Physical and data fusion driven flood simulation method

Abstract

The invention provides a flood simulation method driven by physical and data fusion, and belongs to the technical field of flood simulation, the method comprises the following steps: pre-training to obtain a qualified urban flood simulation model, a target loss function comprising a data supervision loss item and a physical constraint loss item, the physical constraint loss item is composed of a mass conservation residual error loss item, a momentum conservation residual error loss item and an energy conservation residual error loss item; inputting ground elevation data and historical meteorological data of a target area in a city into the hydrological model to obtain water depth grid data and flow velocity grid data, and inputting the water depth grid data and the flow velocity grid data into the city flood simulation model for prediction processing to obtain water depth grid data and flow velocity grid data of the target area at different time points in the future; and based on the data at different time points, generating an urban flood ponding dynamic evolution graph of the target area. The invention aims to improve the efficiency and accuracy of urban flood scene prediction.

Description

Technical Field

[0001] The present application relates to the technical field of flood simulation, and in particular to a flood simulation method driven by physics and data fusion. Background Art

[0002] Urban flooding can have serious impacts on cities and residents, including economic losses, infrastructure damage, and public service disruptions. Therefore, urban flood simulation is particularly important. Traditional urban flood simulation methods primarily rely on physics-based models, which simulate hydrological and hydrodynamic processes to reveal the physical mechanisms of flood formation. These models have been the core of early research, but running them requires various hydrological and geomorphological monitoring datasets and intensive computation. Summary of the Invention

[0003] In view of this, the present application provides a flood simulation method driven by physics and data fusion, aiming to solve or partially solve the problems existing in the background technology.

[0004] In a first aspect of the present application, a flood simulation method driven by physics and data fusion is provided, the method comprising: Pre-training to obtain a qualified urban flood simulation model, wherein the objective loss function in the urban flood simulation model includes a data supervision loss term and a physical constraint loss term, wherein the physical constraint loss term is composed of a mass conservation residual loss term, a momentum conservation residual loss term, and an energy conservation residual loss term; Collect historical meteorological data of a target area in the city during the current rainfall event for a period of time, and input the historical meteorological data and the elevation data of the target area into the LISFLOOD-FP hydrological model for hydrological simulation to obtain water depth raster data and flow velocity raster data for the period of time, wherein the meteorological data includes at least rainfall intensity data and flow data; Inputting the water depth raster data and the flow velocity raster data of the past period into the urban flood simulation model for prediction processing, and obtaining the water depth raster data and flow velocity raster data of the target area at different time points in the future period; Based on the obtained water depth raster data and flow velocity raster data of the target area at different time points in the future, a dynamic evolution map of urban flooding and waterlogging in the target area is generated.

[0005] Compared with the prior art, this application has the following advantages: The embodiment of the present application provides a flood simulation method driven by physics and data fusion. First, a qualified urban flood simulation model is pre-trained, wherein the target loss function in the urban flood simulation model includes a data supervision loss term and a physical constraint loss term, wherein the physical constraint loss term is composed of a mass conservation residual loss term, a momentum conservation residual loss term, and an energy conservation residual loss term; the collected ground elevation data and meteorological data of the target area in the current city are input into the LISFLOOD-FP hydrological model for hydrological simulation to obtain current water depth raster data and flow velocity raster data, wherein the meteorological data at least includes rainfall intensity data and flow data; the current water depth raster data and flow velocity raster data are input into the urban flood simulation model for prediction processing to obtain water depth raster data and flow velocity raster data of the target area at different time points in the future; based on the obtained water depth raster data and flow velocity raster data of the target area at different time points in the future, a dynamic evolution diagram of urban flood water accumulation in the target area is generated.

[0006] Therefore, this application embeds the conservation of mass, momentum, and energy of water flow as physical constraints into the urban flood simulation model. This allows the model to directly integrate physical constraints during training, ensuring that the model's output not only conforms to observed data but also adheres to the basic laws of fluid dynamics. This allows for more accurate dynamic simulation of the evolution of urban flood water over different time periods. Simultaneously, the model is trained using simulation data from the LISFLOOD-FP hydrological model, enabling more accurate and efficient urban flood scenario forecasting.

[0007] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are listed below. 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 following briefly introduces the drawings required for use in the embodiments or the description of the prior art.

[0009] Figure 1 A flow chart of a flood simulation method driven by physics and data fusion provided in an embodiment of the present application; Figure 2 A physical and data fusion driven flood simulation method provided in an embodiment of the present application simulates a dynamic evolution diagram of urban flood water accumulation in a certain urban area under a certain rainfall scenario. DETAILED DESCRIPTION

[0010] Exemplary embodiments of the present application will be described in more detail below with reference to the accompanying drawings.

[0011] Figure 1 A flow chart of a flood simulation method driven by physics and data fusion provided in an embodiment of the present application is shown in FIG. Figure 1 As shown, the method includes: Step S1: Pre-training to obtain a qualified urban flood simulation model, wherein the objective loss function in the urban flood simulation model includes a data supervision loss term and a physical constraint loss term, and the physical constraint loss term is composed of a mass conservation residual loss term, a momentum conservation residual loss term, and an energy conservation residual loss term.

[0012] In this embodiment, the present application obtains a qualified urban flood simulation model through pre-training. The target loss function in the urban flood simulation model includes two loss terms, namely, a data supervision loss term and a physical constraint loss term.

[0013] In this embodiment, this application provides a method for simulating urban flooding. Urban flood simulation in urban settings is relatively complex. Therefore, to ensure the accuracy of urban flood simulation, this application proposes a physical constraint loss term for the objective loss function. This physical constraint loss term consists of a mass conservation residual loss term, a momentum conservation residual loss term, and an energy conservation residual loss term. This physical loss term is proposed because, due to the large number of artificial structures such as buildings, roads, and pipelines in urban areas, water frequently undergoes a conversion between kinetic energy and potential energy during its movement (e.g., when water strikes a building, its velocity decreases but the water level rises, or when potential energy is converted to kinetic energy when it falls from a height). Therefore, by constraining the sum of kinetic energy and gravitational potential energy to be conserved in space and time, the energy conservation loss term avoids non-physical phenomena such as model predictions of abnormally increased velocity without a decrease in water level, and energy violations when water flows over obstacles, thereby ensuring that the simulation results conform to actual energy transfer laws. Therefore, this application simultaneously determines the mass conservation residual loss term, momentum conservation residual loss term and energy conservation residual loss term as physical constraint loss terms.

[0014] Step S2: Collect historical meteorological data of the target area in the city during the current rainfall event for a period of time, and input the historical meteorological data and the elevation data of the target area into the LISFLOOD-FP hydrological model for hydrological simulation to obtain water depth raster data and flow velocity raster data for a period of time, wherein the meteorological data at least includes rainfall intensity data and flow data.

[0015] In this embodiment, when a rainfall event is actually occurring, historical meteorological data for a target area in a city is collected from the onset of the rainfall event to the present time, and the ground elevation data for the target area in the city, recorded in a database, is simultaneously obtained. The obtained ground elevation data for the target area and the collected historical meteorological data are then input into the LISFLOOD-FP hydrological model for hydrological simulation, and water depth raster data and flow velocity raster data for the target area at various time points during the past period are predicted. The target area in the city can be any part of the city, such as a 50 square kilometer area or a 100 square kilometer area. The meteorological data includes at least rainfall intensity data and flow data.

[0016] Step S3: inputting the water depth raster data and the flow velocity raster data of the past period into the urban flood simulation model for prediction processing, and obtaining the water depth raster data and flow velocity raster data of the target area at different time points in the future period.

[0017] In this embodiment, after obtaining a qualified urban flood simulation model through pre-training in step S1, under actual rainfall events, the water depth raster data and flow velocity raster data of the target area at various time points in the past period of time obtained through hydrological simulation using the LISFLOOD-FP hydrological model are input into the qualified urban flood simulation model for prediction processing. The urban flood simulation model will output the water depth raster data and flow velocity raster data of the target area at different time points in the future period of time starting from the current moment. The future period of time can be pre-set according to the actual application scenario, such as within 5 hours, within 10 hours, within 24 hours, etc. The time intervals of different time points can be pre-set according to the actual application scenario, such as a water depth raster data result and flow velocity raster data result every 10 minutes, a water depth raster data result and flow velocity raster data result every 5 minutes, etc.

[0018] Step S4: Based on the obtained water depth raster data and flow velocity raster data of the target area at different time points in the future, a dynamic evolution diagram of urban flooding and waterlogging in the target area is generated.

[0019] In this embodiment, after obtaining the water depth raster data and flow velocity raster data of the target area at different time points in the future starting from the current moment through step S3, the water depth raster data and flow velocity raster data obtained at different time points are processed by corresponding hydrological analysis tools (such as ArcGIS, etc.) to obtain the dynamic evolution map of urban flooding and waterlogging in the target area. Figure 2 As shown, Figure 2The figure shows the dynamic evolution of flood water accumulation in a designated area of ​​a city under a certain rainfall scenario, using the physics and data fusion-driven flood simulation method provided in this application, 60 minutes, 120 minutes, 180 minutes and 240 minutes after the rainfall event. The water depth in the figure is in meters.

[0020] The embodiment of the present application provides a flood simulation method driven by physics and data fusion. First, a qualified urban flood simulation model is pre-trained, wherein the target loss function in the urban flood simulation model includes a data supervision loss term and a physical constraint loss term, wherein the physical constraint loss term is composed of a mass conservation residual loss term, a momentum conservation residual loss term, and an energy conservation residual loss term; the collected ground elevation data and meteorological data of the target area in the current city are input into the LISFLOOD-FP hydrological model for hydrological simulation to obtain current water depth raster data and flow velocity raster data, wherein the meteorological data at least includes rainfall intensity data and flow data; the current water depth raster data and flow velocity raster data are input into the urban flood simulation model for prediction processing to obtain water depth raster data and flow velocity raster data of the target area at different time points in the future; based on the obtained water depth raster data and flow velocity raster data of the target area at different time points in the future, a dynamic evolution diagram of urban flood water accumulation in the target area is generated.

[0021] Therefore, this application embeds the conservation of mass, momentum, and energy of water flow as physical constraints into the urban flood simulation model. This allows the model to directly integrate physical constraints during training, ensuring that the model's output not only conforms to observed data but also adheres to the basic laws of fluid dynamics. This allows for more accurate dynamic simulation of the evolution of urban flood water over different time periods. Simultaneously, the model is trained using simulation data from the LISFLOOD-FP hydrological model, enabling more accurate and efficient urban flood scenario forecasting.

[0022] In combination with the above embodiments, in one implementation, the present application also provides a flood simulation method driven by physics and data fusion. In the flood simulation method driven by physics and data fusion, step S1 may include: Step S11: Construct a training data set and a test data set for model training, wherein the sample data in the data set includes water depth raster data and flow velocity raster data of a local area in the city at a target time, and water depth raster data and flow velocity raster data of the local area at different time points within a period of time after the target time.

[0023] In this embodiment, the present application first constructs a data set required for model training, and divides the data set into a training data set and a test data set according to a preset ratio. The preset ratio is preferably set to 80% of the sample data in the training data set and 20% of the sample data in the test data set. A sample data in the data set includes: water depth raster data and flow velocity raster data of a local area in the city at a target time, and water depth raster data and flow velocity raster data of the local area at different time points within a period of time after the target time. The target time is a certain time after the rainfall event occurs, the duration of the period of time is the same as the above-mentioned future period of time, and the time interval between the two time points is the same. The water depth raster data and flow velocity raster data of the local area at different time points within a period of time after the target time belong to the label of the sample data. The data set of the present application includes sample data of the same local area in the city under the influence of different meteorological data, and sample data of different local areas in the city under the influence of the same meteorological data.

[0024] Step S12: performing model parameter optimization training on the constructed initial urban flood simulation model using the sample data in the training data set.

[0025] In this embodiment, after the training data set and the test data set are constructed and obtained in step S11, the model parameter optimization training is performed on the constructed initial urban flood simulation model using the sample data in the training data set.

[0026] Step S13: Determine the training result of the initial urban flood simulation model through the target loss function.

[0027] In this embodiment, during the training process of the initial urban flood simulation model, the training result of the model is determined by the target loss function of the model, and when the training result does not meet the set conditions, the model parameters are updated to continue the model training.

[0028] Step S14: When the training result satisfies the set conditions, the initial urban flood simulation model after training is evaluated using the sample data in the test data set to obtain a corresponding evaluation result.

[0029] In this embodiment, when the training results obtained by the model's objective loss function satisfy the set conditions, the trained initial urban flood simulation model is further evaluated using sample data from the constructed test dataset to obtain the corresponding evaluation results. The set conditions can include verifying the stability of the loss. When the loss stops decreasing or begins to increase after several consecutive rounds (e.g., 3-5 rounds), the model is determined to have converged or is overfitting, and training is terminated. The set conditions can also include early stopping, with a patience value (e.g., patience = 5). If the loss does not improve within 5 consecutive rounds, training is terminated.

[0030] Step S15: When the evaluation result meets the application requirements, a qualified urban flood simulation model is obtained.

[0031] In this embodiment, when the evaluation result obtained in step S14 meets the corresponding application requirements, it is determined that a qualified urban flood simulation model has been obtained. The application requirements may be that the accuracy, F1 score, MSE, etc. of the model meet corresponding preset thresholds.

[0032] In combination with the above embodiments, in one embodiment, the embodiment of the present application also provides a flood simulation method driven by physics and data fusion. In this flood simulation method driven by physics and data fusion, an initial urban flood simulation model is constructed, including: constructing a physical constraint loss term based on the mass conservation residual loss term, the momentum conservation residual loss term, and the energy conservation residual loss term, as well as the weights corresponding to each residual loss term; constructing a data supervision loss term based on the mean square error between the model's predicted results and the actual results; constructing a target loss function based on the constructed physical constraint loss term and the data supervision loss term, as well as the weights corresponding to each loss term; constructing a total residual vector based on the mass conservation, momentum conservation, and energy conservation of the water flow; and constructing an initial urban flood simulation model based on the target loss function and the total residual vector.

[0033] In this embodiment, one implementation method for constructing the initial urban flood simulation model is as follows: first, a mass conservation residual loss term, a momentum conservation residual loss term, and an energy conservation residual loss term are constructed. A corresponding weight term is then assigned to each residual loss term, i.e., each residual loss term has a corresponding weight term. The mass conservation residual loss term is multiplied by its corresponding weight term, the momentum conservation residual loss term is multiplied by its corresponding weight term, and the energy conservation residual loss term is multiplied by its corresponding weight term. Finally, these three products are added together to obtain the physical constraint loss term.

[0034] In this embodiment, the mean square error between the predicted result obtained by predicting the input data using the urban flood simulation model and the actual result corresponding to the input data is , construct the data supervision loss term. Among them, is the prediction result obtained by the model for the input data, To predict the water depth grid data, To predict the velocity grid data, is the actual result corresponding to the input data, is the real water depth grid data, It is the real flow velocity grid data.

[0035] In this embodiment, corresponding weights are assigned to the constructed physical constraint loss term and data supervision loss term. The physical constraint loss term is multiplied by its corresponding weight, and the data supervision loss term is multiplied by its corresponding weight. Finally, these two products are added together to obtain the target loss function.

[0036] In this embodiment, the present application constructs a total residual vector in the urban flood simulation model based on the conservation of mass, momentum, and energy of the water flow. Ultimately, based on the constructed target loss function and the total residual vector, the final initial urban flood simulation model is constructed.

[0037] In combination with the above embodiments, in one implementation, the embodiment of the present application further provides a flood simulation method driven by physics and data fusion. In the flood simulation method driven by physics and data fusion, a physical constraint loss term is constructed based on the mass conservation residual loss term, momentum conservation residual loss term, and energy conservation residual loss term, as well as the weight corresponding to each residual loss term, including: constructing the mass conservation equation of water flow , where h is the water depth, u is the longitudinal velocity in the x-direction, v is the transverse velocity in the y-direction, and t is time; based on the mass conservation equation for water flow, determine the mass conservation residual loss term ,in, is the residual loss term for mass conservation; construct the momentum conservation equation for water flow ; Based on the momentum conservation equation of water flow, determine the momentum conservation residual loss term ,in, is the residual loss term for momentum conservation; construct the energy conservation equation for water flow ; Based on the energy conservation equation of water flow, determine the energy conservation residual loss term ,in, is the energy conservation residual loss term; based on the determined mass conservation residual loss term, momentum conservation residual loss term and energy conservation residual loss term, as well as the weight corresponding to each residual loss term, a physical constraint loss term is constructed ,in, 、 and is the weight coefficient.

[0038] In this embodiment, the present application first constructs the mass conservation equation of water flow , where h is the water depth, u is the longitudinal velocity in the x-direction, v is the transverse velocity in the y-direction, and t is time. And, construct the momentum conservation equation for the water flow . And, construct the energy conservation equation of water flow .

[0039] In this embodiment, after constructing the mass conservation equation, momentum conservation equation and energy conservation equation of the water flow, the mass conservation residual loss term is determined based on the mass conservation equation of the water flow. ,in, is the residual loss term for mass conservation. And, based on the momentum conservation equation of water flow, determine the residual loss term for momentum conservation ,in, is the residual loss term for conservation of momentum. And, based on the energy conservation equation of water flow, determine the residual loss term for conservation of energy ,in, is the energy conservation residual loss term.

[0040] In this embodiment, after constructing the mass conservation residual loss term, momentum conservation residual loss term, and energy conservation residual loss term, a physical constraint loss term is constructed based on these three residual loss terms and the weights corresponding to each residual loss term. ,in, 、 and is the weight coefficient.

[0041] In combination with the above embodiments, in one embodiment, the present application also provides a flood simulation method driven by physics and data fusion. In the flood simulation method driven by physics and data fusion, an optional implementation method of constructing the total residual vector based on the mass conservation, momentum conservation and energy conservation of water flow is: based on the mass conservation equation of water flow, determine the mass conservation residual term ; Based on the momentum conservation equation of water flow, determine the momentum conservation residual term ; Based on the energy conservation equation of water flow, determine the energy conservation residual term Based on the determined mass conservation residual term, momentum conservation residual term and energy conservation residual term, as well as the weight corresponding to each residual term, a total residual vector is constructed ,in, 、 、 is the weight coefficient.

[0042] In combination with the above embodiments, in one implementation, the embodiment of the present application further provides a flood simulation method driven by physics and data fusion. In the flood simulation method driven by physics and data fusion, an initial urban flood simulation model is constructed based on the target loss function and the total residual vector, including: constructing a basic urban flood simulation model consisting of an input layer of a first number of neurons, a second number of hidden layers, and an output layer of a third number of neurons; setting each neuron in the first hidden layer to ,in, is the i-th neuron in the first hidden layer, is the temporal weight matrix of the i-th neuron in the first hidden layer, is the weight matrix of the i-th neuron in the first hidden layer in the x direction, is the weight matrix of the i-th neuron in the first hidden layer in the y direction, is the bias vector of the i-th neuron in the first hidden layer, is the hyperbolic tangent activation function; each neuron in the other hidden layers is set to ,in, is the jth neuron in all hidden layers except the first hidden layer, is the connection weight from the i-th neuron in the first hidden layer to the j-th neuron in other hidden layers, is the bias vector of the jth neuron in other hidden layers except the first hidden layer; based on the basic urban flood simulation model, the target loss function and the total residual vector, an initial urban flood simulation model is constructed.

[0043] In this embodiment, the present application first constructs a basic urban flood simulation model consisting of an input layer of a first number of neurons, a second number of hidden layers, and an output layer of a third number of neurons. The first number is preferably 3 neurons, the second number is preferably 4 hidden layers, each hidden layer preferably has 50 neurons, and the third number is preferably 3 neurons.

[0044] In this embodiment, each neuron in the first hidden layer is set to ,in, is the i-th neuron in the first hidden layer, is the temporal weight matrix of the i-th neuron in the first hidden layer, is the weight matrix of the i-th neuron in the first hidden layer in the x direction, is the weight matrix of the i-th neuron in the first hidden layer in the y direction, is the bias vector of the i-th neuron in the first hidden layer, is the hyperbolic tangent activation function. At the same time, each neuron in the other hidden layers is set to ,in, is the jth neuron in all hidden layers except the first hidden layer, is the connection weight from the i-th neuron in the first hidden layer to the j-th neuron in other hidden layers, is the bias vector of the jth neuron in all hidden layers except the first hidden layer.

[0045] In this embodiment, a final initial urban flood simulation model is constructed based on the constructed basic urban flood simulation model, the target loss function and the total residual vector.

[0046] In combination with the above embodiments, in one embodiment, the embodiment of the present application also provides a flood simulation method driven by physics and data fusion. In this flood simulation method driven by physics and data fusion, a target loss function is constructed based on the constructed physical constraint loss term and the data supervision loss term, as well as the weights corresponding to each loss term, including: determining the first weight of the physical constraint loss term to be one, and determining the second weight of the data supervision loss term to be one, so as to balance data fitting and physical constraints; and performing weighted summation of the physical constraint loss term and the data supervision loss term based on the determined first weight and second weight to construct a target loss function.

[0047] In this embodiment, the present application adopts a simplified fixed value strategy to balance the importance of data fitting and physical constraints in the model prediction process. That is, the first weight of the physical constraint loss term is set to 1, and the second weight of the data supervision loss term is also set to 1. The physical constraint loss term is then multiplied by its own first weight, and the data supervision loss term is multiplied by its own second weight. The results of these two multiplications are added together to obtain the final target loss function.

[0048] In combination with the above embodiments, in one embodiment, the embodiment of the present application further provides a flood simulation method driven by physics and data fusion. In this flood simulation method driven by physics and data fusion, sample data is constructed, including: inputting ground elevation data and meteorological data of a designated area in a city under a historical rainfall scenario into the LISFLOOD-FP hydrological model for hydrological simulation, obtaining water depth raster data and flow velocity raster data at different time points within a preset time period; determining the water depth raster data and flow velocity raster data at time points other than the starting time point obtained at different time points within the preset time period as sample labels, and together with the water depth raster data and flow velocity raster data at the starting time point, forming sample data.

[0049] In this embodiment, the water depth grid data and flow velocity grid data at each time point in the sample data are obtained by hydrological simulation using the LISFLOOD-FP hydrological model.

[0050] Specifically, the ground elevation data and meteorological data of a specified area in the city at a specific time under the historical rainfall scenario are input into the LISFLOOD-FP hydrological model for hydrological simulation, and the water depth raster data and flow velocity raster data of the specified area at the specific time are obtained. Through the same implementation method, the water depth raster data and flow velocity raster data at different time points within a continuous preset time length can be calculated. Then, the water depth raster data and flow velocity raster data obtained at other time points other than the starting time point within the preset time length are used as sample labels, and the water depth raster data and flow velocity raster data at the starting time point within the preset time length are used as input data that need to be input into the model. The sample label and the input data together constitute a sample data set to participate in subsequent model training. Through the same implementation method, a large amount of sample data is constructed to form a data set for model training. The training data set and test data set required for training the model are obtained by dividing the data set.

[0051] It should be noted that for the 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 embodiments of the present application are not limited by the order of the actions described, because according to the embodiments of 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 involved are not necessarily required by the embodiments of the present application.

[0052] The various embodiments in this specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the various embodiments can be referenced to each other.

[0053] Those skilled in the art will appreciate that the embodiments of the present application can be provided as methods, systems, or computer program products. Therefore, the embodiments of the present application can adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the embodiments of the present application can adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to magnetic disk storage, CD-ROM, optical storage, etc.) that contain computer-usable program code.

[0054] Although preferred embodiments of the present invention have been described, those skilled in the art may make additional changes and modifications to these embodiments once they become aware of the basic inventive concepts. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications that fall within the scope of the embodiments of the present invention.

[0055] The above is a detailed introduction to the flood simulation method driven by physical and data fusion provided by this application. This article uses specific examples to illustrate the principles and implementation methods of this application. The description of the above embodiments is only used to help understand the method of this application and its core idea; at the same time, for general technical personnel in this field, based on the ideas of this application, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as a limitation on this application.

Claims

1. A flood simulation method driven by physics and data fusion, characterized by: The method comprises: Pre-training to obtain a qualified urban flood simulation model, wherein the objective loss function in the urban flood simulation model includes a data supervision loss term and a physical constraint loss term, wherein the physical constraint loss term is composed of a mass conservation residual loss term, a momentum conservation residual loss term, and an energy conservation residual loss term; Collect historical meteorological data of a target area in the city during the current rainfall event for a period of time, and input the historical meteorological data and the elevation data of the target area into the LISFLOOD-FP hydrological model for hydrological simulation to obtain water depth raster data and flow velocity raster data for the period of time, wherein the meteorological data includes at least rainfall intensity data and flow data; Inputting the water depth raster data and the flow velocity raster data of the past period into the urban flood simulation model for prediction processing, and obtaining the water depth raster data and flow velocity raster data of the target area at different time points in the future period; Based on the obtained water depth raster data and flow velocity raster data of the target area at different time points in the future, a dynamic evolution map of urban flooding and waterlogging in the target area is generated.

2. A flood simulation method driven by physics and data fusion according to claim 1, characterized in that: Pre-training to obtain a qualified urban flood simulation model, including: Constructing a training data set and a test data set for model training, wherein the sample data in the data set includes water depth raster data and flow velocity raster data of a local area in the city at a target time, and water depth raster data and flow velocity raster data of the local area at different time points within a period of time after the target time; Performing model parameter optimization training on the constructed initial urban flood simulation model using sample data in the training data set; Determining the training result of the initial urban flood simulation model through the target loss function; When the training result satisfies the set conditions, the initial urban flood simulation model after training is evaluated using the sample data in the test data set to obtain a corresponding evaluation result; When the evaluation results meet the application requirements, a qualified urban flood simulation model is obtained.

3. The flood simulation method driven by physics and data fusion according to claim 2 is characterized in that: Construct an initial urban flood simulation model, including: Construct a physical constraint loss term based on the mass conservation residual loss term, momentum conservation residual loss term, and energy conservation residual loss term, as well as the weights corresponding to each residual loss term; Construct a data supervision loss term based on the mean squared error between the model's predictions and the true results; Constructing a target loss function based on the constructed physical constraint loss term and the data supervision loss term, as well as the weights corresponding to each loss term; Based on the conservation of mass, momentum and energy of water flow, the total residual vector is constructed; An initial urban flood simulation model is constructed based on the target loss function and the total residual vector.

4. The flood simulation method driven by physics and data fusion according to claim 3 is characterized in that: According to the mass conservation residual loss term, momentum conservation residual loss term and energy conservation residual loss term, as well as the weight corresponding to each residual loss term, the physical constraint loss term is constructed, including: Constructing the mass conservation equation for water flow , where h is the water depth, u is the longitudinal velocity in the x direction, v is the transverse velocity in the y direction, and t is the time; Based on the mass conservation equation of water flow, determine the mass conservation residual loss term ,in, is the mass conservation residual loss term; Constructing the momentum conservation equation for water flow ; Based on the momentum conservation equation of water flow, determine the momentum conservation residual loss term ,in, is the residual loss term of momentum conservation; Constructing the energy conservation equation for water flow ; Based on the energy conservation equation of water flow, determine the energy conservation residual loss term ,in, is the energy conservation residual loss term; Based on the determined mass conservation residual loss term, momentum conservation residual loss term and energy conservation residual loss term, as well as the weight corresponding to each residual loss term, a physical constraint loss term is constructed. ,in, 、 and is the weight coefficient.

5. The flood simulation method driven by physics and data fusion according to claim 3 is characterized in that: Based on the conservation of mass, momentum, and energy of water flow, the total residual vector is constructed, including: Determine the mass conservation residual term based on the mass conservation equation for water flow ; Based on the momentum conservation equation of water flow, determine the momentum conservation residual term ; Based on the energy conservation equation of water flow, determine the energy conservation residual term ; Based on the determined mass conservation residual term, momentum conservation residual term and energy conservation residual term, and the weight corresponding to each residual term, a total residual vector is constructed ,in, 、 、 is the weight coefficient.

6. The flood simulation method driven by physics and data fusion according to claim 3 is characterized in that: Based on the target loss function and the total residual vector, an initial urban flood simulation model is constructed, including: constructing a basic urban flood simulation model consisting of an input layer of a first number of neurons, a hidden layer of a second number of neurons, and an output layer of a third number of neurons; Set each neuron in the first hidden layer to ,in, is the i-th neuron in the first hidden layer, is the temporal weight matrix of the i-th neuron in the first hidden layer, is the weight matrix of the i-th neuron in the first hidden layer in the x direction, is the weight matrix of the i-th neuron in the first hidden layer in the y direction, is the bias vector of the i-th neuron in the first hidden layer, is the hyperbolic tangent activation function; Set the neurons in other hidden layers to ,in, is the jth neuron in all hidden layers except the first hidden layer, is the connection weight from the i-th neuron in the first hidden layer to the j-th neuron in other hidden layers, is the bias vector of the jth neuron in all hidden layers except the first hidden layer; Based on the basic urban flood simulation model, the target loss function and the total residual vector, an initial urban flood simulation model is constructed.

7. The flood simulation method driven by physics and data fusion according to claim 3 is characterized in that: According to the constructed physical constraint loss term and the data supervision loss term, as well as the weights corresponding to each loss term, a target loss function is constructed, including: Determining a first weight of the physical constraint loss term to be one, and determining a second weight of the data supervision loss term to be one, so as to balance data fitting and physical constraints; The physical constraint loss term and the data supervision loss term are weightedly summed using the determined first weight and the second weight to construct a target loss function.

8. The flood simulation method driven by physics and data fusion according to claim 2 is characterized in that: Construct sample data, including: The ground elevation data and meteorological data of the designated area in the city under the historical rainfall scenario are input into the LISFLOOD-FP hydrological model for hydrological simulation, and the water depth raster data and flow velocity raster data at different time points within the preset time period are obtained; The water depth raster data and flow velocity raster data at time points other than the starting time point obtained within the preset time length are determined as sample labels, and together with the water depth raster data and flow velocity raster data at the starting time point, constitute sample data.

Images