Total heat balance natural water temperature simulation method based on GPU acceleration

By using a GPU-accelerated full thermal equilibrium method, combined with DEM topographic data and hydrodynamic control equations, the problems of accuracy and speed in simulating natural water temperature in rivers, lakes, and seas were solved, achieving efficient and high-precision water temperature simulation and visualization analysis.

CN119047366APending Publication Date: 2024-11-29XIAN UNIV OF TECH
View PDF 0 Cites 2 Cited by

Patent Information

Application Number
CN202411091208.9
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-08-09
Publication Date
2024-11-29

AI Technical Summary

Technical Problem

Existing technologies are insufficient for accurately simulating the dynamic transport process of natural water temperature in rivers, lakes, and seas, and the simulation speed is relatively slow.

Method used

A GPU-accelerated full thermal equilibrium natural water temperature simulation method is adopted. A two-dimensional water temperature convection-diffusion transport equation is constructed using high-resolution DEM topographic data and coupled with hydrodynamic control equations. The flux term is solved using the central finite volume method and HLLC Riemann solver. Considering thermal equilibrium and water vapor exchange, the hydrostatic reconstruction method is used to handle the wet and dry boundary. GPU acceleration technology is introduced to improve computational efficiency.

Benefits of technology

It achieves efficient and high-precision simulation of natural water temperature in rivers, lakes and seas, and can refine and visualize the surface water temperature distribution in rivers, lakes and seas, accurately exploring the impact of water temperature on the aquatic ecological environment.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119047366A_ABST
    Figure CN119047366A_ABST
Patent Text Reader

Abstract

The invention discloses a total heat balance natural water temperature simulation method based on GPU acceleration, and the method comprises the steps: firstly representing the real terrain of a river, lake and sea area through a high-resolution DEM terrain, and taking the real terrain as model input terrain data; assigning values to the related parameters; constructing a two-dimensional natural water temperature convection diffusion transport equation; secondly, coupling a hydrodynamic control equation with a constructed two-dimensional natural water temperature convection diffusion transport equation, effectively processing a dry and wet boundary problem, and realizing space-time second-order precision of the model; and finally, limiting the time stability by using a CFL condition. And outputting the natural water temperature transportation process of the research area. The device can accurately simulate the dynamic transportation process of natural water temperature in rivers, lakes and sea areas.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention belongs to the technical field of natural water temperature simulation in rivers, lakes and seas, and specifically relates to a full thermal balance natural water temperature simulation method based on GPU acceleration. Background Art

[0002] As an important physical characteristic of surface water, water temperature directly affects the physical, chemical and biological processes in river, lake and sea ecosystems, and becomes one of the key parameters that determine whether the aquatic ecosystem is healthy. Water temperature that is too high or too low will be detrimental to the metabolism, growth, development and reproduction of aquatic species, and may even lead to drastic changes in the biological community and population structure in the aquatic ecosystem. Previous studies on natural water temperature usually took descriptive as the basic state, and could not predict the development trend of water temperature. At the same time, limited by the development level of water temperature measurement technology, it is difficult to obtain model data, resulting in low accuracy of the water temperature model. Therefore, it is particularly important to construct a high-precision natural water temperature model to explore the impact of water temperature on the water ecological environment. At present, with the continuous development of water temperature measurement technology, many high-precision water temperature monitoring technologies have emerged, which makes it possible for deterministic water temperature models to accurately simulate and predict complex flow natural water temperatures. Based on this, a full thermal balance natural water temperature simulation method based on GPU acceleration is proposed. This method uses the principle of total heat balance and establishes a numerical model through the water quality and heat balance principle to simulate the natural water temperature transfer law in rivers, lakes and seas; at the same time, GPU acceleration technology is introduced to further improve the efficiency of model simulation. This method provides an important simulation tool for accurately and efficiently simulating the natural water temperature transfer process in rivers, lakes and seas, so as to more accurately explore the impact of water temperature on the water ecological environment. Summary of the invention

[0003] The purpose of the present invention is to provide a full thermal balance natural water temperature simulation method based on GPU acceleration, which can accurately simulate the dynamic transfer process of natural water temperature in rivers, lakes and seas.

[0004] The technical solution adopted by the present invention is a full thermal balance natural water temperature simulation method based on GPU acceleration, which is specifically implemented according to the following steps:

[0005] Step 1: Use high-resolution DEM terrain to characterize the real terrain of rivers, lakes and seas, and use this as terrain data for model input;

[0006] Step 2: Assign values ​​to relevant parameters;

[0007] Step 3, construct a two-dimensional natural water temperature convection-diffusion transport equation;

[0008] Step 4, coupling the hydrodynamic control equation with the constructed two-dimensional natural water temperature convection-diffusion transport equation to obtain a two-dimensional hydrodynamic-water temperature coupled control equation group, discretizing the above two-dimensional hydrodynamic-water temperature coupled control equation group, and solving the flux term, diffusion term and source term of the coupled control equation group;

[0009] Step 5: Meeting the internal balance of revenue and expenditure in rivers, lakes and seas;

[0010] Step 6: Effectively handle the dry-wet boundary problem and achieve the second-order spatiotemporal accuracy of the model;

[0011] Step 7: Use CFL conditions to limit the stability over time;

[0012] Step 8: Output the natural water temperature transport process in the study area.

[0013] The present invention is also characterized in that:

[0014] The two-dimensional natural water temperature convection-diffusion equation in step 3 is expressed as follows:

[0015]

[0016] Where T is the average water temperature increase along the water depth, in degrees Celsius; E x 、E y are the thermal diffusion coefficients in the x and y directions respectively; h is the water depth, in m; u and v are the flow velocities in the x and y directions, in m / s; t is the time, in s; is the net surface heat flux, in W / m 2 ; ρ is water density, unit: kg / m 3 , C p is the specific heat capacity flux of water, unit: W / (m 2 ·(kg·℃)), T i0 It is the source and sink term, and its unit is ℃.

[0017] The vector form of the two-dimensional hydrodynamic-water temperature coupling control equation in step 4 is as follows:

[0018]

[0019] in,

[0020]

[0021] Where t is time; Q is the variable vector, h is the water depth, unit: m, q x and q y They are the single width flow in the x and y directions, in m 2 / s; F, G are flux vectors in the x and y directions respectively; u, v are flow velocities in the x and y directions respectively, in units of m / s; S is the source term vector, including the bottom slope source term, friction source term, temperature diffusion term and temperature source term; z b is the riverbed bottom elevation, in m; C f is the friction coefficient of the bed surface, calculated as C f =gn 2 / h 1 / 3 , where n is the Manning coefficient; g is the acceleration due to gravity, unit: m / s 2 ; T is the average water temperature increase along the water depth, unit: ℃; E x 、E y are the thermal diffusion coefficients in the x and y directions respectively; h is the water depth in meters; is the net surface heat flux, in W / m 2 ; ρ is water density, unit: kg / m 3 , C p is the specific heat capacity flux of water, unit: W / (m 2 ·(kg·℃)), T i0 It is the source and sink term, and its unit is ℃.

[0022] Step 4 is as follows:

[0023] Step 4.1: For the i-th control unit grid, the central finite volume method is used to integrate the hydrodynamic-water temperature control equations and convert them into line integrals. The specific form is:

[0024]

[0025] Where Ω is the volume of control cell grid i; Γ is the boundary of control volume i; n is the unit vector in the direction of the external normal corresponding to the boundary Γ of the control volume; S is the source term vector, including the bottom slope source term S b , friction source term S f , temperature diffusion term and temperature source term, k is the number of the control unit grid edge; l k is the length of the kth side of the i-th control unit grid;

[0026] The flux term is solved using the HLLC Riemann solver, F(Q)·n is obtained as:

[0027]

[0028] Among them, the control volume unit grid interface intermediate flux F * for,

[0029]

[0030] Control volume unit grid interface flux intermediate wave sides, F * L and F * R for,

[0031]

[0032] In the formula, F L and F R are the fluxes on the left and right sides of the unit grid interface respectively; F * , and are the left wave flux, middle wave flux and right wave flux on both sides of the intermediate wave of the control volume unit grid interface flux respectively; u || is the tangential velocity along the interface, u || =-un y +vn x ,u ||L and u ||R are the left and right tangential velocities of the control volume unit grid interface, in m 3 / s;n x and n y are the indicated directions along x and y, respectively; u and v are the single-width flow rates along x and y, respectively, in m / s; variable Q ⊥ =[h,q x n x +q y n y ];q x and q y are the single width flow along x and y, in m 2 / s;S L , S M , S R They are left wave speed, middle wave speed, and right wave speed, in m / s; u ⊥ is the velocity perpendicular to the grid unit interface, in m / s; u ⊥ =un x +vn y ;h * and is an intermediate variable; g is the acceleration due to gravity, in m / s 2 ; T is the average water temperature increase along the water depth, in °C;

[0033] Step 4.2: Diffusion term As the source term, the central difference method is used to solve it. The value on the grid unit interface is replaced by the average value at the center of the two sides of the unit. The thermal diffusion coefficient E in the formula is x 、E y It is expressed by the following formula:

[0034] E x =αu * h,E y =βu * h (8)

[0035] In the formula, α and β are coefficients; h is the water depth in meters; u * is the friction velocity, in m / s;

[0036] Step 4.3: The bottom slope source term is solved using the bottom slope flux method, and the vector form of the bottom slope flux at a certain side f is obtained as:

[0037]

[0038] In the formula, F sf (Q) is the bottom slope flux at a certain side f; n f is the unit external normal vector of the surface f; n fx and n fy Yes f The component of the unit external normal vector of face f, h L and z bL are the water depth and bottom elevation at the center of the grid cell, respectively. With z ML The water depth and bottom elevation at the edge of the grid. g is the acceleration due to gravity in m / s 2 ;

[0039] Step 4.4: The friction source term is solved using the explicit and implicit friction method. The calculation formula is:

[0040]

[0041] Among them, the calculation formula for single-width flow update to time step n+1 is:

[0042]

[0043] In the formula, S fx and S fy S f Friction source term in x and y directions; C f is the bed surface roughness coefficient, C f =gn 2 / h 1 / 3 , n is the Manning coefficient, unit is s / m 1 / 3 ; h is the water depth, in meters; q x ,q y are the average flow velocities in the x and y directions, in m 2 / s;q x n+1 ,q yn+1 They are the single-width flow rates in the x and y directions at time step n+1, in m 2 / s;q x n ,q y n They are the single-width flow rates in the x and y directions at time step n, in m 2 / s; Δt is the time step, in seconds; Δq x is the difference in single-width flow in the x direction at time step n and time step n+1, that is, Δq x =q x n+1 -q x n , unit is m 2 / s.

[0044] Step 5 is as follows:

[0045] Step 5.1: Calculate the net heat flux of water-gas heat exchange According to the total heat balance temperature method, the solar shortwave net radiation flux, the atmospheric longwave radiation flux and the heat conduction flux from the atmosphere to the water body are taken as the main sources of heat on the water surface, and the water body longwave return radiation flux, the water body evaporation heat flux and the heat conduction flux from the water body to the atmosphere are taken as the loss sources of heat on the water surface. The net heat flux of water-gas heat exchange is obtained as follows:

[0046]

[0047] In the formula, is the net surface heat flux in W / m 2 ; is the surface heat flux caused by the solar shortwave net radiation, in W / m 2 ; is the surface heat flux caused by the solar long-wave net radiation, in W / m 2 ; is the surface flux caused by evaporation or condensation loss, in W / m 2 ; is the surface heat flux caused by water vapor heat conduction, in W / m 2 ;

[0048] Step 5.2, solve the heat flux terms in formula (2) one by one: the solar shortwave net radiation surface heat flux, considering the water surface reflectivity and cloud coverage:

[0049]

[0050] In the formula, is the solar shortwave net radiation surface heat flux, in W / m2 ; RSrad is the measured or calculated solar shortwave radiation, in W / m 2 ;γ is the water surface reflectivity, C is the cloud coverage, and β is the water surface absorption rate;

[0051] The solar long-wave net radiation surface heat flux is obtained according to the Stenfan-Boltzman radiation law:

[0052]

[0053] In the formula, ε a is the long-wave emissivity of air, σ is the Stenfan-Boltzman constant, T a is the temperature in °C; C is the cloud cover rate, ε w is the long-wave emissivity of the water body, T s is the surface temperature of the water body, in °C;

[0054] The surface flux of evaporation or condensation is calculated based on the evaporation pressure of the air and water surfaces:

[0055] RE=f(W z )(e s -e a ) (16)

[0056] f(W z )=9.2+0.46W 2 (17)

[0057] e a =exp(2.3026×7.5T d / (T d +237.3)+0.6609) (18)

[0058] e s =0.611×exp(17.3T s / (T s +237.3)) (19)

[0059] Where RE is the evaporation or condensation surface flux, in W / m 2 ; f(W z ) Wind speed function at height zm, unit is Wm -2 mmHg -1 ;e s is the saturated vapor pressure of the water surface, in mmHg -1 ;e a is the atmospheric vapor pressure, in mmHg -1 ; W z is the wind speed 2m above the water surface;

[0060] Heat flux of heat conduction surface:

[0061] RC=Cf(W z )(T s -T a ) (20)

[0062] Where: C is Bowen's coefficient, f(W z ) Wind speed function at height zm, unit is Wm -2 mmHg -1 ; T s is the surface temperature of the water body, in °C; T a is the temperature, in °C;

[0063] Step 5.3: Based on the known data including solar radiation, air temperature, sunshine duration (cloud cover), wind speed and dew point temperature, a generalized water temperature formula is established based on the principle of heat balance, that is, the net surface heat flux of the water body is proportional to the difference between the surface temperature and the equilibrium temperature:

[0064]

[0065] In the formula, is the net surface heat flux in W / m 2 ; K s is the total heat exchange coefficient, in W / (m 2 ·℃), which is a function of water temperature and wind speed; T e is the equilibrium temperature in °C; T sw is the water surface temperature in °C.

[0066] Among them, the total heat exchange coefficient K s for:

[0067] K s =4.50+0.05T sw +βf(W z )+0.47f(W z ) (twenty two)

[0068]

[0069] T m =(T sw +T d ) / 2 (24)

[0070] Wind function f(W z ) is expressed as:

[0071] f(W z )=9.2+0.46W z 2(25)

[0072] Where, β is the coefficient, mmHg / ℃; f(W z ) is the wind function at a height of zm, in Wm -2 mmHg -1 ; W z is the wind speed zm above the water surface; T d is the dew point temperature, in °C; T m is the average temperature in °C;

[0073] In this equation, the empirical formula for equilibrium temperature is:

[0074]

[0075] Where, T e is the equilibrium temperature in °C; T d is the dew point temperature, in °C; is the solar shortwave net radiation heat flux, in W / m 2 ; K s is the total heat exchange coefficient, in W / (m 2 ·℃).

[0076] Step 6 is as follows:

[0077] Step 6.1: Use the hydrostatic reconstruction method to solve the dry-wet boundary, as follows:

[0078] 1) Based on the hydrostatic reconstruction method, the bottom slope elevation at the center point M of the dry-wet unit boundary edge f is:

[0079]

[0080] 2) Reconstruct the water depth on both sides and make the water level on the right side of the dry-wet interface the same as the elevation of the reconstructed bottom slope, that is:

[0081]

[0082] 3) Reconstruct the traffic on both sides of the interface:

[0083]

[0084] In this process, in order to ensure that the reconstructed bottom slope elevation is not higher than the water level on the left, the bottom slope elevation value at the center point M of the dry-wet boundary unit edge f should be the lowest value of the bottom elevation after reconstruction and the water level on the left, that is:

[0085]

[0086] In the formula, are the bottom slope elevations on the left and right sides of the center point M of the dry and wet unit boundary edge f, in m; They are the water levels on the left and right sides of the center point M of the dry and wet unit boundary edge f, in m; are the flow velocities on the left and right sides of the center point M of the dry and wet unit boundary edge f in the x and y directions, respectively, in m / s; z bM is the bottom slope elevation after reconstruction of the center point M of the dry-wet unit boundary edge f, in meters; are the water depths on the left and right sides after reconstruction, in m; are the water levels after reconstruction, in m; They are the single-width flows on the left and right sides in the x and y directions after reconstruction, in m 2 / s;

[0087] Step 6.2, using the MUSCL reconstruction method to calculate the spatial variation gradient of each variable in the unit grid; obtaining the restricted variable gradient by restricting the variable gradient, interpolating the variable value at the midpoint of the unit boundary based on the restricted gradient, and obtaining the left and right variables of the center M of the boundary surface of the kth edge of the control unit grid i, so as to realize the MUSCL second-order reconstruction;

[0088]

[0089] From this, the values ​​of other variables can be derived:

[0090]

[0091] In the formula, are the single-width flow rates on the left and right sides of the center M of the boundary surface of the kth edge of the unit grid in the x and y directions, respectively, in m 2 / s; and are the bottom slope elevation, water level and water depth on the left and right sides of the center M of the boundary surface of the kth edge of the unit grid in the x and y directions, respectively, in meters; and are the flow velocities on the left and right sides of the center M of the boundary surface of the kth edge of the unit grid in the x and y directions, respectively, in m / s;

[0092] Step 6.3: Obtain the second-order time accuracy by the two-step Runge-Kutta method. The single-width flow rate of the i-th control cell grid at the time step n+1 in the new time step is The values ​​are:

[0093]

[0094] Intermediate variable value for

[0095]

[0096] Where: is the single-width flow increment function at time step n; i is the number of control unit grids; Ω is the volume of the i-th control unit grid, in m 3 ;q n+1 is the single-width flow at time step n+1, in m 2 / s,q n is the single-width flow at time step n, in m 2 / s; Δt is the time step, in seconds; S(q n ) is the source term at time step n.

[0097] Step 7 is as follows:

[0098] Under the spatial step length Δx, the calculation formula for the time step length Δt is as follows:

[0099]

[0100] Where Δt is the time step, i.e. the difference between time t+1 and t, s; Δx min It represents the minimum distance from the center of the grid unit to the corresponding interface, in meters; u is the flow velocity, in meters per second; g is the acceleration due to gravity, and h is the water depth, in meters.

[0101] The beneficial effect of the present invention is that the full thermal balance natural water temperature simulation method based on GPU acceleration is a high-efficiency and high-precision natural water temperature simulation method proposed in view of the characteristics of the current natural water temperature model that the simulation accuracy of water temperature transfer in complex rivers, lakes and seas is not high and the simulation speed is slow. It can effectively simulate the natural water temperature transfer process in rivers, lakes and seas, and realize the refined and visualized simulation of the two-dimensional surface water temperature of rivers, lakes and seas, so as to more accurately explore the impact of water temperature on the water ecological environment. BRIEF DESCRIPTION OF THE DRAWINGS

[0102] Figure 1 It is the location and DEM topographic map of a certain Yellow River mainstream near the Lanzhou hydrological station;

[0103] Figure 2 It is a schematic diagram of the heat budget balance process in water bodies;

[0104] Figure 3 It is the daily average flow, meteorological and measured water temperature data of Lanzhou Hydrological Station from May to September 2009;

[0105] Figure 4 It is the simulation result of natural river water temperature based on the full thermal balance water temperature model. DETAILED DESCRIPTION

[0106] The present invention is described in detail below with reference to the accompanying drawings and specific embodiments.

[0107] The full thermal balance natural water temperature simulation method based on GPU acceleration of the present invention is implemented in the following steps:

[0108] Step 1: Use high-resolution DEM terrain to characterize the real terrain of rivers, lakes and seas, and use this as terrain data for model input;

[0109] Step 2: Discretize the terrain data using rectangular grids and assign values ​​to related parameters according to known boundary conditions;

[0110] Step 3: Based on the thermal balance principle and convection-diffusion theory, the full thermal balance method is adopted to fully consider the heat balance process in the water body and construct a two-dimensional natural water temperature convection-diffusion transport equation;

[0111] The two-dimensional natural water temperature convection-diffusion equation in step 3 is expressed as follows:

[0112]

[0113] Where T is the average water temperature increase along the water depth, in degrees Celsius; E x 、E y are the thermal diffusion coefficients in the x and y directions respectively; h is the water depth, in m; u and v are the flow velocities in the x and y directions, in m / s; t is the time, in s; is the net surface heat flux, in W / m 2 ; ρ is water density, unit: kg / m 3 , ρ is 1000kg / m 3 ; C p is the specific heat capacity flux of water, unit: W / (m 2 ·(kg·℃)), C p Take 4178.4W / (m 2 ·(kg·℃)); T i0 It is the source and sink term, and its unit is ℃.

[0114] Step 4: Considering the water flow factor, the hydrodynamic control equation is coupled with the constructed two-dimensional natural water temperature convection diffusion transport equation to obtain a two-dimensional hydrodynamic-water temperature coupling control equation group, which is expressed in vector form; at the same time, the central finite volume method of the Godunov format is further used to discretize the above two-dimensional hydrodynamic-water temperature coupling control equation group, and the flux term, diffusion term and source term (bottom slope and friction source term) of the coupling control equation group are solved;

[0115] The vector form of the two-dimensional hydrodynamic-water temperature coupling control equation in step 4 is as follows:

[0116]

[0117] in,

[0118]

[0119] Where t is time; Q is the variable vector, h is the water depth, unit: m, q x and q y They are the single width flow in the x and y directions, in m 2 / s; F, G are flux vectors in the x and y directions respectively; u, v are flow velocities in the x and y directions respectively, in units of m / s; S is the source term vector, including the bottom slope source term, friction source term, temperature diffusion term and temperature source term; z b is the riverbed bottom elevation, in m; C f is the friction coefficient of the bed surface, calculated as C f =gn 2 / h 1 / 3 , where n is the Manning coefficient; g is the acceleration due to gravity, unit: m / s 2 ; T is the average water temperature increase along the water depth, unit: ℃; E x 、E y are the thermal diffusion coefficients in the x and y directions respectively; h is the water depth in meters; is the net surface heat flux, in W / m 2 ; ρ is water density, unit: kg / m 3 , ρ is 1000kg / m 3 ; C p is the specific heat capacity flux of water, unit: W / (m 2 ·(kg·℃)), C p Take 4178.4W / (m 2 ·(kg·℃)); T i0 It is the source and sink term, and its unit is ℃.

[0120] Step 4 is as follows:

[0121] Step 4.1: For the i-th control unit grid, the central finite volume method is used to integrate the hydrodynamic-water temperature control equations and convert them into line integrals. The specific form is:

[0122]

[0123] Where Ω is the volume of control cell grid i; Γ is the boundary of control volume i; n is the unit vector in the direction of the external normal corresponding to the boundary Γ of the control volume; S is the source term vector, including the bottom slope source term S b , friction source term S f , temperature diffusion term and temperature source term, k is the number of the control unit grid edge; l k is the length of the kth side of the i-th control unit grid;

[0124] The flux term is solved using the HLLC Riemann solver, F(Q)·n is obtained as:

[0125]

[0126] Among them, the control volume unit grid interface intermediate flux F * for,

[0127]

[0128] Control volume unit grid interface flux intermediate wave sides, F * L and F * R for,

[0129]

[0130] In the formula, F L and F R are the fluxes on the left and right sides of the unit grid interface respectively; F * , and are the left wave flux, middle wave flux and right wave flux on both sides of the intermediate wave of the control volume unit grid interface flux respectively; u || is the tangential velocity along the interface, u || =-un y +vn x ,u ||L and u ||R are the left and right tangential velocities of the control volume unit grid interface, in m 3 / s;n x and n y are the indicated directions along x and y, respectively; u and v are the single-width flow rates along x and y, respectively, in m / s; variable Q ⊥ =[h,q x n x +q y n y ];q x and q y are the single width flow along x and y, in m 2 / s;S L , S M , S R They are left wave speed, middle wave speed, and right wave speed, in m / s; u ⊥ is the velocity perpendicular to the grid unit interface, in m / s; u ⊥ =un x +vn y ;h * and is the intermediate variable; g is the acceleration due to gravity, g is 9.81, and the unit is m / s 2 ; T is the average water temperature increase along the water depth, in °C;

[0131] Step 4.2: Diffusion term As the source term, the central difference method is used to solve it. The value on the grid unit interface is replaced by the average value at the center of the two sides of the unit. The thermal diffusion coefficient E in the formula is x 、E y It is expressed by the following formula:

[0132] E x =αu * h,E y =βu * h (8)

[0133] In the formula, α and β are coefficients; h is the water depth in meters; u * is the friction velocity, in m / s;

[0134] Step 4.3: The bottom slope source term is solved using the bottom slope flux method, and the vector form of the bottom slope flux at a certain side f is obtained as:

[0135]

[0136] In the formula, F sf (Q) is the bottom slope flux at a certain side f; n f is the unit external normal vector of the surface f; n fx and n fy Yes f The component of the unit external normal vector of face f, h L and z bL are the water depth and bottom elevation at the center of the grid cell, respectively. With z ML The water depth and bottom elevation at the edge of the grid. g is the acceleration due to gravity in m / s 2 ;

[0137] Step 4.4: The friction source term is solved using the explicit and implicit friction method. The calculation formula is:

[0138]

[0139] Among them, the calculation formula for single-width flow update to time step n+1 is:

[0140]

[0141]

[0142] In the formula, S fx and Sfy S f Friction source term in x and y directions; C f is the bed surface roughness coefficient, C f =gn 2 / h 1 / 3 , n is the Manning coefficient, unit is s / m 1 / 3 ; h is the water depth, in meters; q x ,q y are the average flow velocities in the x and y directions, in m 2 / s;q x n+1 ,q y n+1 They are the single-width flow rates in the x and y directions at time step n+1, in m 2 / s;q x n ,q y n They are the single-width flow rates in the x and y directions at time step n, in m 2 / s; Δt is the time step, in seconds; Δq x is the difference in single-width flow in the x direction at time step n and time step n+1, that is, Δq x =q x n+1 -q x n , unit is m 2 / s; In order to avoid the denominator being zero, θ is taken as 1.0e -12 .

[0143] Step 5: For the water temperature source term of the above coupled control equations, the heat flux terms such as solar shortwave net radiation, longwave net radiation, air-water evaporation or condensation heat flux, heat conduction flux and riverbed heat flux are fully considered and summed up to meet the balance process of water body in rivers, lakes and seas;

[0144] Combination Figure 2 , step 5 is as follows:

[0145] Step 5.1: Calculate the net heat flux of water-gas heat exchange According to the total heat balance temperature method, the solar shortwave net radiation flux, the atmospheric longwave radiation flux and the heat conduction flux from the atmosphere to the water body are taken as the main sources of heat on the water surface, and the water body longwave return radiation flux, the water body evaporation heat flux and the heat conduction flux from the water body to the atmosphere are taken as the loss sources of heat on the water surface. The net heat flux of water-gas heat exchange is obtained as follows:

[0146]

[0147] In the formula, is the net surface heat flux in W / m 2 ; is the surface heat flux caused by the solar shortwave net radiation, in W / m 2 ; is the surface heat flux caused by the solar long-wave net radiation, in W / m 2 ; is the surface flux caused by evaporation or condensation loss, in W / m 2 ; is the surface heat flux caused by water vapor heat conduction, in W / m 2 ;

[0148] Step 5.2, solve the heat flux terms in formula (2) one by one: the solar shortwave net radiation surface heat flux, considering the water surface reflectivity and cloud coverage:

[0149]

[0150] In the formula, is the solar shortwave net radiation surface heat flux, in W / m 2 ; RSrad is the measured or calculated solar shortwave radiation, in W / m 2 ; γ is the water surface reflectivity, which is generally taken as 0.03; C is the cloud coverage rate, which ranges from 0 to 1, with 0 indicating no cloud coverage and 1 indicating complete cloud coverage; β is the water surface absorption rate, which is generally taken as 0.5.

[0151] The solar long-wave net radiation surface heat flux is obtained according to the Stenfan-Boltzman radiation law:

[0152]

[0153] In the formula, ε a is the long-wave emissivity of air, which is taken as 0.97; σ is the Stenfan-Boltzman constant, which is taken as 5.67×10 -8 W / (m 2 K 4 );T a is the temperature in °C; C is the cloud coverage rate, ranging from 0 to 1, where 0 means no cloud coverage and 1 means all clouds are covered; ε w is the long-wave emissivity of the water body, which is taken as 0.97; T s is the surface temperature of the water body, in °C;

[0154] The surface flux of evaporation or condensation is calculated based on the evaporation pressure of the air and water surfaces:

[0155] RE=f(W z )(es -e a ) (16)

[0156] f(W z )=9.2+0.46W 2 (17)

[0157] e a =exp(2.3026×7.5T d / (T d +237.3)+0.6609) (18)

[0158] e s =0.611×exp(17.3T s / (T s +237.3)) (19)

[0159] Where RE is the evaporation or condensation surface flux, in W / m 2 ; f(W z ) Wind speed function at height zm, unit is Wm -2 mmHg -1 ;e s is the saturated vapor pressure of the water surface, in mmHg -1 ;e a is the atmospheric vapor pressure, in mmHg -1 ; W z is the wind speed 2m above the water surface;

[0160] Heat flux of heat conduction surface:

[0161] RC=Cf(W z )(T s -T a ) (20)

[0162] Where: C is Bowen's coefficient, and the value of C is 0.47 mmHg -1 ℃- 1 ; f(W z ) Wind speed function at height zm, unit is Wm -2 mmHg -1 ; T s is the surface temperature of the water body, in °C; T a is the temperature, in °C;

[0163] Step 5.3: Based on the known data including solar radiation, air temperature, sunshine duration (cloud cover), wind speed and dew point temperature, a generalized water temperature formula is established based on the principle of heat balance, that is, the net surface heat flux of the water body is proportional to the difference between the surface temperature and the equilibrium temperature:

[0164]

[0165] In the formula, is the net surface heat flux in W / m 2 ; K s is the total heat exchange coefficient, in W / (m 2 ·℃), which is a function of water temperature and wind speed; T e is the equilibrium temperature in °C; T sw is the water surface temperature in °C.

[0166] Among them, the total heat exchange coefficient K s for:

[0167] K s =4.50+0.05T sw +βf(W z )+0.47f(W z ) (twenty two)

[0168]

[0169] T m =(T sw +T d ) / 2 (24)

[0170] Wind function f(W z ) is expressed as:

[0171] f(W z )=9.2+0.46W z 2 (25)

[0172] Where, β is the coefficient, mmHg / ℃; f(W z ) is the wind function at a height of zm, in Wm -2 mmHg -1 ; W z is the wind speed zm above the water surface; T d is the dew point temperature, in °C; T m is the average temperature in °C;

[0173] In this equation, the empirical formula for equilibrium temperature is:

[0174]

[0175] Where, T e is the equilibrium temperature in °C; T d is the dew point temperature, in °C; is the solar shortwave net radiation heat flux, in W / m 2 ; Ks is the total heat exchange coefficient, in W / (m 2 ·℃).

[0176] Step 6: Use the hydrostatic reconstruction method, MUSCL second-order accuracy and Runge-Kutta method to effectively deal with the dry-wet boundary problem, and further achieve the second-order spatiotemporal accuracy of the model;

[0177] Step 6 is as follows:

[0178] Step 6.1: Use the hydrostatic reconstruction method to solve the dry-wet boundary to maintain non-negative water depth, thereby ensuring mass conservation and calculation stability on the dry-wet boundary interface; the details are as follows:

[0179] 1) Based on the hydrostatic reconstruction method, the bottom slope elevation at the center point M of the dry-wet unit boundary edge f is:

[0180]

[0181] 2) Reconstruct the water depth on both sides and make the water level on the right side of the dry-wet interface the same as the elevation of the reconstructed bottom slope, that is:

[0182]

[0183] 3) Reconstruct the traffic on both sides of the interface:

[0184]

[0185] In this process, in order to ensure that the reconstructed bottom slope elevation is not higher than the water level on the left, the bottom slope elevation value at the center point M of the dry-wet boundary unit edge f should be the lowest value of the bottom elevation after reconstruction and the water level on the left, that is:

[0186]

[0187] In the formula, are the bottom slope elevations on the left and right sides of the center point M of the dry and wet unit boundary edge f, in m; They are the water levels on the left and right sides of the center point M of the dry and wet unit boundary edge f, in m; are the flow velocities on the left and right sides of the center point M of the dry and wet unit boundary edge f in the x and y directions, respectively, in m / s; z bM is the bottom slope elevation after reconstruction of the center point M of the dry-wet unit boundary edge f, in meters; are the water depths on the left and right sides after reconstruction, in m; are the water levels after reconstruction, in m; They are the single-width flows on the left and right sides in the x and y directions after reconstruction, in m 2 / s;

[0188] Step 6.2, using the MUSCL reconstruction method to calculate the spatial variation gradient of each variable in the unit grid; obtaining the restricted variable gradient by restricting the variable gradient, interpolating the variable value at the midpoint of the unit boundary based on the restricted gradient, and obtaining the left and right variables of the center M of the boundary surface of the kth edge of the control unit grid i, so as to realize the MUSCL second-order reconstruction;

[0189]

[0190] From this, the values ​​of other variables can be derived:

[0191]

[0192] In the formula, are the single-width flow rates on the left and right sides of the center M of the boundary surface of the kth edge of the unit grid in the x and y directions, respectively, in m 2 / s; and are the bottom slope elevation, water level and water depth on the left and right sides of the center M of the boundary surface of the kth edge of the unit grid in the x and y directions, respectively, in meters; and are the flow velocities on the left and right sides of the center M of the boundary surface of the kth edge of the unit grid in the x and y directions, respectively, in m / s;

[0193] Step 6.3: Obtain the second-order time accuracy by the two-step Runge-Kutta method. The single-width flow rate of the i-th control cell grid at the time step n+1 in the new time step is The values ​​are:

[0194]

[0195] Intermediate variable value for

[0196]

[0197] Where: is the single-width flow increment function at time step n; i is the number of control unit grids; Ω is the volume of the i-th control unit grid, in m 3 ;q n+1 is the single-width flow at time step n+1, in m 2 / s,q n is the single-width flow at time step n, in m 2 / s; Δt is the time step, in seconds; S(q n ) is the source term at time step n.

[0198] Step 7: Use CFL conditions to limit the time stability, and introduce GPU acceleration technology into the entire model program to further improve the simulation efficiency of the model;

[0199] Step 7 is as follows:

[0200] Under the spatial step length Δx, the calculation formula for the time step length Δt is as follows:

[0201]

[0202] Where Δt is the time step, i.e. the difference between time t+1 and t, s; Δx min It represents the minimum distance from the center of the grid unit to the corresponding interface, in meters; u is the flow velocity, in meters per second; g is the gravitational acceleration, with a value of 9.81 meters per second 2 ; h is the water depth, in meters;

[0203] During the entire solution process, at the spatial step length Δx and time t, flux terms, diffusion terms, bottom slope source terms, friction source terms and net surface heat flux terms are involved. GPU parallel computing technology is introduced to improve computational efficiency.

[0204] Step 8. Based on MATLAB software, use the MATLAB plotting program to output the natural water temperature transport process in the study area, so as to characterize the two-dimensional dynamic distribution map of the natural water temperature in the study area and further reveal the temporal and spatial variation law of the natural water temperature.

[0205] Example 1

[0206] The full thermal balance natural water temperature simulation method based on GPU acceleration of the present invention is implemented in the following steps:

[0207] Step 1: Use high-resolution DEM terrain to characterize the real terrain of rivers, lakes and seas, and use this as terrain data for model input;

[0208] Step 2: Discretize the terrain data using rectangular grids and assign values ​​to related parameters according to known boundary conditions;

[0209] Step 3: Based on the thermal balance principle and convection-diffusion theory, the full thermal balance method is adopted to fully consider the heat balance process in the water body and construct a two-dimensional natural water temperature convection-diffusion transport equation;

[0210] Step 4: Considering the water flow factor, the hydrodynamic control equation is coupled with the constructed two-dimensional natural water temperature convection diffusion transport equation to obtain a two-dimensional hydrodynamic-water temperature coupling control equation group, which is expressed in vector form; at the same time, the central finite volume method of the Godunov format is further used to discretize the above two-dimensional hydrodynamic-water temperature coupling control equation group, and the flux term, diffusion term and source term (bottom slope and friction source term) of the coupling control equation group are solved;

[0211] Step 5: For the water temperature source term of the above coupled control equations, the heat flux terms such as solar shortwave net radiation, longwave net radiation, air-water evaporation or condensation heat flux, heat conduction flux and riverbed heat flux are fully considered and summed up to meet the balance process of water body in rivers, lakes and seas;

[0212] Step 6: Use the hydrostatic reconstruction method, MUSCL second-order accuracy and Runge-Kutta method to effectively deal with the dry-wet boundary problem, and further achieve the second-order spatiotemporal accuracy of the model;

[0213] Step 7: Use CFL conditions to limit the time stability, and introduce GPU acceleration technology into the entire model program to further improve the simulation efficiency of the model;

[0214] Step 8. Based on MATLAB software, use the MATLAB plotting program to output the natural water temperature transport process in the study area, so as to characterize the two-dimensional dynamic distribution map of the natural water temperature in the study area and further reveal the temporal and spatial variation law of the natural water temperature.

[0215] Example 2

[0216] The full thermal balance natural water temperature simulation method based on GPU acceleration of the present invention is implemented in the following steps:

[0217] Step 1: Use high-resolution DEM terrain to characterize the real terrain of rivers, lakes and seas, and use this as terrain data for model input;

[0218] Step 2: Discretize the terrain data using rectangular grids and assign values ​​to related parameters according to known boundary conditions;

[0219] Step 3: Based on the thermal balance principle and convection-diffusion theory, the full thermal balance method is adopted to fully consider the heat balance process in the water body and construct a two-dimensional natural water temperature convection-diffusion transport equation;

[0220] The two-dimensional natural water temperature convection-diffusion equation in step 3 is expressed as follows:

[0221]

[0222] Where T is the average water temperature increase along the water depth, in degrees Celsius; E x 、E y are the thermal diffusion coefficients in the x and y directions respectively; h is the water depth, in m; u and v are the flow velocities in the x and y directions, in m / s; t is the time, in s; is the net surface heat flux, in W / m 2 ; ρ is water density, unit: kg / m 3 , ρ is 1000kg / m3 ; C p is the specific heat capacity flux of water, unit: W / (m 2 ·(kg·℃)), C p Take 4178.4W / (m 2 ·(kg·℃)); T i0 It is the source and sink term, and its unit is ℃.

[0223] Step 4: Considering the water flow factor, the hydrodynamic control equation is coupled with the constructed two-dimensional natural water temperature convection diffusion transport equation to obtain a two-dimensional hydrodynamic-water temperature coupling control equation group, which is expressed in vector form; at the same time, the central finite volume method of the Godunov format is further used to discretize the above two-dimensional hydrodynamic-water temperature coupling control equation group, and the flux term, diffusion term and source term (bottom slope and friction source term) of the coupling control equation group are solved;

[0224] The vector form of the two-dimensional hydrodynamic-water temperature coupling control equation in step 4 is as follows:

[0225]

[0226] in,

[0227]

[0228] Where t is time; Q is the variable vector, h is the water depth, unit: m, q x and q y They are the single width flow in the x and y directions, in m 2 / s; F, G are flux vectors in the x and y directions respectively; u, v are flow velocities in the x and y directions respectively, in units of m / s; S is the source term vector, including the bottom slope source term, friction source term, temperature diffusion term and temperature source term; z b is the riverbed bottom elevation, in m; C f is the friction coefficient of the bed surface, calculated as C f =gn 2 / h 1 / 3 , where n is the Manning coefficient; g is the acceleration due to gravity, unit: m / s 2 ; T is the average water temperature increase along the water depth, unit: ℃; E x 、E y are the thermal diffusion coefficients in the x and y directions respectively; h is the water depth in meters; is the net surface heat flux, in W / m 2 ; ρ is water density, unit: kg / m 3 , ρ is 1000kg / m 3 ; C p is the specific heat capacity flux of water, unit: W / (m2 ·(kg·℃)), C p Take 4178.4W / (m 2 ·(kg·℃)); T i0 It is the source and sink term, and its unit is ℃.

[0229] Step 5: For the water temperature source term of the above coupled control equations, the heat flux terms such as solar shortwave net radiation, longwave net radiation, air-water evaporation or condensation heat flux, heat conduction flux and riverbed heat flux are fully considered and summed up to meet the balance process of water body in rivers, lakes and seas;

[0230] Step 6: Use the hydrostatic reconstruction method, MUSCL second-order accuracy and Runge-Kutta method to effectively deal with the dry-wet boundary problem, and further achieve the second-order spatiotemporal accuracy of the model;

[0231] Step 6 is as follows:

[0232] Step 6.1: Use the hydrostatic reconstruction method to solve the dry-wet boundary to maintain non-negative water depth, thereby ensuring mass conservation and calculation stability on the dry-wet boundary interface; the details are as follows:

[0233] 1) Based on the hydrostatic reconstruction method, the bottom slope elevation at the center point M of the dry-wet unit boundary edge f is:

[0234]

[0235] 2) Reconstruct the water depth on both sides and make the water level on the right side of the dry-wet interface the same as the elevation of the reconstructed bottom slope, that is:

[0236]

[0237] 3) Reconstruct the traffic on both sides of the interface:

[0238]

[0239] In this process, in order to ensure that the reconstructed bottom slope elevation is not higher than the water level on the left, the bottom slope elevation value at the center point M of the dry-wet boundary unit edge f should be the lowest value of the bottom elevation after reconstruction and the water level on the left, that is:

[0240]

[0241] In the formula, are the bottom slope elevations on the left and right sides of the center point M of the dry and wet unit boundary edge f, in m; They are the water levels on the left and right sides of the center point M of the dry and wet unit boundary edge f, in m; are the flow velocities on the left and right sides of the center point M of the dry and wet unit boundary edge f in the x and y directions, respectively, in m / s; z bMis the bottom slope elevation after reconstruction of the center point M of the dry-wet unit boundary edge f, in meters; are the water depths on the left and right sides after reconstruction, in m; are the water levels after reconstruction, in m; They are the single-width flows on the left and right sides in the x and y directions after reconstruction, in m 2 / s;

[0242] Step 6.2, using the MUSCL reconstruction method to calculate the spatial variation gradient of each variable in the unit grid; obtaining the restricted variable gradient by restricting the variable gradient, interpolating the variable value at the midpoint of the unit boundary based on the restricted gradient, and obtaining the left and right variables of the center M of the boundary surface of the kth edge of the control unit grid i, so as to realize the MUSCL second-order reconstruction;

[0243]

[0244] From this, the values ​​of other variables can be derived:

[0245]

[0246] In the formula, are the single-width flow rates on the left and right sides of the center M of the boundary surface of the kth edge of the unit grid in the x and y directions, respectively, in m 2 / s; and are the bottom slope elevation, water level and water depth on the left and right sides of the center M of the boundary surface of the kth edge of the unit grid in the x and y directions, respectively, in meters; and are the flow velocities on the left and right sides of the center M of the boundary surface of the kth edge of the unit grid in the x and y directions, respectively, in m / s;

[0247] Step 6.3: Obtain the second-order time accuracy by the two-step Runge-Kutta method. The single-width flow rate of the i-th control cell grid at the time step n+1 in the new time step is The values ​​are:

[0248]

[0249] Intermediate variable value for

[0250]

[0251] Where: is the single-width flow increment function at time step n; i is the number of control unit grids; Ω is the volume of the i-th control unit grid, in m 3 ;q n+1 is the single-width flow at time step n+1, in m 2 / s,q n is the single-width flow at time step n, in m 2 / s; Δt is the time step, in seconds; S(q n ) is the source term at time step n.

[0252] Step 7: Use CFL conditions to limit the time stability, and introduce GPU acceleration technology into the entire model program to further improve the simulation efficiency of the model;

[0253] Step 7 is as follows:

[0254] Under the spatial step length Δx, the calculation formula for the time step length Δt is as follows:

[0255]

[0256] Where Δt is the time step, i.e. the difference between time t+1 and t, s; Δx min It represents the minimum distance from the center of the grid unit to the corresponding interface, in meters; u is the flow velocity, in meters per second; g is the gravitational acceleration, with a value of 9.81 meters per second 2 ; h is the water depth, in meters;

[0257] During the entire solution process, at the spatial step length Δx and time t, flux terms, diffusion terms, bottom slope source terms, friction source terms and net surface heat flux terms are involved. GPU parallel computing technology is introduced to improve computational efficiency.

[0258] Step 8. Based on MATLAB software, use the MATLAB plotting program to output the natural water temperature transport process in the study area, so as to characterize the two-dimensional dynamic distribution map of the natural water temperature in the study area and further reveal the temporal and spatial variation law of the natural water temperature.

[0259] Example 3

[0260] A section of the mainstream of the Yellow River near Lanzhou Station was selected as the research object. Combined with the hydrological and meteorological data of Lanzhou Station in relevant literature, the water temperature transport simulation of the natural river channel of this section was carried out to verify and apply the constructed method. Figure 1 Due to the lack of relevant measured data, only the daily average hydrological data (flow) and meteorological data (solar radiation, air temperature, measured water temperature, cloud cover, wind speed and dew point temperature) of Lanzhou Hydrological Station from May to September 2009 were collected. Figure 3 .

[0261] Combining the generalized water temperature formula with the measured values ​​of natural water temperature, and comparing the simulation results of natural water temperature, the simulation values ​​of natural water temperature of the model are consistent with the change trend of the measured values ​​and the calculated values. Figure 4, through statistical analysis, compared with the measured value of natural water temperature, the maximum difference is only 1.1℃, the error is between 0.291 and 5.978%, and the maximum absolute error is only 5.978%; at the same time, the error variables MAE, RMSE and NSE are 0.298, 1.390 and 0.940 respectively. Compared with the calculated value of generalized water temperature, the maximum difference is only +0.38℃, the error is between -1.314 and 2.203%, and the maximum absolute error is only 2.203%; the error variables MAE, RMSE and NSE are 0.143, 0.557, 0.992 respectively. It can be seen that when the measured data is complete, the natural water temperature simulation results of the full thermal balance water temperature model are consistent with the natural water temperature of the actual river.

Claims

1. A full thermal balance natural water temperature simulation method based on GPU acceleration, characterized in that: Follow the steps below to implement it: Step 1: Use high-resolution DEM terrain to characterize the real terrain of rivers, lakes and seas, and use this as terrain data for model input; Step 2: Assign values ​​to relevant parameters; Step 3, construct a two-dimensional natural water temperature convection-diffusion transport equation; Step 4, coupling the hydrodynamic control equation with the constructed two-dimensional natural water temperature convection-diffusion transport equation to obtain a two-dimensional hydrodynamic-water temperature coupled control equation group, discretizing the above two-dimensional hydrodynamic-water temperature coupled control equation group, and solving the flux term, diffusion term and source term of the coupled control equation group; Step 5: Meeting the internal balance of revenue and expenditure in rivers, lakes and seas; Step 6: Effectively handle the dry-wet boundary problem and achieve the second-order spatiotemporal accuracy of the model; Step 7: Use CFL conditions to limit the stability over time; Step 8: Output the natural water temperature transport process in the study area.

2. The method for simulating natural water temperature with full thermal balance based on GPU acceleration according to claim 1, characterized in that: The two-dimensional natural water temperature convection diffusion equation in step 3 is expressed as follows: Where T is the average water temperature increase along the water depth, in degrees Celsius; E x 、E y are the thermal diffusion coefficients in the x and y directions respectively; h is the water depth, in m; u and v are the flow velocities in the x and y directions, in m / s; t is the time, in s; is the net surface heat flux, in W / m 2 ; ρ is water density, unit: kg / m 3 , C p is the specific heat capacity flux of water, unit: W / (m 2 ·(kg·℃)), T i0 It is the source and sink term, and its unit is ℃.

3. The method for simulating natural water temperature with full thermal balance based on GPU acceleration according to claim 2, characterized in that: The vector form of the two-dimensional hydrodynamic-water temperature coupling control equation in step 4 is expressed as follows: in, Where t is time; Q is the variable vector, h is the water depth, unit: m, q x and q y They are the single width flow in the x and y directions, in m 2 / s; F, G are flux vectors in the x and y directions respectively; u, v are flow velocities in the x and y directions respectively, in units of m / s; S is the source term vector, including the bottom slope source term, friction source term, temperature diffusion term and temperature source term; z b is the riverbed bottom elevation, in m; C f is the friction coefficient of the bed surface, calculated as C f =gn 2 / h 1 / 3 , where n is the Manning coefficient; g is the acceleration due to gravity, unit: m / s 2 ; T is the average water temperature increase along the water depth, unit: ℃; E x 、E y are the thermal diffusion coefficients in the x and y directions respectively; h is the water depth in meters; is the net surface heat flux, in W / m 2 ; ρ is water density, unit: kg / m 3 , C p is the specific heat capacity flux of water, unit: W / (m 2 ·(kg·℃)), T i0 It is the source and sink term, and its unit is ℃.

4. The method for simulating natural water temperature with full thermal balance based on GPU acceleration according to claim 3, characterized in that: The step 4 is specifically as follows: Step 4.1: For the i-th control unit grid, the central finite volume method is used to integrate the hydrodynamic-water temperature control equations and convert them into line integrals. The specific form is: Where Ω is the volume of control cell grid i; Γ is the boundary of control volume i; n is the unit vector in the direction of the external normal corresponding to the boundary Γ of the control volume; S is the source term vector, including the bottom slope source term S b , friction source term S f , temperature diffusion term and temperature source term, k is the number of the control unit grid edge; l k is the length of the kth side of the i-th control unit grid; The flux term is solved using the HLLC Riemann solver, F(Q)·n is obtained as: Among them, the control volume unit grid interface intermediate flux F * for, Control volume unit grid interface flux intermediate wave sides, and for, In the formula, F L and F R are the fluxes on the left and right sides of the unit grid interface respectively; F * , and are the left wave flux, middle wave flux and right wave flux on both sides of the intermediate wave of the control volume unit grid interface flux respectively; u || is the tangential velocity along the interface, u || =-un y +vn x ,u ||L and u ||R are the left and right tangential velocities of the control volume unit grid interface, in m 3 / s;n x and n y are the indicated directions along x and y, respectively; u and v are the single-width flow rates along x and y, respectively, in m / s; variable Q ⊥ =[h,q x n x +q y n y ];q x and q y are the single width flow along x and y, in m 2 / s;S L , S M , S R They are left wave speed, middle wave speed, and right wave speed, in m / s; u ⊥ is the velocity perpendicular to the grid unit interface, in m / s; u ⊥ =un x +vn y ;h * and u * ⊥ is an intermediate variable; g is the acceleration due to gravity, in m / s 2 ; T is the average water temperature increase along the water depth, in °C; Step 4.2: Diffusion term As the source term, the central difference method is used to solve it. The value on the grid unit interface is replaced by the average value at the center of the two sides of the unit. The thermal diffusion coefficient E in the formula is x 、E y It is expressed by the following formula: E x =αu * h,E y =βu * h (8) In the formula, α and β are coefficients; h is the water depth in meters; u * is the friction velocity, in m / s; Step 4.3: The bottom slope source term is solved using the bottom slope flux method, and the vector form of the bottom slope flux at a certain side f is obtained as: In the formula, F sf (Q) is the bottom slope flux at a certain side f; n f is the unit external normal vector of the surface f; n fx and n fy Yes f The component of the unit normal vector to the face f, h L and z bL are the water depth and bottom elevation at the center of the grid cell, respectively. With z ML The water depth and bottom elevation at the edge of the grid. g is the acceleration due to gravity in m / s 2 ; Step 4.4: The friction source term is solved using the explicit and implicit friction method. The calculation formula is: Among them, the calculation formula for single-width flow update to time step n+1 is: In the formula, S fx and S fy S f Friction source term in x and y directions; C f is the bed surface roughness coefficient, C f =gn 2 / h 1 / 3 , n is the Manning coefficient, unit is s / m 1 / 3 ; h is the water depth, in meters; q x ,q y are the average flow velocities in the x and y directions, in m 2 / s;q x n +1 ,q y n+1 They are the single-width flow rates in the x and y directions at time step n+1, in m 2 / s;q x n ,q y n They are the single-width flow rates in the x and y directions at time step n, in m 2 / s; Δt is the time step, in seconds; Δq x is the difference in single-width flow in the x direction at time step n and time step n+1, that is, Δq x =q x n+1 -q x n , unit is m 2 / s.

5. The method for simulating natural water temperature with full thermal balance based on GPU acceleration according to claim 4, characterized in that: The step 5 is specifically as follows: Step 5.1: Calculate the net heat flux of water-gas heat exchange According to the total heat balance temperature method, the solar shortwave net radiation flux, the atmospheric longwave radiation flux and the heat conduction flux from the atmosphere to the water body are taken as the main sources of heat on the water surface, and the water body longwave return radiation flux, the water body evaporation heat flux and the heat conduction flux from the water body to the atmosphere are taken as the loss sources of heat on the water surface. The net heat flux of water-gas heat exchange is obtained as follows: In the formula, is the net surface heat flux in W / m 2 ; is the surface heat flux caused by the solar shortwave net radiation, in W / m 2 ; is the surface heat flux caused by the solar long-wave net radiation, in W / m 2 ; is the surface flux caused by evaporation or condensation loss, in W / m 2 ; is the surface heat flux caused by water vapor heat conduction, in W / m 2 ; Step 5.2, solve the heat flux terms in formula (2) one by one: the solar shortwave net radiation surface heat flux, considering the water surface reflectivity and cloud coverage: In the formula, is the solar shortwave net radiation surface heat flux, in W / m 2 ; RSrad is the measured or calculated solar shortwave radiation, in W / m 2 ;γ is the water surface reflectivity, C is the cloud coverage, and β is the water surface absorption rate; The solar long-wave net radiation surface heat flux is obtained according to the Stenfan-Boltzman radiation law: In the formula, ε a is the long-wave emissivity of air, σ is the Stenfan-Boltzman constant, T a is the temperature in °C; C is the cloud cover rate, ε w is the long-wave emissivity of the water body, T s is the surface temperature of the water body, in °C; The surface flux of evaporation or condensation is calculated based on the evaporation pressure of the air and water surfaces: RE=f(W z )(have been s -have been a ) (16) f(W z )=9.2+0.46W 2 (17) e a =exp(2.3026×7.5T d / (T d +237.3)+0.6609) (18) yes s =0.611×exp(17.3T s / (T s +237.3)) (19) Where RE is the evaporation or condensation surface flux, in W / m 2 ; f(W z ) Wind speed function at height zm, unit is Wm - 2 mmHg -1 ;e s is the saturated vapor pressure of the water surface, in mmHg -1 ;e a is the atmospheric vapor pressure, in mmHg -1 ; W z is the wind speed 2m above the water surface; Heat flux of heat conduction surface: RC=Cf(W z )(T s -T a ) (20) Where: C is Bowen's coefficient, f(W z ) Wind speed function at height zm, unit is Wm -2 mmHg -1 ; T s is the surface temperature of the water body, in °C; T a is the temperature, in °C; Step 5.3: Based on the known data including solar radiation, air temperature, sunshine duration (cloud cover), wind speed and dew point temperature, a generalized water temperature formula is established based on the principle of heat balance, that is, the net surface heat flux of the water body is proportional to the difference between the surface temperature and the equilibrium temperature: In the formula, is the net surface heat flux in W / m 2 ; K s is the total heat exchange coefficient, in W / (m 2 ·℃), which is a function of water temperature and wind speed; T e is the equilibrium temperature in °C; T sw is the surface temperature of the water body, in °C; Among them, the total heat exchange coefficient K s for: K s =4.50+0.05T sw +βf(W z )+0.47f(W z ) (22) T m =(T sw +T d ) / 2 (24) Wind function f(W z ) is expressed as: f(W z )=9.2+0.46W z 2 (25) Where, β is the coefficient, mmHg / ℃; f(W z ) is the wind function at a height of zm, in Wm -2 mmHg -1 ; W z is the wind speed zm above the water surface; T d is the dew point temperature, in °C; T m is the average temperature in °C; In this equation, the empirical formula for equilibrium temperature is: Where, T e is the equilibrium temperature in °C; T d is the dew point temperature, in °C; is the solar shortwave net radiation heat flux, in W / m 2 ; K s is the total heat exchange coefficient, in W / (m 2 .℃).

6. The method for simulating natural water temperature with full thermal balance based on GPU acceleration according to claim 5, characterized in that: The step 6 is specifically as follows: Step 6.1: Use the hydrostatic reconstruction method to solve the dry-wet boundary, as follows: 1) Based on the hydrostatic reconstruction method, the bottom slope elevation at the center point M of the dry-wet unit boundary edge f is: 2) Reconstruct the water depth on both sides and make the water level on the right side of the dry-wet interface the same as the elevation of the reconstructed bottom slope, that is: 3) Reconstruct the traffic on both sides of the interface: In this process, in order to ensure that the reconstructed bottom slope elevation is not higher than the water level on the left, the bottom slope elevation value at the center point M of the dry-wet boundary unit edge f should be the lowest value of the bottom elevation after reconstruction and the water level on the left, that is: In the formula, are the bottom slope elevations on the left and right sides of the center point M of the dry and wet unit boundary edge f, in m; They are the water levels on the left and right sides of the center point M of the dry and wet unit boundary edge f, in m; are the flow velocities on the left and right sides of the center point M of the dry and wet unit boundary edge f in the x and y directions, respectively, in m / s; z bM is the bottom slope elevation after reconstruction of the center point M of the dry-wet unit boundary edge f, in meters; are the water depths on the left and right sides after reconstruction, in m; are the water levels after reconstruction, in m; They are the single-width flows on the left and right sides in the x and y directions after reconstruction, in m 2 / s; Step 6.2, using the MUSCL reconstruction method to calculate the spatial variation gradient of each variable in the unit grid; obtaining the restricted variable gradient by restricting the variable gradient, interpolating the variable value at the midpoint of the unit boundary based on the restricted gradient, and obtaining the left and right variables of the center M of the boundary surface of the kth edge of the control unit grid i, so as to realize the MUSCL second-order reconstruction; From this, the values ​​of other variables can be derived: In the formula, are the single-width flow rates on the left and right sides of the center M of the boundary surface of the kth edge of the unit grid in the x and y directions, respectively, in m 2 / s; and are the bottom slope elevation, water level and water depth on the left and right sides of the center M of the boundary surface of the kth edge of the unit grid in the x and y directions, respectively, in meters; and are the flow velocities on the left and right sides of the center M of the boundary surface of the kth edge of the unit grid in the x and y directions, respectively, in m / s; Step 6.3: Obtain the second-order time accuracy by the two-step Runge-Kutta method. The single-width flow rate of the i-th control cell grid at the time step n+1 in the new time step is The values ​​are: Intermediate variable value for Where: is the single-width flow increment function at time step n; i is the number of control unit grids; Ω is the volume of the i-th control unit grid, in m 3 ;q n+1 is the single-width flow at time step n+1, in m 2 / s,q n is the single-width flow at time step n, in m 2 / s; Δt is the time step, in seconds; S(q n ) is the source term at time step n.

7. The method for simulating natural water temperature with full thermal balance based on GPU acceleration according to claim 6, characterized in that: The step 7 is specifically as follows: Under the spatial step length Δx, the calculation formula for the time step length Δt is as follows: Where Δt is the time step, i.e. the difference between time t+1 and t, s; Δx min It represents the minimum distance from the center of the grid unit to the corresponding interface, in meters; u is the flow velocity, in meters per second; g is the acceleration due to gravity, and h is the water depth, in meters.

Citation Information

Cited By

  • Wetland large-scale plant and biogenic element circulation coupling simulation method and device

    CN122334114A

  • Wetland macrophyte and biogenic element circulation coupling simulation method and device

    CN122334114B