CN115293037A - Hydrodynamic modeling method of river and lake composite system based on machine learning - Google Patents

Hydrodynamic modeling method of river and lake composite system based on machine learning Download PDF

Info

Publication number
CN115293037A
CN115293037A CN202210926817.6A CN202210926817A CN115293037A CN 115293037 A CN115293037 A CN 115293037A CN 202210926817 A CN202210926817 A CN 202210926817A CN 115293037 A CN115293037 A CN 115293037A
Authority
CN
China
Prior art keywords
river
lake
flow
machine learning
boundary
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
CN202210926817.6A
Other languages
Chinese (zh)
Inventor
黄绳
夏军
王月玲
佘敦先
王纲胜
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Wuhan University WHU
Institute of Geographic Sciences and Natural Resources of CAS
Original Assignee
Wuhan University WHU
Institute of Geographic Sciences and Natural Resources of CAS
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Wuhan University WHU, Institute of Geographic Sciences and Natural Resources of CAS filed Critical Wuhan University WHU
Priority to CN202210926817.6A priority Critical patent/CN115293037A/en
Publication of CN115293037A publication Critical patent/CN115293037A/en
Pending legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F30/00—Computer-aided design [CAD]
    • G06F30/20—Design optimisation, verification or simulation
    • G06F30/28—Design optimisation, verification or simulation using fluid dynamics, e.g. using Navier-Stokes equations or computational fluid dynamics [CFD]
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F30/00—Computer-aided design [CAD]
    • G06F30/20—Design optimisation, verification or simulation
    • G06F30/27—Design optimisation, verification or simulation using machine learning, e.g. artificial intelligence, neural networks, support vector machines [SVM] or training a model
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F2113/00—Details relating to the application field
    • G06F2113/08—Fluids

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Evolutionary Computation (AREA)
  • Geometry (AREA)
  • General Engineering & Computer Science (AREA)
  • Computer Hardware Design (AREA)
  • Software Systems (AREA)
  • Medical Informatics (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Artificial Intelligence (AREA)
  • Algebra (AREA)
  • Computing Systems (AREA)
  • Fluid Mechanics (AREA)
  • Mathematical Analysis (AREA)
  • Mathematical Optimization (AREA)
  • Mathematical Physics (AREA)
  • Pure & Applied Mathematics (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)

Abstract

The embodiment of the application provides a hydrodynamic modeling method of a river and lake composite system based on machine learning, wherein the boundary condition of a one-dimensional hydrodynamic model provides input for the machine learning; the output of the machine learning provides complex boundary conditions and downstream boundary conditions at river and lake junctions for the one-dimensional hydrodynamic model. The method improves the operating efficiency of hydrodynamic simulation of the river and lake composite system, can quickly calculate the water level and the flow of the important section position of the river and lake composite system, simultaneously reduces the requirement on terrain data, does not need to input complex lake terrain elevation survey data, and can effectively improve the water level simulation precision of the section near the downstream boundary.

Description

Hydrodynamic modeling method of river and lake composite system based on machine learning
Technical Field
The application relates to the field of hydrodynamic modeling, can also be used in the field of machine learning, and particularly relates to a method for hydrodynamic modeling of a river and lake composite system based on machine learning.
Background
Machine learning is currently an important tool for studying watershed hydrological simulations. When machine learning is used for hydrological modeling, rainfall or upstream flow in a watershed is generally used as an input, and downstream water level or flow is used as an output. The parameters of the machine learning model are trained according to the measured data, and the model for completing parameter training can simulate and predict the water level and the flow of rivers or lakes in a river area. Machine learning is a hydrographic hydrodynamic simulation method with high running speed and high precision, but can only be used for sequentially simulating a single station, and is difficult to model hydrographic hydrodynamic elements of each section of a large-range river reach.
The hydrodynamic numerical simulation of the river and lake composite system can establish a one-dimensional, two-dimensional or three-dimensional physical hydrodynamic model. The choice of different model dimensions is essentially a trade-off between solution accuracy and computational cost, and is also limited by data availability. For river simulation in a long river section and a long period, the one-dimensional model along the flow direction of the river channel is adopted, so that the calculated amount is greatly reduced. The lake area cannot be generalized by a one-dimensional model because the length and the width in the horizontal direction are both large, and a two-dimensional model or a three-dimensional model which is averaged along the water depth direction is often adopted. However, for large lakes, high-resolution bathymetric data or digital elevation data required for constructing a two-dimensional or three-dimensional model are difficult to obtain, and the problems of low model operation efficiency, repeated iteration of water level flow at a coupling boundary, easy calculation instability and the like exist. On the premise of ensuring the calculation accuracy, the lake model is reasonably generalized to reduce the calculation cost, and the method is a major challenge for simulating the hydromechanical and hydrodynamic process of a river and lake composite system. In addition, in a changing environment without hydraulic engineering regulation, the downstream boundary conditions of the river and lake composite system are unknown. The usually adopted water level-flow relation curve method is often unstable due to the interference of various factors such as terrain, vegetation and the like, and the accuracy of hydrodynamic simulation of a river and lake composite system is easily influenced.
The machine learning has the advantage of exploring an implicit relation in a nonlinear system, and a new thought can be provided for solving complex problems in hydrodynamic modeling of a river and lake composite system by coupling the machine learning and a hydrodynamic method. Practical application and intensive research show that the current river and lake composite system modeling method related to machine learning and hydrodynamic force has some defects: (1) The pure machine learning method has limited calculation sites, and is difficult to model all units of the whole research area simultaneously; (2) The pure water power method is used for calculating the time consumption and has higher requirements on topographic data, particularly on lake parts; (3) The method for performing the alternative modeling by driving the machine learning by using the water power model has limited effect, needs to update parameters frequently according to terrain changes and the like, has poor expandability and does not fully combine the advantages of the machine learning and the water power method.
Disclosure of Invention
Aiming at the problems in the prior art, the hydrodynamic modeling method of the river and lake composite system based on machine learning is provided, and the operation efficiency of hydrodynamic simulation of the river and lake composite system can be effectively improved.
In order to solve at least one of the above problems, the present application provides the following technical solutions:
the application provides a hydrodynamic modeling method of a river and lake composite system based on machine learning, which comprises the following steps:
collecting flow data of upstream boundary, lateral boundary and river and lake intersection position stations of a research area and water level data of downstream boundary stations;
training and predicting the flow of the river and lake intersection position by adopting a machine learning method according to the flow of the upstream boundary station and the flow of the lateral branch station which possibly influences the water exchange between rivers and lakes;
training and predicting the water level of the downstream boundary by adopting a machine learning method according to the flow of the upstream and lateral boundary stations and the flow exchanged between rivers and lakes;
discretizing a research river reach according to the topographic data of the river course, dividing a section on the river course at intervals, and sorting the starting point distance and elevation data on the bank side on the section;
assigning the initial flow of each section as the flow of the upstream boundary in the first day, and performing linear interpolation calculation on the initial water level according to the water level of the downstream boundary in the first day and the river slope;
dispersing the holy-Venen equation set, and inputting initial and boundary conditions of each river channel section to perform one-dimensional hydrodynamic calculation of the river channel.
Further, the method for training and predicting the flow of the river and lake intersection position by adopting the machine learning method comprises the following steps:
the specific calculation paradigm adopted is:
Figure BDA0003779856860000021
wherein,
Figure BDA0003779856860000022
is the amount of water exchanged by the jth lake and river; q in Is the upstream water inflow; q B The river side branch flow is adopted, and the water quantity exchanged between the jth lake and the river is influenced by i river branches;
Figure BDA0003779856860000023
the flow of the branch born by the jth lake per se is determined, and k lake branches are shared; f (g) represents a preset machine learning method.
Further, the training and predicting the water level of the downstream boundary by adopting the machine learning method comprises the following steps:
the adopted downstream boundary water level concrete calculation paradigm is as follows:
Figure BDA0003779856860000031
wherein Z is a downstream boundary water level; q in Is the upstream water inflow; q B Is the flow of the side branch of the river; q L Is the amount of water exchanged between the lake and the river; m and n are the number of tributaries connected with the river and the number of lakes in the river; f (g) generationThe table presets a machine learning method.
According to the technical scheme, the hydrodynamic modeling method of the river and lake composite system based on machine learning can improve the operation efficiency of hydrodynamic simulation of the river and lake composite system, quickly calculate the water level and the flow of the important section position of the river and lake composite system, reduce the requirement on terrain data, avoid inputting complex lake terrain elevation survey data, and effectively improve the water level simulation precision of sections near the downstream boundary.
Drawings
In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings needed to be used in the description of the embodiments or the prior art will be briefly introduced below, and it is obvious that the drawings in the following description are some embodiments of the present application, and it is obvious for those skilled in the art to obtain other drawings based on these drawings without creative efforts.
Fig. 1 is a schematic flow chart of a hydrodynamic modeling method of a river and lake composite system based on machine learning in an embodiment of the application;
FIG. 2 is a schematic diagram illustrating measured and simulated water levels and flow rates at a primary site according to an embodiment of the present application;
FIG. 3 is a schematic diagram illustrating a comparison between measured and simulated downstream boundary water levels in an embodiment of the present application.
Detailed Description
In order to make the objects, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be described clearly and completely with reference to the drawings in the embodiments of the present application, and it is obvious that the described embodiments are some embodiments of the present application, but not all embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present application.
According to the technical scheme, the data acquisition, storage, use, processing and the like meet relevant regulations of national laws and regulations.
In order to effectively improve the operation efficiency of hydrodynamic simulation of a river and lake composite system, the present application provides an embodiment of a hydrodynamic modeling method of a river and lake composite system based on machine learning, and referring to fig. 1, the hydrodynamic modeling method of a river and lake composite system based on machine learning specifically includes the following contents:
step 1, hydrologic data collection: and collecting the flow data of the upstream boundary, the lateral boundary and the river and lake intersection position station of the research area and the water level data of the downstream boundary station.
In particular, according to data sources such as the hydrological annale certificate, upstream boundary flow (Yichang station flow), river branch flow (Hanjiang peach station flow), lake branch flow (Shimen station, taojiang station and Hunan pond station flow of the Changji lake, exzhou station of the Po Yanghu, li home station, mei harbor station, hushan station, weijin station, wanjia port and Du Peak pit station flow), river and lake rendezvous point flow (city mountain station flow of the Po Yanhu, lake outlet station flow of the Po Yanghu) and downstream boundary water level (Datong station) in 2007-2016 are collected, and all flow or hydrological data are Japanese scale hydrological data.
Step 2, calculating the amount of river and lake exchanged water: and training and predicting the flow of the intersection position of the rivers and the lakes by adopting a machine learning method according to the flow of the upstream boundary station and the flow of the lateral branch stations which possibly influence water exchange between the rivers and the lakes. The specific calculation paradigm is as follows:
Figure BDA0003779856860000041
wherein,
Figure BDA0003779856860000042
is the amount of water exchanged by the jth lake and river; q in The upstream water flow, namely the flow of the Yichang station; q B The flow of the side branch of the river is adopted, and the water quantity exchanged between the jth lake and the river is influenced by i river branches;
Figure BDA0003779856860000043
the flow of the branch born by the jth lake per se is determined, and k lake branches are shared; RNN (g) represents a machine learning method called a recurrent neural network.
Taking the flow at the junction between the rivers and the lakes in 2007-2013 as the output of machine learning, taking the upstream boundary flow (Yichang station flow), the branch flows of the corresponding lakes (the flux of a Shimen station, a Taojiang station, a peach source station and a Hunan pool station of the Dongting lake, the flux of an Exzhou station, a Li home ferry station, a Meigang station, a Hushan station, a Yangtze station, a Wanjia port station and a Dufeng pit station of the Po Yangtze lake) and the upstream branch flows of rivers capable of influencing the junction between the rivers and the lakes (for example, the lake mouth station at the junction between the Po Yangtze lake and the Changjiang river trunk flow is influenced by the flux of the Hanjiang river) as the input of machine learning, and training the parameters of a machine learning model. The flow at the river-lake junction in 2014-2016 (flow at the Chengling rock station in Dongting lake, flow at the lake outlet in Poyang lake) was predicted using an RNN model with trained parameters.
And 3, calculating the downstream boundary water level: and (4) training and predicting the water level of the downstream boundary by adopting a machine learning method according to the flow of the upstream and lateral boundary stations and the flow exchanged between rivers and lakes. The specific calculation paradigm is as follows:
Figure BDA0003779856860000051
wherein Z is a downstream boundary water level, namely a large communication station water level; q in The upstream inflow amount, namely Yichang flow; q B The river side branch flow is the flow of the Hanjiang peach station; q L Is the amount of water exchanged between the lake and the river; m and n are the number of tributaries connected with rivers and the number of lakes in the river, and are 1 and 2 respectively; RNN (g) represents a machine learning method called a recurrent neural network.
And (3) taking the water level of a downstream boundary station (Datong station water level) in 2007-2013 as a target variable of the machine learning model, and taking upstream boundary flow (Yichang station flow), all upstream river branch flow (Hanjiang Xiantao station flow) and lake in-out river flow (Chengling rock station flow and lake outlet station flow) as the input of the machine learning model, and training parameters of the machine learning model. And (5) predicting the water level of the boundary station (water level of the Datong station) in the downstream of 2014-2016 by using an RNN model with trained parameters.
Step 4, dividing river channel sections: discretizing the research river reach according to the topographic data of the river reach, and dividing a section on the river reach at intervals. 535 sections are divided, and the distance between the sections is 1-5 kilometers. And (5) arranging the shoreside starting point distance and the elevation data on the sections.
Step 5, setting initial conditions: assigning initial flow of 535 sections as the flow of the first day (1 month and 1 day in 2014) of the upstream boundary Yichang station, and performing linear interpolation calculation on the initial water level according to the water level and the river slope of the first day (1 month and 1 day in 2014) of the downstream boundary Datong station;
and 6, dispersing a Saint-Venant equation set (Saint-Venant equation set, namely a basic equation of the one-dimensional hydrodynamic model) in the following form:
Figure BDA0003779856860000052
in the formula: a is the area, m; t is time, s; q is flow, m 3 S; x is a process, m; q is the lateral flow per unit flow path, positive values indicate inflow, m 2 S; g is gravity acceleration, and is 9.81m/s 2 (ii) a B is the water surface width m; z is water level, m; n is roughness and dimensionless; r is wet week, m.
And solving a one-dimensional hydrodynamic model Saint-Vinem equation set according to the boundary conditions obtained in the step 1-3 and the initial conditions obtained in the step 5. The results of the water level and flow simulation of the major river and lake sites are shown in fig. 2. The simulation and actual measurement values of the downstream boundary water level are shown in fig. 3. The method well simulates the water level and the flow rate of the river and lake composite system, and is suitable for hydrodynamic modeling of the river and lake composite system. In addition, the method can simulate the water level and the flow of a research area within 1 year in 9.4 minutes, and compared with the traditional two-dimensional coupling hydrodynamic model which takes 6 hours for simulating the same area and the same time, the operation efficiency of the River and lake composite system hydrodynamic model is improved by 38 times (Xijun Lai, jianhu Jiang, qiahua Liang, qun Huang,2013, large-scale hydraulic model of the middle Yangzze River Basin with complex-like interfaces, journal of Hydrology,492, pp.228-243).
From the above description, the hydrodynamic modeling method of the river and lake composite system based on machine learning provided by the embodiment of the application can improve the operation efficiency of hydrodynamic simulation of the river and lake composite system through flexible modeling, quickly calculate the water level and flow of the important cross section position of the river and lake composite system, reduce the requirement on terrain data, avoid inputting complex survey data of lake terrain elevation, and effectively improve the water level simulation accuracy of the cross section near the downstream boundary.
The principle and the implementation mode of the invention are explained by applying specific embodiments in the invention, and the description of the embodiments is only used for helping to understand the method and the core idea of the invention; meanwhile, for a person skilled in the art, according to the idea of the present invention, there may be variations in the specific embodiments and the application scope, and in summary, the content of the present specification should not be construed as a limitation to the present invention.

Claims (3)

1. A hydrodynamic modeling method of a river and lake composite system based on machine learning is characterized by comprising the following steps:
collecting flow data of upstream boundary, lateral boundary and river and lake intersection position stations of a research area and water level data of downstream boundary stations;
training and predicting the flow of the intersection position of the rivers and the lakes by adopting a machine learning method according to the flow of the upstream boundary station and the flow of the lateral branch stations which possibly influence the water exchange between the rivers and the lakes;
training and predicting the water level of the downstream boundary by adopting a machine learning method according to the flow of the upstream and lateral boundary stations and the flow exchanged between rivers and lakes;
discretizing a research river reach according to the topographic data of the river course, dividing a section on the river course at intervals, and sorting the starting point distance and elevation data on the bank side on the section;
assigning the initial flow of each section as the flow of the upstream boundary in the first day, and performing linear interpolation calculation on the initial water level according to the water level of the downstream boundary in the first day and the river slope;
dispersing the holy-Venen equation set, and inputting initial and boundary conditions of each river channel section to perform one-dimensional hydrodynamic calculation of the river channel.
2. The machine learning-based hydrodynamic modeling method for a river and lake composite system according to claim 1, wherein the training and predicting the flow of the river and lake junction position by using the machine learning method comprises:
the specific calculation paradigm adopted is:
Figure FDA0003779856850000011
wherein,
Figure FDA0003779856850000012
is the amount of water exchanged between the jth lake and the river; q in Is the upstream water inflow; q B The flow of the side branch of the river is adopted, and the water quantity exchanged between the jth lake and the river is influenced by i river branches;
Figure FDA0003779856850000013
the flow of the branch born by the jth lake per se is determined, and k lake branches are shared; f (g) represents a preset machine learning method.
3. The machine learning based river and lake complex system hydrodynamic modeling method according to claim 1, wherein the training and predicting the water level of the downstream boundary using the machine learning method comprises:
the adopted downstream boundary water level specific calculation paradigm is as follows:
Figure FDA0003779856850000014
wherein Z is a downstream boundary water level; q in Is the upstream water inflow; q B The side branch flow of the river; q L Is the amount of water exchanged between the lake and the river; m and n are the number of tributaries connected with the river and the number of lakes in the river; f (g) represents a preset machine learning method.
CN202210926817.6A 2022-08-03 2022-08-03 Hydrodynamic modeling method of river and lake composite system based on machine learning Pending CN115293037A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202210926817.6A CN115293037A (en) 2022-08-03 2022-08-03 Hydrodynamic modeling method of river and lake composite system based on machine learning

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202210926817.6A CN115293037A (en) 2022-08-03 2022-08-03 Hydrodynamic modeling method of river and lake composite system based on machine learning

Publications (1)

Publication Number Publication Date
CN115293037A true CN115293037A (en) 2022-11-04

Family

ID=83825852

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202210926817.6A Pending CN115293037A (en) 2022-08-03 2022-08-03 Hydrodynamic modeling method of river and lake composite system based on machine learning

Country Status (1)

Country Link
CN (1) CN115293037A (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115496015A (en) * 2022-11-18 2022-12-20 珠江水利委员会珠江水利科学研究院 Hydrodynamic analysis decision method based on flow gradient change

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115496015A (en) * 2022-11-18 2022-12-20 珠江水利委员会珠江水利科学研究院 Hydrodynamic analysis decision method based on flow gradient change
CN115496015B (en) * 2022-11-18 2023-02-28 珠江水利委员会珠江水利科学研究院 Hydrodynamic analysis decision method based on flow gradient change

Similar Documents

Publication Publication Date Title
CN105893763B (en) Southern tidal rivers network flow water quality coupled simulation system
Liang et al. A reduced-complexity model for river delta formation–Part 1: Modeling deltas with channel dynamics
Abbott et al. An introduction to the European Hydrological System—Systeme Hydrologique Europeen,“SHE”, 1: History and philosophy of a physically-based, distributed modelling system
Mizukami et al. mizuRoute version 1: a river network routing tool for a continental domain water resources applications
CN113569438B (en) Urban flood model construction method based on multisource rainfall fusion and real-time correction
CN107590354A (en) Basin rainwash runoff concentration simulation method based on stationary flow field
CN106498896B (en) A kind of reservoir operation process analogy method being coupled with Sedimentation
CN104933268B (en) A kind of flood risk analysis method based on d Unsteady Water Flow numerical model
CN112784502B (en) Hydrologic-hydraulic dynamic bidirectional coupling flood prediction method
CN110362925A (en) A kind of unrestrained top flood due to dam-break method for numerical simulation of the earth and rockfill dam comprising reservoir area
CN112464584B (en) Method for estimating water level and flow of free surface flow
CN110135069B (en) Method and device for acquiring silt characteristics during water delivery of water delivery tunnel and computer equipment
CN112651118B (en) Full-coupling simulation method for climate-land-hydrologic process
Isaac et al. Sediment management studies of a run-of-the-river hydroelectric project using numerical and physical model simulations
CN116151152B (en) Hydrodynamic force numerical simulation calculation method based on gridless calculation
Zhang et al. Integrated hydrodynamic model for simulation of river-lake-sluice interactions
CN115293037A (en) Hydrodynamic modeling method of river and lake composite system based on machine learning
CN115659859A (en) Flood simulation calculation method for river network structure containing flood area
CN114925624A (en) Natural river channel three-dimensional water flow numerical simulation method
CN113158556B (en) Short-time high-precision forecasting method for regional water level
Zhang et al. Numerical simulation of flow and suspended sediment transport in the distributary channel networks
Shaad Development of a distributed surface-subsurface interaction model for river corridor hydrodynamics
Balas et al. Hydrodynamic Modelling Using HİDROTÜRK Model
Luijendijk et al. Pilot application of Delft3D Flexible Mesh: Assisting a field campaign at the Sand Engine
CN117893179B (en) Digital twinning-based hydraulic power engineering construction management and control method and system

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination