Database accelerated retrieval method suitable for flame surface model

Abstract

The invention belongs to the field of engine combustion and computational fluid mechanics, and discloses a database accelerated retrieval method suitable for a flame surface model. The retrieval method comprises the following steps: obtaining retrieval parameters of a local grid from flow field calculation; determining initial values of retrieval parameters and a retrieval method according to the retrieval values of the adjacent grids and the time steps; obtaining a final retrieval value of each retrieval parameter of the local grid; obtaining the mass fraction of each component of the local grid through interpolation; and traversing all grids until all grid points are retrieved. According to the database accelerated retrieval method suitable for the flame surface model, the initial value of the retrieval variable is determined according to the retrieval values of the adjacent grids and the adjacent time steps based on the characteristic that the adjacent grids and the adjacent time steps have similar physical characteristics, so that the range of variable retrieval is greatly reduced, the retrieval time is shortened, and the retrieval efficiency is improved. And the calculation efficiency of the flame surface combustion model is obviously improved.

Description

Technical Field

[0001] The present invention belongs to the fields of engine combustion and computational fluid dynamics, and particularly relates to a database acceleration retrieval method applicable to a flamelet model. Background Art

[0002] The key to numerical simulation of an engine combustion chamber is the turbulent combustion model. Due to its high calculation accuracy and efficiency, the flamelet combustion model is most widely used in numerical simulation of an engine combustion chamber.

[0003] Before calculating the flow field, the flamelet combustion model needs to establish a flamelet database based on different retrieval variables. During calculation, the mass fractions of each component are retrieved and interpolated according to the flow field parameters of the local grid. There are typically 4 retrieval variables, namely the mean and variance of the mixture fraction, and the mean and variance of the reaction progress variable. With the continuous development of the flamelet model, the number of retrieval variables has also been increasing. For example, considering an unsteady flamelet requires adding the scalar dissipation rate, considering radiation requires adding enthalpy, and considering compressibility requires adding pressure.

[0004] When calculating with the flamelet combustion model, all grids need to be traversed, and each retrieval variable needs to be retrieved sequentially at each grid. The more retrieval variables there are, the longer the retrieval time required. Therefore, it is necessary to improve the existing flamelet database retrieval method to enhance the calculation efficiency of the flamelet combustion model.

[0005] Currently, there is an urgent need to develop a database acceleration retrieval method applicable to the flamelet model. Summary of the Invention

[0006] The technical problem to be solved by the present invention is to provide a database acceleration retrieval method applicable to the flamelet model to overcome the defects of the prior art.

[0007] The database acceleration retrieval method applicable to the flamelet model of the present invention assumes that a flamelet database has 、 、 and four retrieval variables: ; wherein, is the mass fraction of each component, and are respectively the mean and variance of the mixture fraction, and are respectively the mean and variance of the reaction progress variable; Assume that the discrete numbers of each retrieval variable when establishing the database are respectively Nf , Nvarf ,NC , NvarC , that is: ; Since the adjacent grids on the computing node are physically adjacent, according to the locality and continuity characteristics of the grids, the adjacent grids have similar physical characteristics. At the same time, the same grid also has similar physical characteristics in the current time step and the previous time step; The database acceleration retrieval method applicable to the flamelet model includes the following steps: S10. Obtain the retrieval parameters of the local grid from the flow field calculation; Obtain the mean value of the mixture fraction of the local grid from the flow field calculation f m and variance varf m as well as the mean value of the reaction progress variable C m and variance varC m ; S20. Determine the initial values of the retrieval parameters and the retrieval method according to the retrieval values of the adjacent grids and time steps; Assume that the retrieval values of the mean mixture fraction in the previous time step are respectively ipref 0 and ipref 1 , and the retrieval values of the previous grid point are respectively iupf 0 and iupf 1 , then assign the initial values to the retrieval parameters according to the preset method and perform the retrieval; S30. Obtain the final retrieval values of each retrieval parameter of the local grid; Retrieve the mean value and variance of the mixture fraction, and the mean value and variance of the reaction progress variable respectively according to the preset method in S20 to obtain the final retrieval values, so that the finally obtained retrieval values satisfy , and ; , and ; , and ; , and ; S40. Obtain the mass fractions of each component of the local grid by interpolation; According to the retrieval values obtained in S20 and S30, obtain the mass fractions of each component of the local grid by interpolation ; S50. Traverse all grids until the retrieval of all grid points is completed; Repeat S10~S40 for each grid point until the retrieval of all grid points is completed.

[0008] Further, the preset method in S20 assigns initial values to the retrieval variables according to the following rules and conducts the retrieval; for the mixture fraction: S21. If , then let , and conduct the retrieval in ascending order; S22. If , then let , and conduct the retrieval in descending order; S23. In the remaining cases, then let , , and conduct the retrieval using the bisection method.

[0009] The database acceleration retrieval method applicable to the flamelet model of the present invention is based on the characteristics that adjacent grids and adjacent time steps have similar physical characteristics, and determines the initial values and retrieval methods of the retrieval variables according to the retrieval values of adjacent grids and adjacent time steps, greatly narrowing the range of variable retrieval, reducing the retrieval time, and significantly improving the calculation efficiency of the flamelet combustion model. BRIEF DESCRIPTION OF THE DRAWINGS

[0010] Figure 1 is a flowchart of the database acceleration retrieval method applicable to the flamelet model of the present invention; Figure 2 is a schematic diagram of a Bunsen burner flame in an embodiment; Figure 2 In d p is the diameter of the pilot flame; D is the inner diameter of the central jet; x is the axial direction; r is the radial direction; U is the axial velocity; V is the radial velocity; Figure 3a is a comparison curve of the typical cross-section temperature distributions obtained by using two flamelet database retrieval methods (x = 6.5D); Figure 3b is a comparison curve of the typical cross-section temperature distributions obtained by using two flamelet database retrieval methods (x = 8.5D). DETAILED DESCRIPTION OF THE EMBODIMENTS

[0011] The present invention will be described in detail below with reference to the drawings and embodiments.

[0012] Embodiment: As Figure 1 shown, for the database acceleration retrieval method applicable to the flamelet model in this embodiment, assume that a flamelet database has , , and four retrieval variables: ; wherein, is the mass fraction of each component, and are the mean and variance of the mixture fraction respectively, and are the mean and variance of the reaction progress variable respectively; Assume that the discrete numbers of each retrieval variable when establishing the database are respectively Nf , Nvarf , NC , NvarC , that is: ; Since the adjacent grids on the computing node are physically adjacent, according to the locality and continuity characteristics of the grids, the adjacent grids have similar physical characteristics. At the same time, the same grid also has similar physical characteristics at the current time step and the previous time step; The database acceleration retrieval method applicable to the flamelet model includes the following steps: S10. Obtain the retrieval parameters of the local grid from the flow field calculation; Obtain the mean f m and variance varf m of the mixture fraction and the mean C m and variance varC m of the reaction progress variable of the local grid from the flow field calculation; S20. Determine the initial values of the retrieval parameters and the retrieval method according to the retrieval values of the adjacent grids and time steps; Assume that the retrieval values of the mixture fraction mean at the previous time step are respectively ipref 0 and ipref 1 , and the retrieval values of the previous grid point are respectively iupf 0 and iupf 1 , then assign initial values to the retrieval parameters according to the preset method and perform retrieval; S30. Obtain the final retrieval values of each retrieval parameter of the local grid; Retrieve the mean and variance of the mixture fraction and the mean and variance of the reaction progress variable respectively according to the preset method in S20 to obtain the final retrieval values, so that the finally obtained retrieval values satisfy , and ; , and ; , and ; , and ; S40. Obtain the mass fractions of each component in the local grid by interpolation; Obtain the mass fractions of each component in the local grid by interpolation according to the retrieval values obtained in S20 and S30 ; S50. Traverse all grids until the retrieval of all grid points is completed; Repeat S10 - S40 for each grid point until the retrieval of all grid points is completed.

[0013] Furthermore, the preset method in S20 assigns initial values to the retrieval variables and performs retrieval according to the following rules; for the mixture fraction: S21. If , then let Retrieve in ascending order; S22. If , then let , retrieve in descending order; S23. In the remaining cases, then let , , retrieve by the bisection method.

[0014] Comparative example: The retrieval method of the flamelet database in this comparative example is the bisection method, and the specific steps are as follows: Step1. Obtain the mean value of the mixture fraction f m and variance varf m of the local grid from the flow field calculation, as well as the mean value of the reaction progress variable C m and variance varC m ; Step2. Set the initial retrieval values of the mean and variance of the mixture fraction, and the mean and variance of the reaction progress variable to be if 0 = 1, if 1 = Nf , ivarf 0 = 1, ivarf 1 = Nvarf , iC 0 = 1, iC 1 = NC , ivarC 0 = 1, ivarC 1 = NvarC ; Step 3. Obtain the final search value by looking up the table using the bisection method, such that , and ; , and ; , and ; , and ; Step 4. According to the search value obtained in Step 3, interpolate to obtain the mass fractions of each component in the local grid ; Step 5. Repeat Steps 1 to 4 for each grid point until the search of each grid point is completed.

[0015] Next, the database acceleration search method applicable to the flamelet model of the present invention and the bisection method of the comparative example are tested through a Bunsen burner flame.

[0016] The Bunsen burner flame is shown in Figure 2 . The central jet is a methane / air premixed gas with an equivalence ratio of . The inner diameter D = 12 mm, and the average velocity of the jet is U 0 = 30 m / s. The Reynolds number Re of the central jet is approximately 24,200. Surrounding the jet is a pilot flame of methane / air burned gas with an equivalence ratio ϕ = 1. The diameter of the pilot flame is d p = 68 mm, and the average velocity is 1.5 m / s. Water cooling is provided outside the pilot flame to make the inlet temperature of the pilot flame lower than the corresponding adiabatic flame temperature. The outside of the burner is an air coflow with an average velocity of 0.22 m / s.

[0017] To reduce the computational amount, the computational domain is one-fourth of a cylinder, that is, a sector cylinder with a bottom surface of 90°. The number of grids is about 140,000. The flamelet combustion model and the methane Grimech 2.11 reaction mechanism are used for calculation. The search variables of the flamelet database include the mean and variance of the mixture fraction and the mean and variance of the reaction progress variable. The database acceleration search method applicable to the flamelet model of the present invention and the bisection method of the comparative example are respectively used for searching.

[0018] Figure 3a 、 Figure 3bThe contrast curve of the database acceleration retrieval method applicable to the flamelet model of the present invention and the typical cross-sectional temperature distribution obtained by the dichotomy of the comparative example is given. It can be seen that the calculation results obtained by the two retrieval methods completely coincide and are in good agreement with the experimental results, indicating that the database acceleration retrieval method applicable to the flamelet model of the present invention has no impact on the accuracy of numerical simulation. At the same time, Table 1 gives the time comparison of numerical simulation under the two retrieval methods. It can be seen that the calculation time using the database acceleration retrieval method applicable to the flamelet model of the present invention is shortened by about 24% compared with the dichotomy of the comparative example, and the calculation efficiency is greatly improved.

[0019] Table 1 Comparison Table of Calculation Times

[0020] Although the embodiments of the present invention have been disclosed as above, they are not limited to the applications listed in the specification and embodiments. For those skilled in the art, without departing from the principle of the present invention, all the features disclosed in the present invention, or all the steps in the disclosed methods or processes, except for mutually exclusive features and / or steps, can be combined in any way. The present invention is not limited to the specific details and the illustrated examples here.

Claims

1. A database accelerated retrieval method suitable for flame surface models, characterized in that: Assume that a flame surface database has , , and Four search variables: ； in, is the mass fraction of each component, and are the mean and variance of the mixture fraction, and are the mean and variance of the reaction progress variable, respectively; Assume that when establishing the database, the discrete number of each search variable is Nf , Nv , NC , NvC ,Right now: ； Since adjacent grids on the computing nodes are physically adjacent, according to the locality and continuity characteristics of the grids, adjacent grids have similar physical characteristics. At the same time, the same grid also has similar physical characteristics in the current time step and the previous time step; The database accelerated retrieval method applicable to the flame surface model comprises the following steps: S10. Obtaining retrieval parameters of the local grid from the flow field calculation; Obtain the mean value of the mixture fraction on the local grid from the flow calculation f m and variance varf m and the mean of the reaction progress variable C m and variance varC m ; S20. Determine the initial value and retrieval method of the retrieval parameter according to the retrieval value of the adjacent grid and the time step; Assume that the retrieval values ​​of the mean of the mixture fraction at the previous time step are ipref 0 and ipref 1, the retrieval values ​​of the previous grid point are iupf 0 and iupf 1, then assign initial values ​​to the search parameters according to the preset method and perform the search; S30. Obtain the final search value of each search parameter of the local grid; According to the preset method of S20, the mean and variance of the mixed fraction and the mean and variance of the reaction progress variable are retrieved to obtain the final retrieval value, so that the final retrieval value satisfies ,and ; ,and ; ,and ; ,and ; S40. Obtain the mass fraction of each component of the local grid by interpolation; According to the retrieved values ​​obtained by S20 and S30, the mass scores of each component in the local grid are obtained by interpolation. ; S50. Traverse all grids until all grid points are retrieved; Repeat S10 to S40 for each grid point until all grid points are retrieved.

2. The database accelerated retrieval method applicable to the flame surface model according to claim 1, characterized in that: The preset method in S20 is to assign initial values ​​to the search variables and perform the search according to the following rules; for mixed fractions: S21. If , then let , search in order from small to large; S22. If , then let , search in order from large to small; S23. In the remaining cases, let , , search according to the binary method.

Images