Rapid construction method of computational fluid mechanics solver

Abstract

The invention provides a quick construction method for a computational fluid mechanics solver, which relates to the field of numerical simulation and comprises the following steps of: determining a control equation required by calculation; presetting an operator and a function interface thereof; each item in the control equation is converted to be represented in the form of an operator; and calling a corresponding function interface to solve the control equation according to an operator in the control equation, and calculating to obtain an original variable. According to the method, the problems that in the prior art, the flexibility is limited, the fluid mechanics numerical simulation problem cannot be deeply customized and solved, and the customization process is complex can be solved; the flexibility of solving the fluid mechanics numerical simulation problem is improved, and the customizing process is simple.

Description

Technical Field

[0001] The present invention relates to the field of numerical simulation, and particularly to a method for rapidly constructing a computational fluid dynamics solver. Background Art

[0002] When performing numerical simulation of fluid mechanics, it is mainly carried out through commercial software and open-source software. However, with the replacement of equipment and the diversification and complexity of simulation scenarios, commercial software and open-source software often cannot adapt to the rapid development and efficient implementation of specific scenarios.

[0003] When using commercial software or open-source software to solve numerical simulation problems of fluid mechanics, the flexibility of commercial software is limited, and users can only select models and solution methods at the upper layer and cannot deeply customize; while open-source software is too low-level, difficult to reuse, and the customization process is complex. Summary of the Invention

[0004] In view of the above problems, a method for rapidly constructing a computational fluid dynamics solver provided by the present invention can solve the problems of limited flexibility in the prior art, inability to deeply customize the numerical simulation problems of fluid mechanics, and complex customization process; and realizes improving the flexibility of solving numerical simulation problems of fluid mechanics and simplifying the customization process.

[0005] To achieve the above object, the technical solution adopted by the present invention is as follows:

[0006] A method for rapidly constructing a computational fluid dynamics solver provided by the present invention includes the following steps:

[0007] S101: Determine the governing equations required for the calculation;

[0008] S102: Preset operators and their function interfaces; convert each term in the governing equations into a form represented by operators;

[0009] S103: Generate a solution program by calling the corresponding function interfaces according to the operators in the governing equations, solve the governing equations according to the solution program, and calculate and obtain the primitive variables.

[0010] In the method for rapidly constructing a computational fluid dynamics solver provided by the present invention, preferably, the operators in step S102 include:

[0011] Ordinary divergence of a single variable Higher-order interpolation divergence of two variables Limiter interpolation divergence of two variables Scalar divergence Ordinary Laplacian of a single variable Δφ; ordinary Laplacian of two variables Vector component source term q and scalar source term f i ; left-end transient term Right - hand transient term Gradient Calculate the time step Δt; calculate the molecular viscous stress τ vis ; calculate the sub - grid stress τ sgs ;

[0012] Among them, the function interface of the univariate ordinary divergence is div(Field <vector>phi);

[0013] The function interface for the divergence of bivariate high-order interpolation is div(Field <vector>rhoU, Field<scalar, T> phi, Field <vector>grad);

[0014] The function interface for the divergence of the bivariate limiter interpolation is div(Field <vector>rhoU,Field <scalar>phi,Field <vector>grad,Field <scalar>limter);

[0015] The function interface of the scalar divergence is div(Field <scalar>phi,vector <t>orient);

[0016] The function interface of the univariate ordinary Laplace is laplas(Field <scalar>phi);

[0017] The function interface of the bivariate ordinary Laplace is laplas(Field <scalar>gamma,Field <scalar>phi);

[0018] The function interface for the vector component source term and the scalar source term is src(Field<scalar, T> q);

[0019] The function interface for the left - hand - side transient term is lddt(Field<scalar, T> phi);

[0020] The function interface for the right - hand - side transient term is rddt(Field<scalar, T> phi);

[0021] The function interface for the gradient is grad(Field <scalar>phi);

[0022] The function interface for calculating the time step is calcTimeStepFlowCFL();

[0023] The function interface for calculating the molecular viscous stress is calcVisStress(Field <tensor>&gradU,Field <scalar>nu);

[0024] The function interface for calculating sub-grid stress is calcSGStress(Field <tensor>&gradUTilda,Field <scalar>nuTilda).

[0025] The method for quickly constructing a computational fluid dynamics solver provided by the present invention. Preferably, in step S103, when solving the control equation, the control equation needs to be iteratively updated, and the method for updating the control equation includes the following steps:

[0026] S201: Calculate the time step;

[0027] S202: Calculate the original quantity or conserved quantity required for updating the control equation;

[0028] S203: Calculate the gradient required for updating the control equation;

[0029] S204: Initialize the array for the current time step; allocate space and set the values of the array for the current time step to zero;

[0030] S205: Update the spatial terms, including the source term and the flux term, and stack the content into the array step by step;

[0031] S206: Update the time term, process the sum of the spatial terms, and stack the value of the previous step array;

[0032] S207: Update the numerical value of the previous step result;

[0033] S208: Calculate the output quantity.

[0034] The method for quickly constructing a computational fluid dynamics solver provided by the present invention. Preferably, the types of control equations applicable to the operator include:

[0035] Continuity equation:

[0036] Pressure elliptic equation:

[0037] Pressure parabolic equation:

[0038] Compressible momentum equation, u - component, pressure gradient component form:

[0039] Compressible momentum equation, u - component, pressure divergence form:

[0040] Incompressible momentum equation, u - component:

[0041] Incompressible energy equation:

[0042] Scalar equation:

[0043] Total energy equation:

[0044] Total enthalpy equation:

[0045] RANS turbulence model k equation:

[0046] RANS turbulence model ∈ equation:

[0047] Chemical reaction component equation:

[0048] VOF volume fraction equation:

[0049] Where ρ is density; t is time; ν is velocity; u is the x - direction component of velocity; is the Laplace operator; p is pressure; f is the force source term; T is temperature; h is total enthalpy; φ is a scalar; k is the turbulent kinetic energy; Y k is the component mass fraction; α is the multiphase flow volume fraction.

[0050] The above - mentioned technical solution has the following advantages or beneficial effects:

[0051] The method for quickly constructing a computational fluid dynamics solver provided by the present invention. In step S101, a physical model and a mathematical model are formed through mathematical derivation, and the control equations (equations) to be used are determined to construct the solver through different control equations; in step S102, preset operators (operators) and their function interfaces are set, and each term (term) in the control equation is transformed into a form represented by an operator, and different operators can be flexibly selected to transform the control equation; in step S103, a solution process is designed according to the control equation, and a solution procedure can be directly generated according to the control equation and the function interfaces of its operators, thus constituting the main framework of the solver. Further, the control equation is solved through the solution procedure to calculate and obtain the original variables; by freely choosing the control equation, independently designing the solution process, and flexibly selecting operators, the solver can be adapted to the fluid dynamics numerical simulation problem to be solved, improving the adaptability of the solver. Further, independently constructing the solution process can optimize the calculation method of the solver by itself, effectively improving the calculation efficiency of the solver to reduce the time cost required for fluid dynamics numerical simulation; by transforming the control equation into the form of an operator, the control equation can be simplified, and the function interface corresponding to the operator can be directly called to generate the solution procedure, reducing the amount of code used to construct the solver, so as to quickly and simply construct a solver that meets the actual requirements of the application scenario;

[0052] When using the existing technology, there are problems such as limited technical flexibility, inability to deeply customize the solution of computational fluid dynamics numerical simulation problems, and complex customization processes. By using the method for quickly constructing a computational fluid dynamics solver provided by the present invention, the flexibility of solver construction is improved by freely selecting control equations, independently designing the solution process, and flexibly selecting operators. The control equation is simplified by transforming it into the form of an operator, and the function interface corresponding to the operator is directly called to generate a solution program, so as to reduce the complexity of constructing the solver. Description of the Drawings

[0053] By reading the following detailed description of the non-limiting embodiments with reference to the accompanying drawings, the present invention and its features, shape, and advantages will become more obvious. The same reference numerals indicate the same parts in all the drawings. The drawings are not deliberately drawn to scale, and the focus is on showing the gist of the present invention.

[0054] Figure 1 It is a flowchart block diagram of a method for quickly constructing a computational fluid dynamics solver provided in Embodiment 1 of the present invention.

[0055] Figure 2 It is a schematic diagram of the main framework of the solver constructed in the method for quickly constructing a computational fluid dynamics solver provided in Embodiment 1 of the present invention.

[0056] Figure 3 It is a flowchart block diagram of a method for updating the control equation in the method for quickly constructing a computational fluid dynamics solver provided in Embodiment 1 of the present invention. Detailed Embodiments

[0057] The following further illustrates the present invention with reference to the accompanying drawings and specific embodiments, but it is not a limitation of the present invention.

[0058] Embodiment 1:

[0059] As Figure 1 shown, a method for quickly constructing a computational fluid dynamics solver provided in Embodiment 1 of the present invention includes the following steps:

[0060] S101: Determine the control equation required for the calculation;

[0061] S102: Preset the operator and its function interface; transform each term in the control equation into the form represented by the operator;

[0062] S103: Call the corresponding function interface according to the operator in the control equation to generate a solution program, solve the control equation according to the solution program, and calculate to obtain the original variables.

[0063] The method for quickly constructing a computational fluid dynamics solver provided in Embodiment 1 of the present invention. In step S101, a physical model and a mathematical model are formed through mathematical derivation, and the control equation (equation) to be used is determined to construct the solver by different control equations; in step S102, preset operators (operators) and their function interfaces are set, and each term (term) in the control equation is transformed into a form represented by an operator, and different operators can be flexibly selected to transform the control equation; in step S103, the solution process is designed according to the control equation, and the solution program (solution procedure) can be directly generated according to the control equation and the function interfaces of its operators, thereby constituting the main framework of the solver. Further, the control equation is solved through the solution program, and the original variables are calculated and obtained; by freely selecting the control equation, independently designing the solution process, and flexibly selecting the operator, the solver can be adapted to the hydrodynamics numerical simulation problem to be solved, improving the adaptability of the solver. Further, independently constructing the solution process can optimize the calculation method of the solver by itself, effectively improving the calculation efficiency of the solver to reduce the time cost required for hydrodynamics numerical simulation; by transforming the control equation into the form of an operator, the control equation can be simplified, and the function interface corresponding to the operator can be directly called to generate the solution program, reducing the amount of code used to construct the solver, thereby quickly and simply constructing a solver that meets the actual needs of the application scenario;

[0064] When using the prior art, there are problems such as limited technical flexibility, inability to deeply customize the solution for hydrodynamics numerical simulation problems, and complex customization processes. By using the method for quickly constructing a computational fluid dynamics solver provided in Embodiment 1 of the present invention, the flexibility of solver construction is improved by freely selecting the control equation, independently designing the solution process, and flexibly selecting the operator. The control equation is simplified by transforming it into the form of an operator, and the function interface corresponding to the operator is directly called to generate the solution program to reduce the complexity of constructing the solver.

[0065] The method for quickly constructing a computational fluid dynamics solver provided in Embodiment 1 of the present invention, as Figure 2 shown, the main framework of the solver is the solution program, and the solution program can be jointly composed of several control equations and several terms. The terms constituting the solution program are used to calculate parameters required for different flow models and simulation methods such as viscous stress, sub-grid stress, stress work, and slope limiter; the solution equation is composed of several terms, and the terms are composed of several operators.

[0066] For the method for quickly constructing a computational fluid dynamics solver provided in Embodiment 1 of the present invention, when selecting the governing equations to be used, the governing equations and boundary conditions should be clarified; when constructing the solver, the solution method should be clarified. Specifically, it is necessary to have the final representation of the primitive variables and the operator granularity; it is necessary to have the conservation and vector form equations; the variables need to be set inside the derivative, and the constants need to be set outside the derivative; when using the solver, the numerical format should be clarified.

[0067] For the method for quickly constructing a computational fluid dynamics solver provided in Embodiment 1 of the present invention, preferably, the operators in step S102 include:

[0068] Ordinary divergence of single variable High-order interpolation divergence of two variables Limiter interpolation divergence of two variables ψ; Scalar divergence Ordinary Laplacian of single variable Δφ; Ordinary Laplacian of two variables Vector component source term q and scalar source term f i ; Left-end transient term Right-end transient term Gradient Calculate the time step Δt; Calculate the molecular viscous stress τ vis ; Calculate the subgrid stress τ sgs ;

[0069] Among them, the function interface of the ordinary divergence of single variable is div(Field <vector>phi), for momentum divergence, velocity divergence, and surface force work;

[0070] The function interface of the bivariate high-order interpolation divergence is div(Field <vector>rhoU, Field<scalar, T> phi, Field <vector>grad), for the convective term;

[0071] The function interface for the divergence of the bivariate limiter interpolation is div(Field <vector>rhoU,Field <scalar>phi,Field <vector>grad,Field <scalar>limter), for the convective term;

[0072] The function interface for the scalar divergence is div(Field <scalar>phi,vector <t>orient), for the pressure gradient in the momentum equation;

[0073] The function interface for the univariate ordinary Laplacian is laplas(Field <scalar>phi), for the pressure term and the incompressible energy heat flux;

[0074] The function interface for the bivariate ordinary Laplacian is laplas(Field <scalar>gamma,Field <scalar>phi), for the diffusion term;

[0075] The function interface for the vector component source term and the scalar source term is src(Field<scalar,T> q), which is used for body forces, reaction rates, and thermal radiation;

[0076] The function interface for the left - hand - side transient term is lddt(Field<scalar,T> phi), which is used for general transient terms;

[0077] The function interface for the right - hand - side transient term is rddt(Field<scalar,T> phi), which is used for the pressure time derivative in the total enthalpy equation;

[0078] The function interface for the gradient is grad(Field <scalar>phi), for various stresses, velocity gradients and heat fluxes required for high-order upwinding;

[0079] The function interface for calculating the time step is calcTimeStepFlowCFL(), which is used to calculate the time step;

[0080] The function interface for calculating the molecular viscous stress is calcVisStress(Field <tensor>&gradU,Field <scalar>nu), which is used to calculate the molecular viscous stress;

[0081] The function interface for calculating the sub-grid stress is calcSGStress(Field <tensor>&gradUTilda,Field <scalar>nuTilda), which is used to calculate the sub-grid stress;

[0082] According to different application scenarios, different operators can be selected to construct the selected governing equations, and the function interfaces corresponding to the operators can be directly called to generate the solution program, so as to quickly and simply construct a solver that meets the actual needs of the application scenario.

[0083] Such as Figure 3 As shown, for the method for quickly constructing a computational fluid dynamics solver provided in Embodiment 1 of the present invention, preferably, in step S103, when solving the governing equation, the governing equation needs to be iteratively updated, and the method for updating the governing equation includes the following steps:

[0084] S201: Calculate the time step;

[0085] S202: Calculate the original quantity or conserved quantity required for updating the governing equation;

[0086] S203: Calculate the gradient required for updating the governing equation;

[0087] S204: Initialize the array of the current time step; allocate space and set the values of the array of the current time step to zero;

[0088] S205: Update the spatial terms, including the source term and the flux term, and stack the content into the array step by step;

[0089] S206: Update the time term, process the sum of the spatial terms, and stack the value of the previous step array;

[0090] S207: Update the numerical value of the previous step result;

[0091] S208: Calculate the output quantity.

[0092] For the method for quickly constructing a computational fluid dynamics solver provided in Embodiment 1 of the present invention, in steps S201 to S203, the pre-parameters required for iteratively updating the governing equation are calculated; in steps S204 to S207, the governing equation is iteratively calculated; in step S208, the original variables obtained by calculating the governing equation are output.

[0093] For the method for quickly constructing a computational fluid dynamics solver provided in Embodiment 1 of the present invention, preferably, the types of governing equations applicable to the operator include:

[0094] Continuity equation:

[0095] Pressure elliptic equation:

[0096] Pressure parabolic equation:

[0097] Compressible momentum equation, u - component, pressure - gradient form:

[0098] Compressible momentum equation, u - component, pressure - divergence form:

[0099] Incompressible momentum equation, u - component:

[0100] Incompressible energy equation:

[0101] Scalar equation:

[0102] Total - energy equation:

[0103] Total - enthalpy equation:

[0104] RANS turbulence - model k - equation:

[0105] RANS turbulence - model ε - equation:

[0106] Chemical - reaction species equation:

[0107] VOF volume - fraction equation:

[0108] Where ρ is density; t is time; v is velocity; u is the x - direction component of velocity; is the Laplace operator; p is pressure; f is the force source term; T is temperature; h is total enthalpy; φ is scalar; k is turbulent kinetic energy; Y k is the species mass fraction; α is the multiphase - flow volume fraction;

[0109] By selecting the governing equations in the above - mentioned covered fields and through mathematical derivation to form physical and mathematical models, the governing equations used to solve the fluid - mechanics numerical - simulation problems can be determined, and a solver can be constructed through different solution processes.

[0110] Example 2:

[0111] Taking the incompressible solution algorithm of the artificial - compressibility method as an example, a computational fluid - dynamics solver is constructed by using the fast - construction method of the computational fluid - dynamics solver provided in Example 1.

[0112] The governing equations are determined as

[0113] The solution process is determined, including:

[0114] term1: Calculate the velocity gradient;

[0115] term2: Calculate the stress tensor, and the three stress tensor components in the x - direction are

[0117] equation1: Solve the pressure equation:

[0118] equation2: Solve the momentum equation:

[0119] term3: Calculate the primitive variables;

[0120] Generate a solution program according to the solution process and the function interfaces corresponding to the operators:

[0121] 1. Declare the flow - field variables

[0122] Declare the iterative variables:

[0123] ScalarDField rho("rho”);

[0124] VectorDField rhoU();

[0125] Declare the non - iterative variables:

[0126] ScalarBField p;

[0127] VectorBField U;

[0128] 2. Solve the pressure equation

[0129] using namespace EPL;

[0130] Calculate the velocity divergence term:

[0131] p -= RT * div(U);

[0132] Calculate the pressure Laplacian term:

[0133] p += nu * lapl(p);

[0134] Calculate the transient term:

[0135] p += lddt(p);

[0136] 3. Solve the momentum equation

[0137] Calculate the convection term:

[0138] rhoU -= div(rhoU, u, gradu);

[0139] Calculate the pressure gradient term:

[0140] rhoU- = grid(p);

[0141] Calculate the diffusion term:

[0142] rhoU+ = div(tau);

[0143] Calculate the transient term:

[0144] rhoU+ = lddt(rhoU);

[0145] 4. Construct the solver

[0146] Declare the data field and parameters:

[0147] #include "declareFieldPrimitive.h"

[0148] #include "declareFieldDerived.h"

[0149] #include "declareProperties.h"

[0150] Iterative solver:

[0151] while (runController.advance()){

[0152] Calculate the time step:

[0153] deltaT = calaTimeStep();

[0154] Calculate the gradient:

[0155] gradU = FVM::grad(U);

[0156] Calculate the molecular viscous stress:

[0157] tau = calcVisStress(gradU, mu);

[0158] Update the pressure equation:

[0159] #include "pEquation.h"

[0160] Update the momentum equation:

[0161] #include "UEquation.h"

[0162] Update the primitive variables:

[0163] rho = p / RT;

[0164] u = rhou / u;

[0165] v = rhov / v;

[0166] w = rhow / w;}

[0167] Through the above solution procedure, the density rho and the velocities u, v, and w of the fluid in the x, y, and z directions can be calculated.

[0168] In summary, the method for quickly constructing a computational fluid dynamics solver provided by the present invention can solve the problems of limited flexibility in the prior art, inability to deeply perform customized solutions for numerical simulation problems of fluid mechanics, and complex customization processes; it can improve the flexibility in solving numerical simulation problems of fluid mechanics and simplify the customization process.

[0169] Those skilled in the art should understand that those skilled in the art can implement the described variations in combination with the prior art and the above embodiments, which will not be elaborated here. Such variations do not affect the essence of the present invention and will not be elaborated here.

[0170] The preferred embodiments of the present invention have been described above. It should be understood that the present invention is not limited to the above specific embodiments. The devices and structures not described in detail should be understood to be implemented in a common manner in the art; any person skilled in the art can make many possible changes and modifications without departing from the technical solution of the present invention, or modify it into an equivalent embodiment with equivalent changes, which does not affect the essence of the present invention. Therefore, any simple modification, equivalent change, and modification made to the above embodiments based on the technical essence of the present invention without departing from the technical solution of the present invention still fall within the scope of protection of the technical solution of the present invention.< / scalar> < / tensor> < / scalar> < / tensor> < / scalar> < / scalar> < / scalar> < / scalar> < / t> < / scalar> < / scalar> < / vector> < / scalar> < / vector> < / vector> < / vector> < / vector> < / scalar> < / tensor> < / scalar> < / tensor> < / scalar> < / scalar> < / scalar> < / scalar> < / t> < / scalar> < / scalar> < / vector> < / scalar> < / vector> < / vector> < / vector> < / vector>

Claims

1. A method for quickly constructing a computational fluid dynamics solver, characterized in that: The following steps are involved: S101: Determine the control equation required for calculation; S102: Preset the operator and its function interface; convert each term in the control equation into an operator; S103: Generate a solution program by calling the corresponding function interface according to the operator in the control equation, solve the control equation according to the solution program, and calculate and obtain the original variables.

2. The method for quickly constructing a computational fluid dynamics solver according to claim 1, characterized in that: The operator in step S102 includes: Univariate General Divergence Bivariate High-Order Interpolation Divergence Bivariate Limiter Interpolation Divergence Scalar divergence Univariate ordinary Laplace Δφ; bivariate ordinary Laplace Vector component source term q and scalar source term f i ; Left-hand transient term Right-hand transient term gradient Calculate the time step Δt; calculate the molecular viscosity stress τ vis ; Calculate the subgrid stress τ sgs ; Among them, the function interface of the single variable common divergence is div(Field <vector> phi);< / vector> The function interface of the bivariate high-order interpolation divergence is div(Field <vector>rhoU,Field<scalar,T>phi,Field <vector> grad);< / vector> < / vector> The function interface of the bivariate limiter interpolation divergence is div(Field <vector>rhoU,Field <scalar>field <vector>grad,Field <scalar> limter);< / scalar> < / vector> < / scalar> < / vector> The function interface of scalar divergence is div(Field <scalar>phi,vector <t> orientation);< / t> < / scalar> The function interface of the univariate ordinary Laplace is laplas(Field <scalar> phi);< / scalar> The function interface of the two-variable ordinary Laplace is laplas(Field <scalar>gamma,Field <scalar> phi);< / scalar> < / scalar> The function interface for vector component source items and scalar source items is src(Field<scalar,T> q); The function interface of the left transient item is lddt(Field<scalar,T> phi); The function interface of the transient item on the right is rddt(Field<scalar,T> phi); The gradient function interface is grad(Field <scalar> phi);< / scalar> The function interface for calculating the time step is calcTimeStepFlowCFL(); The function interface for calculating molecular viscous stress is calcVisStress(Field <tensor> &gradU,< / tensor> Field <scalar> nu);< / scalar> The function interface for calculating subgrid stress is calcSGStress(Field <tensor>&gradUTilda,Field <scalar> nuTilda).< / scalar> < / tensor> 3. The method for quickly constructing a computational fluid dynamics solver according to claim 1, characterized in that: In step S103, when solving the control equation, the control equation needs to be iteratively updated. The method for updating the control equation includes the following steps: S201: Calculate time step; S202: Calculate the input original quantity or conservation quantity required for updating the control equation; S203: Calculate the gradient required for updating the control equation; S204: Initialize the array of the current time step; create space and set the value of the array of the current time step to zero; S205: updating the spatial items, including the source items and the flux items, and adding the contents to the array step by step; S206: Update the time item, process the sum of the space item, and add the value of the array in the previous step; S207: Update the result value of the previous step; S208: Calculate the output amount.

4. The method for quickly constructing a computational fluid dynamics solver according to claim 2, wherein: The types of control equations applicable to the operator include: Continuity equation: Pressure ellipse equation: Pressure parabolic equation: The compressible momentum equation u component, pressure gradient component form: The compressible momentum equation u component, pressure divergence form: The u component of the incompressible momentum equation: Incompressible energy equation: Scalar equation: Total energy equation: Total enthalpy equation: RANS turbulence model k equation: RANS turbulence model∈Equation: Chemical reaction component equation: VOF volume fraction equation: Among them, ρ is density; t is time; v is velocity; u is the component of velocity in the x direction; is the Laplace operator; p is the pressure; f is the force source term; T is the temperature; h is the total enthalpy; φ is a scalar; k is the turbulent kinetic energy; Y k is the component mass fraction; α is the multiphase flow volume fraction.

Images