ideas, the present application provides a method and device for quickly determining the parameters of an urban waterlogging model based on deep learning. The deep learning model is applied to the determination of the parameters of an urban rainstorm waterlogging model with a clear physical mechanism, which can realize the rapid and accurate determination of the model parameters of the urban rainstorm waterlogging model, thereby improving the accuracy and timeliness of urban rainstorm waterlogging warnings.

[0023] The technical solution of the present application and how the technical solution of the present application solves the above-mentioned technical problems are described in detail below with specific embodiments. The following specific embodiments can be combined with each other, and the same or similar concepts or processes may not be repeated in some embodiments. The embodiments of the present application will be described below in conjunction with the accompanying drawings.

[0024] See also Figure 1 , Figure 1 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present application. For the sake of convenience, only the parts related to the embodiment of the present application are shown. For specific technical details not disclosed, please refer to the method part of the embodiment of the present application. The electronic device 1 can be any terminal device including a mobile phone, a computer, a PDA, a POS, a car computer, etc. Figure 1 As shown, the electronic device 1 includes a memory 20, a processor 10, a communication bus 40, a communication interface 30 and one or more programs 21, the one or more programs 21 are stored in the memory 20 and are configured to be executed by the processor 10, and the one or more programs 21 include instructions for executing any step in the following method embodiments. In a specific implementation, the processor 10 is used to execute any step in the following method embodiments, and when performing data transmission such as sending, the communication interface 30 can be selectively called to complete the corresponding operation.

[0025] In the embodiment of the present application, the processor 10 included in the computer device may have the function corresponding to any method step in the embodiment.

[0026] Those skilled in the art will understand that Figure 1 The computer device structure shown in the figure does not constitute a limitation on the computer device, and may include more or less components than shown in the figure, or combine certain components, or arrange the components differently.

[0027] See also Figure 2 , Figure 2 A flowchart of a method for quickly determining parameters of an urban waterlogging model based on deep learning provided in an embodiment of the present application is provided. The method is applied to Figure 1 The processor 10 shown, the method includes:

[0028] Step S201, the processor obtains underlying surface data, pipe network data, rainfall data and measured hydrological data of multiple study areas.

[0029] In some embodiments, the underlying surface data includes terrain data, elevation data and land use type, and the land use type includes roads, buildings and green spaces. The pipeline network data includes pipeline data, manhole data and the connection mode between pipelines and manholes. The pipeline data includes the spatial distribution of the pipeline network, the length of the pipeline, the shape of the pipeline cross section, the pipe diameter, the upstream and downstream bottom elevations, and the manhole data includes the depth and size of the manhole; the rainfall data includes the rainfall of the measured field; the measured hydrological data includes the runoff data of the measured field of the pipeline, and the monitoring water level of the manhole and the waterlogging point during the measured rainfall; wherein, the rainfall data and the measured hydrological data of the same study area have a corresponding relationship, and the rainfall data and the measured hydrological data of the same study area include data of at least one measured field, and the start time and test duration of the same measured field are the same.

[0030] Among them, the underlying surface data can be obtained by obtaining high-resolution remote sensing images based on satellite remote sensing technology, and then the terrain data, elevation data and land use type can be extracted using software such as InfoWorks ICM and GIS. In addition, the underlying surface data, pipe network data, rainfall data and measured hydrological data of each study area in this embodiment can also be obtained through the local water affairs department. The underlying surface data and pipe network data of each study area need to be restricted to the same coordinate system.

[0031] Among them, the rainfall data and measured hydrological data of each study area are a collection of test data of multiple measured sessions, and each rainfall data of a measured session corresponds to the measured hydrological data of a measured session. For example, the rainfall data and measured hydrological data of the first study area include rainfall data and measured hydrological data of 100 sessions in 12 months of 2023. The start time of rainfall data and measured hydrological data of the same session is the same as the test time. For example, the start time of rainfall data of the first measured session is 8:00 on January 5, and the test duration is 1 hour. Then the start time of measured hydrological data of the first measured session is 8:00 on January 5, and the test duration is 1 hour.

[0032] Step S202: The processor constructs a rainstorm waterlogging model for each study area based on the underlying surface data, the pipe network data, and the rainfall data.

[0033] In some embodiments, constructing a rainstorm waterlogging model for each study area based on the underlying surface data, the pipe network data, and the rainfall data includes: inputting the underlying surface data, the pipe network data, and the rainfall data into InfoWorks ICM software to obtain a rainstorm waterlogging model for each study area.

[0034] After the underlying surface data, pipe network data, and rainfall data are input into InfoWorks ICM, InfoWorks ICM first constructs a one-dimensional drainage pipe network model, calculates the water volume process gathered in the catchment area through the one-dimensional drainage pipe network model, and simulates the water flow process in the drainage pipe network. The specific calculation process includes the following formula:

[0035]

[0036] Among them, A is the cross-sectional area of ​​the pipe; t is time; Q is flow rate; x is the length of the pipe along the direction of water flow; g is the acceleration of gravity; h is the water depth; S0 is the slope of the pipe bottom; and K is the flow resistance coefficient.

[0037] Then, a two-dimensional urban land surface hydrodynamic model is constructed and coupled with a one-dimensional drainage network model to simulate the complex water exchange process between the network and the surface. The formulas are:

[0038]

[0039] Where h is the water depth; u and v are the velocity components in the x and y directions respectively; S 0，x , S 0，y are the bottom slope components in the x and y directions respectively; S f，x , S f，y are the friction components in the x and y directions respectively; q 1D is the outflow per unit area; u 1D 、v 1D They are q 1D The velocity components in the x and y directions.

[0040] Step S203: the processor calibrates and verifies the corresponding rainstorm waterlogging model according to the rainfall data and measured hydrological data of each study area to obtain a model parameter database.

[0041] The model parameter database includes optimized model parameters of each rainstorm waterlogging model.

[0042] The model parameters include but are not limited to: fixed runoff coefficients of various underlying surfaces, initial loss values, initial infiltration rates, stable infiltration rates, attenuation coefficients, as well as pipe Manning's coefficients, surface Manning's coefficients, etc.

[0043] The fixed runoff coefficient refers to the proportion of rainfall converted to runoff on different land types. For example, the runoff coefficient of urban roads may be high (usually between 0.7 and 0.9), while the runoff coefficient of green spaces may be low (possibly 0.1 to 0.3).

[0044] Initial loss refers to the amount of water absorbed by the underlying surface at the beginning of rainfall, and is usually used to describe the water loss at the beginning of rainfall. For example, the initial loss value of soil may be large, indicating that the soil can absorb a certain amount of water, while the initial loss value of hardened surfaces such as roads is small.

[0045] Initial infiltration rate refers to the rate at which water penetrates into the soil at the beginning of rainfall. This value depends on factors such as soil type, moisture, and rainfall intensity. Usually, the initial infiltration rate is higher because the surface layer of the soil is loose and can absorb water faster.

[0046] The stable infiltration rate refers to the rate at which water penetrates into the soil after a period of rainfall. As soil moisture accumulates, the infiltration rate gradually slows down and enters a stable state. The stable infiltration rate is lower than the initial infiltration rate.

[0047] The attenuation coefficient is used to describe the change in infiltration rate over time during rainfall, and is usually related to factors such as soil type and soil moisture. The larger the attenuation coefficient, the faster the infiltration rate decreases.

[0048] The Manning coefficient of a pipe is used to describe the friction characteristics of water flow in a pipe, which affects the water flow velocity and flow rate in the pipe. The Manning coefficient is related to factors such as the material, roughness, and shape of the pipe. For example, a smooth pipe (such as a steel pipe) has a smaller Manning coefficient, while a rougher pipe (such as a brick pipe) has a larger Manning coefficient. The larger the Manning coefficient, the greater the resistance to the water flow and the lower the flow rate.

[0049] The surface Manning coefficient is used to describe the friction characteristics of the ground surface, which mainly affects the speed of surface runoff and the distribution of water flow. In rainstorm simulation, when the surface Manning coefficient is large, it means that the friction of the surface is large and the water flow is slow; while a small Manning coefficient means that the water flow is fast.

[0050] In some embodiments, the corresponding rainstorm waterlogging model is calibrated and verified according to the rainfall data and measured hydrological data of each study area to obtain a model parameter database, including:

[0051] Step a, for a single study area, inputting the rainfall data of at least one measured event into a rainstorm waterlogging model to obtain at least one simulated hydrological data;

[0052] Step b, calculating the simulation accuracy of the rainstorm waterlogging model according to the at least one simulated hydrological data and the at least one measured hydrological data;

[0053] Step c, if the simulation accuracy meets the preset simulation accuracy, the model parameters of the rainstorm waterlogging model are determined as target model parameters and written into the model parameter database;

[0054] Step d: If the simulation accuracy does not meet the preset simulation accuracy, the model parameters of the rainstorm waterlogging model are adjusted, and step a "inputting the rainfall data of at least one measured event into the rainstorm waterlogging model to obtain at least one simulated hydrological data" is re-executed.

[0055] The simulation accuracy includes two indicators: Nash efficiency coefficient (NSE) and relative error. In some embodiments, the simulation accuracy includes Nash efficiency coefficient, and the calculation formula of Nash efficiency coefficient is as follows:

[0056]

[0057] Among them, NSE is the Nash efficiency coefficient, Yi is the measured hydrological data of the ith measured field, is the simulated hydrological data of the ith measured event, is the average value of the measured hydrological data of all measured sessions, n is the total number of measured sessions, and the value range of NSE is (-∞,1]. The closer NSE is to 1, the higher the accuracy.

[0058] Among them, the relative error (ε) refers to the ratio of the absolute error to the measured hydrological data, and the absolute error (E) refers to the absolute value of the difference between the measured hydrological data and the simulated hydrological data. The calculation formula is ε=E / measured hydrological data=|measured hydrological data-simulated hydrological data| / measured hydrological data.

[0059] Among them, the Nash efficiency coefficient is higher than 0.60, and the absolute value of the relative error is lower than 10%.

[0060] For this embodiment, illustratively, the first study area includes two measured rainfall information as shown in Table 1 below:

[0061] Table 1 is the two measured rainfall information of the first study area provided in the embodiment of the present application

[0062]

[0063] The measured rainfall data of the first rainstorm waterlogging model in the first study area is input to obtain the first hydrological simulation data, see Figure 3 (a); input the measured rainfall data of the second rainstorm waterlogging model into the first rainstorm waterlogging model to obtain the second hydrological simulation data, see Figure 3 (b)

[0064] The relative error and Nash efficiency coefficient are calculated based on the first hydrological simulation data, the second hydrological simulation data, the first measured hydrological data, and the second measured hydrological data. If the relative error is less than 10% and the Nash efficiency coefficient is higher than 0.6, it indicates that the model parameter 1 used in the current first rainstorm waterlogging model meets the requirements of simulation accuracy, and the simulation parameter 1 can be directly determined as the target simulation parameter of the first rainstorm waterlogging model. If the relative error is higher than 10%, and / or the Nash efficiency coefficient is lower than 0.6, it indicates that the model parameter 1 used in the current first rainstorm waterlogging model does not meet the requirements of simulation accuracy, then the model parameter 1 used in the first rainstorm waterlogging model is adjusted to model parameter 2, and the rainfall data of the measured session 1 is input into the adjusted first rainstorm waterlogging model again to obtain the third hydrological simulation data, and the rainfall data of the measured session 2 is input into the adjusted first rainstorm waterlogging model to obtain the fourth hydrological simulation data, and the simulation accuracy is calculated based on the third hydrological simulation data, the fourth hydrological simulation data, the first measured hydrological data, and the second measured hydrological data until the simulation accuracy meets the requirements of the preset simulation accuracy.

[0065] Step S204: the processor builds a deep learning model.

[0066] In some embodiments, constructing a deep learning model includes: using the Keras library in the Python language to construct a convolutional neural network (CNN) model, the CNN model including an input layer, a first convolutional layer, a second convolutional layer, a first pooling layer, a second pooling layer, a flattening layer, a first fully connected layer, a second fully connected layer, a third fully connected layer, a Dropout layer and an output layer; wherein the input layer, the first convolutional layer, the first pooling layer, the second convolutional layer, the second pooling layer, the flattening layer, the first fully connected layer, the second fully connected layer, the third fully connected layer, the Dropout layer and the output layer are connected in sequence; the input layer is used to input the input data database; the output layer is used to output the model parameters of the rainstorm waterlogging model for each study area.

[0067] Among them, the flattening layer is used to convert the output of the convolution and pooling layers into a one-dimensional vector to facilitate transmission to the fully connected layer. The number of neurons in the flattening layer is 256.

[0068] Among them, the number of neurons in the first fully connected layer, the second fully connected layer, and the third fully connected layer are 64, 256, and 512, respectively.

[0069] The Dropout layer is used to reduce the overfitting of the model and enhance the generalization ability. The random inactivation rate of the Dropout layer is set to 0.1.

[0070] In the specific implementation, in a one-dimensional convolutional neural network, the previous one-dimensional convolutional layer is forward propagated to the neuron input of the next convolutional layer. The convolution process is as follows:

[0071]

[0072] Where: It is the input of the kth neuron in the lth convolutional layer; is the bias of the kth neuron in the lth convolutional layer; is the output of the i-th neuron in the (l-1)-th convolutional layer; is the convolution kernel from the i-th neuron in the (l-1)th convolutional layer to the k-th neuron in the l-th convolutional layer; ConvlD represents a one-dimensional convolution operation;

[0073] for The feature vector of the kth neuron in the lth convolutional layer obtained after the activation function F and the downsampling function SS, the pooling process is as follows:

[0074]

[0075] Finally, a deep convolutional neural network model is constructed by adding a fully connected layer.

[0076] In some embodiments, the first convolutional layer and the second convolutional layer are respectively followed by an activation function, the activation function is a ReLU function, and the ReLU function is in the following form:

[0077]

[0078] Among them, R(x) is the output value of the activation function, and x is the input value of the activation function, that is, the output value of the previous connection layer.

[0079] Step S205: the processor constructs an input data database based on the underlying surface data, pipe network data and rainfall data of each study area.

[0080] In some embodiments, after generating an input data database and a model parameter database, before training the deep learning model with the input data database as an independent variable and the model parameter database as a dependent variable, the method further includes: normalizing the input data database and the model parameter database; wherein the normalization calculation formula is as follows:

[0081]

[0082] Among them, X i is the normalized value, x is the initial value, X max With X min are the maximum and minimum values ​​for each database, respectively.

[0083] In step S206, the processor trains the deep learning model using the input data database as an independent variable and the model parameter database as a dependent variable.

[0084] Step S207: The processor inputs the underlying surface data, pipe network data, and rainfall data of the target study area into the trained deep learning model to obtain model parameters of the rainstorm waterlogging model of the target study area.

[0085] It can be seen that in the embodiment of the present application, the processor obtains the underlying surface data, pipeline network data, rainfall data and measured hydrological data of multiple study areas, and constructs a rainstorm waterlogging model for each study area based on the underlying surface data, pipeline network data and rainfall data; calibrates and verifies the corresponding rainstorm waterlogging model based on the rainfall data and measured hydrological data of each study area to obtain a model parameter database, which includes the optimized model parameters of each rainstorm waterlogging model; constructs a deep learning model, and constructs an input data database based on the underlying surface data, pipeline network data and rainfall data of each study area; trains the deep learning model with the input data database as the independent variable and the model parameter database as the dependent variable; inputs the underlying surface data, pipeline network data and rainfall data of the target study area into the trained deep learning model to obtain the model parameters of the rainstorm waterlogging model of the target study area. It can be seen that in this application, the deep learning model is applied to the parameter determination of the urban rainstorm waterlogging model with a clear physical mechanism. Compared with the traditional manual debugging method, the model parameters of the urban rainstorm waterlogging model can be quickly and accurately determined, thereby improving the accuracy and timeliness of urban rainstorm waterlogging warnings.

[0086] The above mainly introduces the scheme of the embodiment of the present application from the perspective of the execution process on the method side. It is understandable that, in order to realize the above functions, the controller includes a hardware structure and / or software module corresponding to the execution of each function. Those skilled in the art should easily realize that, in combination with the units and algorithm steps of each example described in the embodiment provided herein, the present application can be implemented in the form of hardware or a combination of hardware and computer software. Whether a function is executed in the form of hardware or computer software driving hardware depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of the present application.

[0087] The embodiment of the present application can divide the controller into functional units according to the above method example. For example, each functional unit can be divided according to each function, or two or more functions can be integrated into one processing module. The above integrated unit can be implemented in the form of hardware or in the form of a software program module. It should be noted that the division of units in the embodiment of the present application is schematic, which is only a logical function division, and there may be other division methods in actual implementation.

[0088] In the case of integrated units, see Figure 4 , Figure 4 This is a functional unit block diagram of a device for quickly determining parameters of an urban waterlogging model based on deep learning provided in an embodiment of the present application. The device 4 for quickly determining parameters of an urban waterlogging model includes:

[0089] An acquisition unit 401 is used to acquire underlying surface data, pipe network data, rainfall data and measured hydrological data of multiple study areas;

[0090] Processing unit 402 is used to construct a rainstorm waterlogging model for each study area based on the underlying surface data, the pipe network data, and the rainfall data; and calibrate and verify the corresponding rainstorm waterlogging model based on the rainfall data and measured hydrological data of each study area to obtain a model parameter database, wherein the model parameter database includes optimized model parameters of each rainstorm waterlogging model; and construct a deep learning model, and construct an input data database based on the underlying surface data, pipe network data, and rainfall data of each study area; and train the deep learning model using the input data database as an independent variable and the model parameter database as a dependent variable; and input the underlying surface data, pipe network data, and rainfall data of the target study area into the trained deep learning model to obtain the model parameters of the rainstorm waterlogging model of the target study area.

[0091] It can be seen that in the embodiment of the present application, the processor obtains the underlying surface data, pipeline network data, rainfall data and measured hydrological data of multiple study areas, and constructs a rainstorm waterlogging model for each study area based on the underlying surface data, pipeline network data and rainfall data; calibrates and verifies the corresponding rainstorm waterlogging model based on the rainfall data and measured hydrological data of each study area to obtain a model parameter database, which includes the optimized model parameters of each rainstorm waterlogging model; constructs a deep learning model, and constructs an input data database based on the underlying surface data, pipeline network data and rainfall data of each study area; trains the deep learning model with the input data database as the independent variable and the model parameter database as the dependent variable; inputs the underlying surface data, pipeline network data and rainfall data of the target study area into the trained deep learning model to obtain the model parameters of the rainstorm waterlogging model of the target study area. It can be seen that in this application, the deep learning model is applied to the parameter determination of the urban rainstorm waterlogging model with a clear physical mechanism. Compared with the traditional manual debugging method, the model parameters of the urban rainstorm waterlogging model can be quickly and accurately determined, thereby improving the accuracy and timeliness of urban rainstorm waterlogging warnings.

[0092] In some embodiments, the underlying surface data includes terrain data, elevation data and land use type, and the land use type includes roads, buildings and green spaces; the pipeline network data includes pipeline data, manhole data and the connection method between pipelines and manholes, and the pipeline data includes the spatial distribution of the pipeline network, pipeline length, pipeline cross-sectional shape, pipe diameter, upstream and downstream bottom elevations, and the manhole data includes the depth and size of the manhole; the rainfall data includes the rainfall amount of the measured times; the measured hydrological data includes the runoff data of the pipeline at the measured times, and the monitoring water level of the manholes and waterlogging points during the measured rainfall; wherein, there is a corresponding relationship between the measured times of rainfall data and measured hydrological data in the same study area, and the rainfall data and measured hydrological data in the same study area include data of at least one measured time, and the start time and test duration of the same measured time are the same.

[0093] In some embodiments, the processing unit 402 constructs a rainstorm waterlogging model for each study area based on the underlying surface data, the pipe network data, and the rainfall data, including: inputting the underlying surface data, the pipe network data, and the rainfall data into InfoWorks ICM software to obtain the rainstorm waterlogging model for each study area.

[0094] In some embodiments, the processing unit 402 calibrates and verifies the corresponding rainstorm waterlogging model according to the rainfall data and measured hydrological data of each study area to obtain the model parameter database, including: for a single study area, inputting the rainfall data of at least one measured session into the rainstorm waterlogging model to obtain at least one simulated hydrological data; calculating the simulation accuracy of the rainstorm waterlogging model according to the at least one simulated hydrological data and at least one measured hydrological data; if the simulation accuracy meets the preset simulation accuracy, determining the model parameters of the rainstorm waterlogging model as the target model parameters and writing them into the model parameter database; if the simulation accuracy does not meet the preset simulation accuracy, adjusting the model parameters of the rainstorm waterlogging model, and re-executing the step of "inputting the rainfall data of at least one measured session into the rainstorm waterlogging model to obtain at least one simulated hydrological data".

[0095] In some embodiments, the simulation accuracy includes a Nash efficiency coefficient, and the calculation formula of the Nash efficiency coefficient is as follows:

[0096]

[0097] Among them, NSE is the Nash efficiency coefficient, Yi is the measured hydrological data of the ith measured field, is the simulated hydrological data of the ith measured event, is the average value of the measured hydrological data of all measured sessions, n is the total number of measured sessions, and the value range of NSE is (-∞,1].

[0098] In some embodiments, before the processing unit 402 trains the deep learning model using the input data database as an independent variable and the model parameter database as a dependent variable, the method further includes: normalizing the input data database and the model parameter database; wherein the normalization calculation formula is as follows:

[0099]

[0100] Among them, X i is the normalized value, x is the initial value, X max With X min They are the maximum and minimum values ​​of a single database, respectively.

[0101] In some embodiments, constructing a deep learning model includes: using the Keras library in the Python language to construct a convolutional neural network (CNN) model, the CNN model including an input layer, a first convolutional layer, a second convolutional layer, a first pooling layer, a second pooling layer, a flattening layer, a first fully connected layer, a second fully connected layer, a third fully connected layer, a Dropout layer and an output layer; wherein the input layer, the first convolutional layer, the first pooling layer, the second convolutional layer, the second pooling layer, the flattening layer, the first fully connected layer, the second fully connected layer, the third fully connected layer, the Dropout layer and the output layer are connected in sequence; the input layer is used to input the input data database; the output layer is used to output the model parameters of the rainstorm waterlogging model for each study area.

[0102] In some embodiments, the first convolutional layer and the second convolutional layer are respectively followed by an activation function, the activation function is a ReLU function, and the ReLU function is in the following form:

[0103]

[0104] Among them, R(x) is the output value of the activation function, and x is the input value of the activation function, that is, the output value of the previous connection layer.

[0105] An embodiment of the present application provides a computer-readable storage medium having a computer program / instruction stored thereon. When the computer program / instruction is executed by a processor, the steps of the method described in any possible embodiment are implemented.

[0106] It should be noted that, for the aforementioned method embodiments, for the sake of simplicity, they are all expressed as a series of action combinations, but those skilled in the art should be aware that the present application is not limited by the described order of actions, because according to the present application, certain steps can be performed in other orders or simultaneously. Secondly, those skilled in the art should also be aware that the embodiments described in the specification are all preferred embodiments, and the actions and modules involved are not necessarily required by the present application.

[0107] In the above embodiments, the description of each embodiment has its own emphasis. For parts that are not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0108] In the several embodiments provided in the present application, it should be understood that the disclosed device can be implemented in other ways. For example, the device embodiments described above are only schematic, such as the division of the above-mentioned units, which is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, and the indirect coupling or communication connection of devices or units can be electrical or other forms.

[0109] The units described above as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0110] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0111] If the above-mentioned integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable memory. Based on this understanding, the technical solution of the present application is essentially or the part that contributes to the prior art or all or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a memory, including a number of instructions to enable a computer device (which can be a personal computer, server or network device, etc.) to perform all or part of the steps of the above-mentioned methods of each embodiment of the present application. The aforementioned memory includes: U disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), mobile hard disk, disk or CD-ROM and other media that can store program codes.

[0112] A person skilled in the art can understand that all or part of the steps in the various methods of the above embodiments can be completed by instructing related hardware through a program, and the program can be stored in a computer-readable memory, and the memory can include: a flash drive, a read-only memory (English: Read-Only Memory, abbreviated as: ROM), a random access memory (English: Random Access Memory, abbreviated as: RAM), a magnetic disk or an optical disk, etc.

[0113] The embodiments of the present application are introduced in detail above. Specific examples are used in this article to illustrate the principles and implementation methods of the present application. The description of the above embodiments is only used to help understand the method and core idea of ​​the present application. At the same time, for general technical personnel in this field, according to the idea of ​​the present application, there will be changes in the specific implementation method and application scope. In summary, the content of this specification should not be understood as a limitation on the present application.

Claims

1. A method for quickly determining parameters of urban waterlogging model based on deep learning, characterized in that: The method comprises: Obtain underlying surface data, pipe network data, rainfall data and measured hydrological data for multiple study areas; Constructing a rainstorm waterlogging model for each study area according to the underlying surface data, the pipe network data, and the rainfall data; Calibrate and verify the corresponding rainstorm waterlogging model according to the rainfall data and measured hydrological data of each study area to obtain a model parameter database, wherein the model parameter database includes optimized model parameters of each rainstorm waterlogging model; Build a deep learning model and construct an input data database based on the underlying surface data, pipeline network data, and rainfall data for each study area; Training the deep learning model using the input data database as an independent variable and the model parameter database as a dependent variable; The underlying surface data, pipeline network data and rainfall data of the target study area are input into the trained deep learning model to obtain the model parameters of the rainstorm waterlogging model of the target study area.

2. The method according to claim 1, characterized in that The underlying surface data includes terrain data, elevation data and land use types, and the land use types include roads, buildings and green spaces; The pipe network data includes pipe data, inspection well data and the connection mode between the pipe and the inspection well. The pipe data includes the spatial distribution of the pipe network, the pipe length, the pipe cross-sectional shape, the pipe diameter, and the upstream and downstream bottom elevations. The inspection well data includes the depth and size of the inspection well. The rainfall data includes the rainfall of the measured sessions; The measured hydrological data include measured runoff data of pipelines, and water level monitoring of inspection wells and waterlogging points during measured rainfall events; Among them, there is a corresponding relationship between the measured times of rainfall data and measured hydrological data in the same study area, and the rainfall data and measured hydrological data in the same study area include data of at least one measured time, and the start time and test duration of the same measured time are the same.

3. The method according to claim 1, characterized in that The constructing of a rainstorm waterlogging model for each study area according to the underlying surface data, the pipe network data, and the rainfall data includes: The underlying surface data, the pipe network data, and the rainfall data are input into InfoWorks ICM software to obtain a rainstorm waterlogging model for each study area.

4. The method according to claim 1, characterized in that: The corresponding rainstorm waterlogging model is calibrated and verified according to the rainfall data and measured hydrological data of each study area to obtain a model parameter database, including: For a single study area, inputting the rainfall data of at least one measured event into a rainstorm waterlogging model to obtain at least one simulated hydrological data; Calculating the simulation accuracy of the rainstorm waterlogging model according to the at least one simulated hydrological data and the at least one measured hydrological data; If the simulation accuracy meets the preset simulation accuracy, the model parameters of the rainstorm waterlogging model are determined as target model parameters and written into the model parameter database; If the simulation accuracy does not meet the preset simulation accuracy, the model parameters of the rainstorm waterlogging model are adjusted, and the step of "inputting the rainfall data of at least one measured event into the rainstorm waterlogging model to obtain at least one simulated hydrological data" is re-executed.

5. The method according to claim 4, characterized in that The simulation accuracy includes the Nash efficiency coefficient, and the calculation formula of the Nash efficiency coefficient is as follows: Among them, NSE is the Nash efficiency coefficient, Yi is the measured hydrological data of the ith measured field, is the simulated hydrological data of the ith measured event, is the average value of the measured hydrological data of all measured sessions, n is the total number of measured sessions, and the value range of NSE is (-∞,1].

6. The method according to claim 1, characterized in that Before training the deep learning model using the input data database as an independent variable and the model parameter database as a dependent variable, the method further includes: Normalizing the input data database and the model parameter database; The normalized calculation formula is as follows: Among them, X i is the normalized value, x is the initial value, X max With X min They are the maximum and minimum values ​​of a single database, respectively.

7. The method according to claim 1, characterized in that The deep learning model is constructed, comprising: A convolutional neural network (CNN) model is constructed using the Keras library in Python language, wherein the CNN model includes an input layer, a first convolutional layer, a second convolutional layer, a first pooling layer, a second pooling layer, a flattening layer, a first fully connected layer, a second fully connected layer, a third fully connected layer, a Dropout layer, and an output layer; Among them, the input layer, the first convolution layer, the first pooling layer, the second convolution layer, the second pooling layer, the flattening layer, the first fully connected layer, the second fully connected layer, the third fully connected layer, the Dropout layer and the output layer are connected in sequence; An input layer, used for inputting the input material database; The output layer is used to output the model parameters of the rainstorm waterlogging model for each study area.

8. The method according to claim 7, characterized in that The first convolutional layer and the second convolutional layer are respectively followed by an activation function, and the activation function is a ReLU function, and the ReLU function is as follows: Among them, R(x) is the output value of the activation function, and x is the output value of the previous connection layer of the activation function.

9. A device for quickly determining parameters of urban waterlogging model based on deep learning, characterized in that: include: An acquisition unit is used to acquire underlying surface data, pipe network data, rainfall data and measured hydrological data of multiple study areas; A processing unit is used to construct a rainstorm waterlogging model for each study area based on the underlying surface data, the pipe network data, and the rainfall data; and to calibrate and verify the corresponding rainstorm waterlogging model based on the rainfall data and measured hydrological data of each study area to obtain a model parameter database, wherein the model parameter database includes optimized model parameters of each rainstorm waterlogging model; and to construct a deep learning model, and to construct an input data database based on the underlying surface data, pipe network data, and rainfall data of each study area; and to train the deep learning model using the input data database as an independent variable and the model parameter database as a dependent variable; and to input the underlying surface data, pipe network data, and rainfall data of the target study area into the trained deep learning model to obtain the model parameters of the rainstorm waterlogging model of the target study area.

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

Images