WO2024098950A1

WO2024098950A1 - Hydrodynamic model parameter optimization method and apparatus, and water level and flow change process simulation method and apparatus

Info

Publication number: WO2024098950A1
Application number: PCT/CN2023/118520
Authority: WO
Inventors: 刘肖廷; 庞国飞; 戴会超; 刘志武; 蒋定国; 翟俨伟; 吕超楠
Original assignee: 中国长江三峡集团有限公司
Priority date: 2022-11-10
Filing date: 2023-09-13
Publication date: 2024-05-16
Also published as: CN115563907A

Abstract

The present invention provides a hydrodynamic model parameter optimization method and apparatus, and a water level and flow change process simulation method and apparatus. The hydrodynamic model parameter optimization method comprises: establishing an optimization objective function by combining a first initial neural network model, a second initial neural network model and a hydrodynamic model, the roughness in the hydrodynamic model being determined by means of the second initial neural network model; solving the optimization objective function, and optimizing a first network model parameter and a second network model parameter to obtain a first network model optimized parameter and a second network model optimized parameter which enable the value of the optimization objective function to be minimum; and determining the second initial neural network model containing the second network model optimized parameter as a roughness optimization model. The roughness determined by the roughness optimization model optimized by using the method provided by embodiments of the present invention is more in line with an actual physical process.

Description

Hydrodynamic model parameter optimization, water level and flow rate change process simulation method and device

Technical Field

The present invention relates to the technical field of engineering simulation and numerical simulation, and in particular to a method and device for optimizing hydrodynamic model parameters and simulating water level and flow rate variation processes.

Background technique

In nature, river systems are complex and water levels are greatly affected by the external environment. Changes in topography, rainfall conditions, human activities, etc. will have a significant impact on the measurement of water levels. For complex large-scale river water level simulation, a distributed hydrodynamic model is usually used, and the roughness parameter is a key factor affecting the simulation results of the hydrodynamic model. Because it is a non-constant parameter affected by factors such as water level and riverbed section, and the distributed model needs to determine multiple roughness parameters, it is very easy to have multiple solutions, which makes it extremely difficult to accurately simulate river water levels.

The roughness parameter of the traditional hydrodynamic model in the calculation of the water surface line is highly dependent on the user's personal experience and debugging. The selection of this parameter is related to the current water level and the status of the section and is in constant change. The general neural network used to fit the roughness parameter mainly selects the optimal solution based on mathematical statistical methods, lacks strong constraints on the physical process, has no physical meaning, and often has non-unique solutions and overfitting that does not conform to the actual physical process. At the same time, the fitting process requires continuous solution of the control equation, which requires a large amount of calculation, limiting the further development and application of the model.

Summary of the invention

Therefore, the technical problem to be solved by the present invention is to overcome the defects in the prior art that the parameters fitted by the neural network have no physical meaning, and the output results of the optimized neural network model do not conform to the actual physical process, thereby providing a method and device for optimizing hydrodynamic model parameters and simulating water level and flow change processes.

The invention provides a method for optimizing parameters of a hydrodynamic model, comprising the following steps: establishing an optimization objective function by combining a first initial neural network model, a second initial neural network model and a hydrodynamic model, wherein water and sand parameters in the hydrodynamic model are determined by the first initial neural network model, and the roughness in the hydrodynamic model is determined by the second initial neural network model, and the optimization objective function is determined according to the sum of simulation residuals of each hydrodynamic model; solving the optimization objective function, optimizing the first network model parameters and the second network model parameters, and obtaining the first network model optimization parameters and the second network model optimization parameters that minimize the value of the optimization objective function, wherein the first network model parameters are parameters in the first initial neural network model, and the second network model parameters are parameters in the second initial neural network model; and determining the second initial neural network model including the second network model optimization parameters as a roughness optimization model.

Optionally, in the hydrodynamic model parameter optimization method provided by the present invention, the simulation residual of the hydrodynamic model includes the residual of the water flow continuity equation and the residual of the water flow motion equation.

Optionally, in the hydrodynamic model parameter optimization method provided by the present invention, the water and sediment parameters determined by the first initial neural network model include flow and water level, and the optimization objective function also includes the approximation error of the flow output by the first initial neural network model to the actual flow, and the approximation error of the water level output by the first initial neural network model to the actual water level.

Optionally, in the hydrodynamic model parameter optimization method provided by the present invention, the simulation residual of the hydrodynamic model includes:

Among them, _e1 represents the residual of the water flow continuity equation, _e2 represents the residual of the water flow motion equation, B represents the water surface width, _Zs represents the water level, _Qs represents the flow rate, t represents time, x represents space, _qL represents the lateral inflow flow rate per unit river length, A represents the water cross-sectional area, g represents the gravitational acceleration, _njs represents the roughness, and R represents the hydraulic radius. Among them, _Zs = _Zs (x, t; _θu ), _Qs = _Qs (x, t; _θu ), _njs = _njs (x; _θp ), _θu represents the first network model parameters, and _θp represents the second network model parameters.

The second aspect of the present invention provides a method for simulating a water level and flow change process, comprising: dividing a target river channel into multiple river sections, and obtaining water level and flow data of each river section; determining the roughness of each river section based on a roughness optimization model corresponding to each river section, wherein the roughness optimization model is determined according to the hydrodynamic model parameter optimization method provided by the first aspect of the present invention; respectively inputting the water level and flow data and roughness of each river channel into the hydrodynamic model to obtain the water level and flow change process of the target river channel.

Optionally, in the method for simulating the water level and flow rate change process provided by the present invention, the hydrodynamic model input data of the river section at the current moment also includes the hydrodynamic model output data of the adjacent upstream river section at the previous moment.

The third aspect of the present invention provides a hydrodynamic model parameter optimization device, including: an optimization objective function establishment module, which is used to establish an optimization objective function in combination with a first initial neural network model, a second initial neural network model, and a hydrodynamic model, wherein the water and sand parameters in the hydrodynamic model are determined by the first initial neural network model, and the roughness in the hydrodynamic model is determined by the second initial neural network model, and the optimization objective function is determined according to the sum of the simulation residuals of each hydrodynamic model; a network parameter optimization module, which is used to solve the optimization objective function, optimize the first network model parameters and the second network model parameters, and obtain the first network model optimization parameters and the second network model optimization parameters that minimize the value of the optimization objective function, wherein the first network model parameters are the parameters in the first initial neural network model, and the second network model parameters are the parameters in the second initial neural network model; an optimization model determination module, which is used to determine the second initial neural network model containing the second network model optimization parameters as the roughness optimization model.

The fourth aspect of the present invention provides a device for simulating water level and flow change processes, comprising: a data acquisition module, used to divide a target river channel into multiple river sections and obtain water level and flow data of each river section; a parameter determination module, used to determine the roughness of each river section based on a roughness optimization model corresponding to each river section, wherein the roughness optimization model is determined according to the hydrodynamic model parameter optimization method provided by the first aspect of the present invention; a simulation module, used to input the water level and flow data and roughness of each river channel into the hydrodynamic model respectively, to obtain the water level and flow change process of the target river channel.

The fifth aspect of the present invention provides a computer device, comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to execute the hydrodynamic model parameter optimization method provided in the first aspect of the present invention, or the water level flow change process simulation method provided in the second aspect of the present invention.

The sixth aspect of the present invention provides a computer-readable storage medium, which stores computer instructions, and the computer instructions are used to enable a computer to execute the hydrodynamic model parameter optimization method provided in the first aspect of the present invention, or the water level flow change process simulation method provided in the second aspect of the present invention.

The technical solution of the present invention has the following advantages:

1. The hydrodynamic model parameter optimization method provided by the present invention is based on the hydrodynamic model. The roughness in the hydrodynamic model is determined by the second initial neural network model. The second network model optimization parameters of the second neural network model determined in the process of solving the optimization objective function can minimize the sum of the simulation residuals of the hydrodynamic model. It can be seen that the method provided in the embodiment of the present invention optimizes the second initial neural network model on the basis of strong constraints on the physical process. The roughness determined by the roughness optimization model optimized by the method provided in the embodiment of the present invention is more in line with the actual physical process. In addition, the method provided in the embodiment of the present invention can directly optimize the objective function using a gradient optimization algorithm by placing the control equation of the hydrodynamic model into the optimized objective function without iterating the equation. At the same time, a physical driving term is introduced in the objective function, so that the optimization algorithm requires less observation data. This method can obtain the optimal solution of the roughness parameter with physical significance without solving the control equation, which can effectively improve the simulation efficiency and accuracy of the river water level and flow change process.

2. The method for simulating the water level and flow change process provided by the present invention divides the target river channel into multiple river sections, determines the roughness of each river section according to the roughness optimization model corresponding to each river section, and inputs the water level and flow data and roughness of each river channel into the hydrodynamic model respectively to obtain the water level and flow change process of the target river channel. Since the roughness optimization model used in the embodiment of the present invention is optimized by implementing the second initial neural network model based on the strong constraints of the physical process, the roughness suitable for the water level and flow change process of the river channel can be obtained in the embodiment of the present invention, and finally the accurate simulation of the water level and flow change process of the river channel can be achieved.

BRIEF DESCRIPTION OF THE DRAWINGS

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

FIG1 is a flow chart of a specific example of a method for optimizing hydrodynamic model parameters in an embodiment of the present invention;

FIG2 is a flow chart of a specific example of a method for simulating a water level flow rate change process according to an embodiment of the present invention;

FIG3 is a schematic diagram of the topological structure of a Bushi river section according to an embodiment of the present invention;

FIG4 is a principle block diagram of a specific example of a hydrodynamic model parameter optimization device according to an embodiment of the present invention;

FIG5 is a principle block diagram of a specific example of a device for simulating a water level and flow rate variation process according to an embodiment of the present invention;

FIG6 is a principle block diagram of a specific example of a computer device in an embodiment of the present invention.

Detailed ways

The technical solution of the present invention will be described clearly and completely below in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

In the description of the present invention, it should be noted that the terms “first”, “second” and “third” are only used for descriptive purposes and cannot be understood as indicating or implying relative importance.

In the description of the present invention, it should be noted that, unless otherwise clearly specified and limited, the terms "installed", "connected", and "connected" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be indirectly connected through an intermediate medium, it can also be the internal connection of two components, it can be a wireless connection, or it can be a wired connection. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

In addition, the technical features involved in the different embodiments of the present invention described below can be combined with each other as long as they do not conflict with each other.

The embodiment of the present invention provides a method for optimizing parameters of a hydrodynamic model, as shown in FIG1 , comprising the following steps:

Step S11: An optimization objective function is established by combining the first initial neural network model, the second initial neural network model, and the hydrodynamic model. The water and sand parameters in the hydrodynamic model are determined by the first initial neural network model, and the roughness in the hydrodynamic model is determined by the second initial neural network model. The optimization objective function is determined based on the sum of the simulation residuals of each hydrodynamic model.

In an optional embodiment, the hydrodynamic model used to establish the optimization objective function is a one-dimensional hydrodynamic model, and the output of the first initial neural network model includes one or more of _Zs = _Zs (x,t; _θu ), _Qs = _Qs (x,t; _θu ), etc., wherein _Zs represents water level, _Qs represents flow, x represents space, t represents time, and _θu represents the first network model parameters. When solving the optimization objective function, the first network model parameters need to be optimized, and the space represented by x refers to the distance along the river direction in the one-dimensional coordinate system established by the model.

In an optional embodiment, the output of the second initial neural network model includes n _js =n _js (x; θ _p ), wherein n _js represents the roughness of the j-th river section, θ _p represents the second network model parameters, and when solving the optimization objective function, the second network model parameters need to be optimized.

Step S12: Solve the optimization objective function, optimize the first network model parameters and the second network model parameters, and obtain the first network model optimization parameters and the second network model optimization parameters that minimize the value of the optimization objective function, the first network model parameters are the parameters in the first initial neural network model, and the second network model parameters are the parameters in the second initial neural network model.

In an optional embodiment, a classical simulated annealing method for finding the global optimum may be used to solve the optimization objective function, thereby obtaining the first network model optimization parameters and the second network model optimization parameters.

Step S13: determining the second initial neural network model including the second network model optimization parameters as the roughness optimization model.

In the hydrodynamic model parameter optimization method provided in the embodiment of the invention, since the optimization objective function is determined according to the sum of the simulation residuals of the hydrodynamic model, wherein the roughness in the hydrodynamic model is determined by the second initial neural network model, the second network model optimization parameters of the second neural network model determined in the process of solving the optimization objective function can minimize the sum of the simulation residuals of the hydrodynamic model. It can be seen that the method provided in the embodiment of the invention optimizes the second initial neural network model on the basis of the strong constraints of the physical process, and the roughness determined by the roughness optimization model optimized by the method provided in the embodiment of the invention is more in line with the actual physical process. In addition, the method provided in the embodiment of the invention can directly optimize the objective function using a gradient optimization algorithm by placing the control equation of the hydrodynamic model into the optimized objective function without iterating the equation. At the same time, a physical driving term (i.e., the simulation residuals of each hydrodynamic model) is introduced into the objective function, so that the optimization algorithm requires less observation data. The method can obtain the optimal solution of the roughness parameter with physical significance without solving the control equation, which can effectively improve the simulation efficiency and accuracy of the river water level and flow change process.

In an optional implementation embodiment, the simulation residuals of the hydrodynamic model used in constructing the optimization objective function include the residuals of the water flow continuity equation and the residuals of the water flow motion equation, that is, the optimization objective function is established based on the sum of the residuals of the water flow continuity equation and the residuals of the water flow motion equation.

In an optional embodiment, the simulation residual of the hydrodynamic model when establishing the optimization objective function includes:

In an optional embodiment, the optimization objective function also includes an approximation error e ₃ =Q _s -Q ^* of the flow output by the first initial neural network model to the actual flow, and an approximation error e ₄ =Z _s -Z ^* of the water level output by the first initial neural network model to the actual water level, wherein Q _s represents the flow output by the first initial neural network model, Q ^* represents the actual flow, Z _s represents the water level output by the first initial neural network model, and Z ^* represents the actual water level, and the actual flow and the actual water level are obtained through observation.

In an optional embodiment, the optimization objective function is:

In an embodiment of the present invention, the optimization objective function is solved until the first network model optimization parameters and the second network model optimization parameters are obtained, so that the outputs of the two neural network models satisfy the control equations as much as possible and are as close to the observed data as possible.

In an optional embodiment, since the water levels and cross-sectional conditions of different river sections in the same river channel are different, the roughness is naturally different. Therefore, even for the same river channel, it is necessary to divide the river channel into multiple river sections. For each river section, the method provided in the above embodiment is executed respectively to obtain the roughness optimization model corresponding to each river section.

The embodiment of the present invention provides a method for simulating a water level flow change process, as shown in FIG2 , comprising:

Step S21: Divide the target river into multiple river sections, and obtain water level and flow data of each river section.

In an optional embodiment, the water level and flow data include the water level and flow of the water body.

In an optional embodiment, the water level and flow data are obtained based on monitoring stations established on the main and tributary rivers.

In an optional embodiment, based on the historical cross-sectional data of the main and tributary rivers along the target river channel, the changes in cross-sectional morphology caused by sediment deposition are analyzed, and the target river channel is divided into multiple river sections based on the analysis results.

In an optional embodiment, the target river channel may be divided according to the cross-sectional width of each section of the target river channel, and sections with similar cross-sectional widths may be divided into the same river section. A schematic diagram of the topological structure of the distributed river section is shown in FIG3 .

Step S22: determining the roughness of each river section based on the roughness optimization model corresponding to each river section, wherein the roughness optimization model is determined according to the hydrodynamic model parameter optimization method provided in the above embodiment.

In the embodiment of the present invention, since the cross-section of each river section is similar, an approximate roughness can be selected in the same river section.

Step S23: inputting the water level flow data and roughness of each river channel into the hydrodynamic model respectively to obtain the water level flow change process of the target river channel.

In the embodiment of the present invention, the water level flow rate data of the river section at the previous moment is input into the hydrodynamic model corresponding to the river section, and the water level flow rate and sediment content of the river section at the current moment can be obtained. The water level flow rate and sediment content of the river section at the current moment are input into the hydrodynamic model corresponding to the adjacent downstream river section, and the water level flow rate and sediment content of the river section at the next moment can be predicted. The water level, flow and sediment content of the target river can be predicted by analogy.

The hydrodynamic model used in the embodiment of the present invention is the same as the hydrodynamic model used in the above step S11.

The method for simulating the water level and flow change process provided in the embodiment of the present invention divides the target river channel into multiple river sections, determines the roughness of each river section according to the roughness optimization model corresponding to each river section, and inputs the water level and flow data and the roughness of each river channel into the hydrodynamic model respectively to obtain the water level and flow change process of the target river channel. Since the roughness optimization model used in the embodiment of the present invention optimizes the second initial neural network model based on the strong constraints of the physical process, the roughness suitable for the water level and flow change process of the river channel can be obtained in the embodiment of the present invention, and finally the accurate simulation of the water level and flow change process of the river channel can be achieved.

In an optional embodiment, a finite difference method is used to solve the hydrodynamic model to obtain the water level and flow change process of the target river channel.

In an optional embodiment, since the calculation of each river section is relatively independent, the calculation result of the upstream river section at the previous moment is the input condition for the numerical calculation of the adjacent river section at the current moment. By simulating the long-term history of water level and flow in multiple river sections, the water level and flow and the water level and flow change process of the entire river channel can be obtained.

The embodiment of the present invention provides a hydrodynamic model parameter optimization device, as shown in FIG4 , comprising:

The optimization objective function establishment module 11 is used to establish the optimization objective function in combination with the first initial neural network model, the second initial neural network model, and the hydrodynamic model. The water and sand parameters in the hydrodynamic model are determined by the first initial neural network model, and the roughness in the hydrodynamic model is determined by the second initial neural network model. The optimization objective function is determined based on the sum of the simulation residuals of each hydrodynamic model. For details, please refer to the description of step S11 in the above embodiment and will not be repeated here.

The network parameter optimization module 12 is used to solve the optimization objective function, optimize the first network model parameters and the second network model parameters, and obtain the first network model optimization parameters and the second network model optimization parameters that minimize the value of the optimization objective function. The first network model parameters are the parameters in the first initial neural network model, and the second network model parameters are the parameters in the second initial neural network model. For details, please refer to the description of step S12 in the above embodiment and will not be repeated here.

The optimization model determination module 13 is used to determine the second initial neural network model including the second network model optimization parameters as the roughness optimization model. The details are described in the description of step S13 in the above embodiment and will not be repeated here.

The embodiment of the present invention provides a water level flow rate change process simulation device, as shown in FIG5 , comprising:

The data acquisition module 21 is used to divide the target river into multiple river sections and obtain the water level and flow data of each river section. The details are described in the above embodiment of step S21 and will not be repeated here.

The parameter determination module 22 is used to determine the roughness of each river section based on the roughness optimization model corresponding to each river section. The roughness optimization model is determined according to the hydrodynamic model parameter optimization method provided in the above embodiment. For details, please refer to the description of step S22 in the above embodiment and will not be repeated here.

The simulation module 23 is used to input the water level flow data and roughness of each river channel into the hydrodynamic model to obtain the water level flow change process of the target river channel. The details are described in the description of step S23 in the above embodiment and will not be repeated here.

An embodiment of the present invention provides a computer device, as shown in FIG6 , the computer device mainly includes one or more processors 31 and a memory 32 , and FIG6 takes one processor 31 as an example.

The computer device may further include: an input device 33 and an output device 34 .

The processor 31, the memory 32, the input device 33 and the output device 34 may be connected via a bus or other means, and FIG6 takes the connection via a bus as an example.

The processor 31 may be a central processing unit (CPU). The processor 31 may also be other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, or a combination of the above chips. The general-purpose processor may be a microprocessor or the processor may be any conventional processor. The memory 32 may include a program storage area and a data storage area, wherein the program storage area may store an operating system and application programs required for at least one function; the data storage area may store data created according to the use of a hydrodynamic model parameter optimization device or a water level flow change process simulation device. In addition, the memory 32 may include a high-speed random access memory, and may also include a non-transient memory, such as at least one disk storage device, a flash memory device, or other non-transient solid-state storage devices. In some embodiments, the memory 32 may optionally include a memory remotely arranged relative to the processor 31, and these remote memories may be connected to the hydrodynamic model parameter optimization device, or the water level flow change process simulation device through a network. The input device 33 may receive a calculation request (or other digital or character information) input by a user, and generate key signal input related to the hydrodynamic model parameter optimization device, or the water level flow change process simulation device. The output device 34 may include a display device such as a display screen to output the calculation results.

The embodiment of the present invention provides a computer-readable storage medium, which stores computer instructions. The computer storage medium stores computer executable instructions, which can execute the hydrodynamic model parameter optimization method in any of the above method embodiments, or the water level flow change process simulation method. Among them, the storage medium can be a disk, an optical disk, a read-only memory (ROM), a random access memory (RAM), a flash memory (Flash Memory), a hard disk (Hard Disk Drive, abbreviated: HDD) or a solid-state drive (SSD), etc.; the storage medium can also include a combination of the above types of memory.

Obviously, the above embodiments are merely examples for clear explanation, and are not intended to limit the implementation methods. For those skilled in the art, other different forms of changes or modifications can be made based on the above description. It is not necessary and impossible to list all the implementation methods here. The obvious changes or modifications derived therefrom are still within the protection scope of the invention.

Claims

A method for optimizing hydrodynamic model parameters, characterized in that it comprises the following steps:

An optimization objective function is established by combining the first initial neural network model, the second initial neural network model, and the hydrodynamic model, wherein the water and sand parameters in the hydrodynamic model are determined by the first initial neural network model, the roughness in the hydrodynamic model is determined by the second initial neural network model, and the optimization objective function is determined according to the sum of the simulation residuals of each hydrodynamic model;

Solving the optimization objective function, optimizing the first network model parameters and the second network model parameters, and obtaining the first network model optimization parameters and the second network model optimization parameters that minimize the value of the optimization objective function, wherein the first network model parameters are parameters in the first initial neural network model, and the second network model parameters are parameters in the second initial neural network model;

The second initial neural network model including the second network model optimization parameters is determined as the roughness optimization model.
The method for optimizing hydrodynamic model parameters according to claim 1, characterized in that:

The simulation residuals of the hydrodynamic model include the residuals of the water flow continuity equation and the residuals of the water flow motion equation.
The method for optimizing hydrodynamic model parameters according to claim 1 or 2 is characterized in that the water and sediment parameters determined by the first initial neural network model include flow and water level,

The optimization objective function also includes an approximation error between the flow output by the first initial neural network model and the actual flow, and an approximation error between the water level output by the first initial neural network model and the actual water level.
The method for optimizing the parameters of a hydrodynamic model according to claim 1 or 2, wherein the simulation residual of the hydrodynamic model comprises:

Among them, e1 represents the residual of the water flow continuity equation, e2 represents the residual of the water flow motion equation, B represents the water surface width, Zs represents the water level, Qs represents the flow rate, t represents time, x represents space, qL represents the lateral inflow flow rate per unit river length, A represents the water cross-sectional area, g represents the gravitational acceleration, njs represents the roughness, and R represents the hydraulic radius. Among them, Zs = Zs (x, t; θu ), Qs = Qs (x, t; θu ), njs = njs (x; θp ), θu represents the first network model parameters, and θp represents the second network model parameters.
A method for simulating a water level flow change process, characterized by comprising:

Divide the target river into multiple river sections and obtain water level and flow data of each river section;

Determine the roughness of each river section based on a roughness optimization model corresponding to each river section, wherein the roughness optimization model is determined by the hydrodynamic model parameter optimization method according to any one of claims 1 to 4;

The water level flow data and the roughness of each river channel are respectively input into the hydrodynamic model to obtain the water level flow change process of the target river channel.
The method for simulating water level and flow rate changes according to claim 5 is characterized in that:

The hydrodynamic model input data of the river section at the current moment also includes the hydrodynamic model output data of the adjacent upstream river section at the previous moment.
A hydrodynamic model parameter optimization device, characterized by comprising:

The optimization objective function establishment module is used to establish the optimization objective function by combining the first initial neural network model, the second initial neural network model, and the hydrodynamic model. The water and sand parameters in the hydrodynamic model are obtained by the first initial neural network model. The roughness in the hydrodynamic model is determined by the second initial neural network model, and the optimization objective function is determined according to the sum of the simulation residuals of each hydrodynamic model;

A network parameter optimization module, used to solve the optimization objective function, optimize the first network model parameters and the second network model parameters, and obtain the first network model optimization parameters and the second network model optimization parameters that minimize the value of the optimization objective function, wherein the first network model parameters are the parameters in the first initial neural network model, and the second network model parameters are the parameters in the second initial neural network model;

The optimization model determination module is used to determine the second initial neural network model including the optimization parameters of the second network model as the roughness optimization model.
A water level flow change process simulation device, characterized in that it comprises:

The data acquisition module is used to divide the target river into multiple river sections and obtain the water level and flow data of each river section;

A parameter determination module, used for determining the roughness of each river section based on a roughness optimization model corresponding to each river section, wherein the roughness optimization model is determined according to the hydrodynamic model parameter optimization method according to any one of claims 1 to 4;

The simulation module is used to input the water level flow data and the roughness of each river channel into the hydrodynamic model to obtain the water level flow change process of the target river channel.