A grid-adaptive turbulence simulation method based on a Vreman dynamic coefficient coupled RSM model

Abstract

The present invention discloses a grid adaptive turbulence simulation method based on the Vreman dynamic coefficient coupled RSM model, and the main process of its implementation includes: step one, judging whether to apply the shielding function; step two, calculating the Vreman dynamic coefficient; step three, determining the local dynamic grid length scale; step four, constructing the scale-related adjustment function; step five, reconstructing the Reynolds stress tensor and turbulent viscosity of the RSM model; step six, using the reconstructed Reynolds stress tensor and turbulent viscosity to simulate turbulence. The present invention constructs the Vreman dynamic coefficient based on the local velocity gradient tensor, identifies the local grid size, and determines the local grid length scale Δ * , the Reynolds stress tensor and turbulent viscosity are reconstructed by constructing scale-related functions through turbulent energy spectrum integration, and grid adaptive simulation is realized, which effectively overcomes the problem that the existing turbulent mixing model is highly dependent on grid experience, broadens the grid adaptability, improves the calculation accuracy, reduces the calculation cost, and speeds up the turbulence simulation process.

Description

Technical Field

[0001] The present invention relates to the field of complex flow calculation of aircraft engines and gas turbines, and in particular to a grid adaptive turbulence simulation method based on a Vreman dynamic coefficient coupled RSM model. Background Art

[0002] Aircraft engines and gas turbines are a major national demand and a key development area for my country's high-end equipment manufacturing. Turbulence is ubiquitous in nature and engineering, and accurate prediction of turbulence is a major challenge in studying complex flow problems. Turbulent flows in aircraft engines and gas turbines are extremely complex, with multi-scale, nonlinear, and unsteady characteristics. These complex turbulences have a great impact on the performance of aircraft engines and gas turbines. Therefore, in engineering design, it is urgent to develop turbulence simulation methods with high prediction accuracy and high computational efficiency, so that they can accurately simulate turbulent flows in complex engineering problems such as the internal flow of aircraft engines and gas turbines.

[0003] The turbulence simulation method commonly used in existing engineering projects is mainly based on solving the Reynolds-averaged NS equations (RANS) method. Although the amount of calculation is small, it is based on simple basic flows and has poor prediction accuracy for complex turbulent flows such as multi-scale, unsteady, and large separation in engineering problems. Therefore, it is difficult to achieve a detailed study of the complex flow mechanism in engineering problems such as the internal flow of aircraft engines and gas turbines, which seriously restricts the improvement of the design level in the fields of aircraft engines and gas turbines. As a high-precision numerical simulation method, the large eddy simulation (LES) method has a high requirement for the number of grids, which makes its computational cost increase exponentially compared to the RANS method, which is much higher than the computational cost that can be tolerated in engineering applications. For engineering flow problems such as the internal flow of aircraft engines and gas turbines, the Reynolds number is often high. Under the existing computing power, the LES method is still difficult to apply to the flow prediction in complex engineering fields at the engineering design level.

[0004] In the past two decades, the RANS-LES hybrid simulation method was born. By using the RANS method in the near-wall area and the LES method in the mainstream area, the calculation accuracy and efficiency are balanced, which provides a good solution to the problem of high-precision simulation and high computational cost of complex flows. However, the current classic RANS-LES hybrid model has relatively strict requirements on the grid, requiring users to have rich experience in high-precision numerical simulation. The existing RANS-LES hybrid simulation method is highly dependent on the experience-based grid distribution for the prediction accuracy of complex engineering flows such as the internal flow of aircraft engines and gas turbines, and the grid demand is large, making it difficult to apply to daily engineering design. Therefore, reducing the empirical dependence of the RANS-LES hybrid model on the grid is of great significance to better realize the accurate, efficient and high-precision simulation of multi-scale, nonlinear, unsteady and other turbulent flow phenomena in complex engineering flow problems.

[0005] The present invention further improves the prediction accuracy and grid adaptability of the grid adaptive turbulence simulation method by introducing the Vreman dynamic coefficient. Compared with the disclosed invention patent CN 115186608 A, the prediction accuracy of the grid adaptive turbulence simulation method can be further improved and the calculation cost can be greatly reduced. Summary of the invention

[0006] 1. Technical issues to be resolved

[0007] The purpose of the present invention is to propose a grid-adaptive turbulence simulation method based on the Vreman dynamic coefficient coupled RSM model, which further improves the grid adaptability by introducing the Vreman dynamic coefficient, effectively overcomes the problem of the existing RANS-LES hybrid model's high empirical dependence on the grid, and while improving the calculation accuracy, greatly reduces the calculation cost, significantly speeds up the turbulence simulation process, and more quickly stimulates the unsteady characteristics of the flow, providing an efficient numerical simulation method for solving the fast and high-precision simulation of multi-scale, nonlinear, unsteady and other turbulent flows in complex engineering flow problems.

[0008] (II) Technical solution

[0009] In order to solve the above technical problems, the present invention provides a grid adaptive turbulence simulation method based on the Vreman dynamic coefficient coupled RSM model, comprising the following steps:

[0010] Step 1, determine whether to apply the shielding function;

[0011] Step 2, calculate the Vreman dynamic coefficient;

[0012] Step 3, determine the local dynamic grid length scale;

[0013] Step 4: construct a scale-dependent adjustment function;

[0014] Step 5, reconstruct the Reynolds stress tensor and turbulent viscosity of the RSM model;

[0015] Step 6, use the reconstructed Reynolds stress tensor and turbulent viscosity to perform turbulence simulation;

[0016] ① The determination of whether to apply the shielding function includes:

[0017] Combined with the type of flow state being simulated, determine whether to use the shielding function F GAS Specifically, when the flow state type is free shear flow, the shielding function is not used. In this case, the shielding function F GAS = 0; when the flow state type is near-wall flow, a shielding function is used, and the shielding function F GAS Choose one from the F in the DDES-SST model 1 Shielding function, F 2 The screening function and F from the DDES-SA model d Masking function;

[0018] ② The calculation of the Vreman dynamic coefficient includes:

[0019] Based on the local velocity gradient tensor α ij , construct the Vreman dynamic coefficient C DGAS ;

[0020] The Vreman dynamic coefficient C DGAS It is obtained from the following formula:

[0021]

[0022] β ij =α mi α mj

[0023]

[0024] Where S is the strain rate, C GAS The empirical coefficient is taken as 0.6;

[0025] ③ The determination of the local dynamic grid length scale includes:

[0026] Combined with the shielding function F described in step 1 GAS , and the Vreman dynamic coefficient C described in step 2 DGAS , determine the local dynamic grid length scale Δ * ;

[0027] The local dynamic grid length scale Δ* is given by:

[0028] Δ * =C DGAS [(1-F GAS )Δ vol +F GAS Δ max ]

[0029] Δ max =max(Δ x ,Δ y ,Δ z )

[0030]

[0031] Among them, Δ x is the length of the local hexahedral grid, Δ y is the width of the local hexahedral grid, Δ z is the height of the local hexahedral grid;

[0032] ④ The structural scale-related adjustment functions include:

[0033] According to the modeling method of turbulent kinetic energy in the RSM model, by solving the trace of the Reynolds stress tensor, the original modeled turbulent kinetic energy k is obtained. m ;

[0034] According to the local dynamic grid length scale Δ * Based on the turbulence energy spectrum, the actual simulated turbulent kinetic energy k is obtained by integration u ;

[0035] The actual simulated turbulent kinetic energy k u It is obtained from the following formula:

[0036]

[0037] Among them, C k is the Kolmogorov constant coefficient, which is taken as 1.5; ε is the actual turbulence dissipation rate; κ c is the solvable turbulence cutoff wave number, which is the dynamic grid length scale Δ obtained in step 3 * Determine as follows:

[0038]

[0039] Among them, π is the ratio of pi, which is 3.14;

[0040] According to the actual simulated turbulent kinetic energy k u , the original modeled turbulent kinetic energy k m and the shielding function F described in step 1 GAS, construct the dynamic scale-dependent regulation function D f ; Define the scale ratio as the actual simulated turbulent kinetic energy k u and the original modeled turbulent kinetic energy k m The dynamic scale-dependent regulation function D f is a function related to the ratio of the scales, which is obtained by the following formula:

[0041]

[0042] l GAS =(1-F GAS ) u +F GAS l m

[0043]

[0044] Among them, l u is the grid-related scale, l m is the turbulence length scale given by the RSM model;

[0045] ⑤ The Reynolds stress tensor and turbulent viscosity of the reconstructed RSM model include:

[0046] Adopt the dynamic scale-dependent adjustment function D described in step 4 f The Reynolds stress tensor in the RSM model Regulate and obtain the reconstructed Reynolds stress tensor It is derived from the following formula:

[0047]

[0048] Use the dynamic scale-dependent adjustment function D described in step 4 f For the turbulent viscosity ν in the RSM model t By adjusting the turbulent viscosity ν, we can obtain the reconstructed turbulent viscosity ν. sfs , obtained from the following formula:

[0049] ν sfs =D f ·ν t

[0050] ⑥ The turbulence simulation using the reconstructed Reynolds stress tensor and turbulent viscosity includes:

[0051] Using the reconstructed Reynolds stress tensor described in step 5 and turbulent viscosity ν sfs , the transport equation of the RSM model is updated, replacing the Reynolds stress tensor in the original transport equation of the RSM model and turbulent viscosity ν t, and combined with the RSM model to obtain the grid adaptive turbulence simulation method based on the Vreman dynamic coefficient coupled RSM model.

[0052] (III) Beneficial effects

[0053] The grid adaptive turbulence simulation method based on the Vreman dynamic coefficient coupled RSM model provided by the present invention has the following beneficial effects: constructing the Vreman dynamic coefficient based on the local velocity gradient tensor, determining the local grid length scale Δ by identifying the local grid size * , and then the scale-related function is constructed by integrating the turbulent energy spectrum to reconstruct the Reynolds stress tensor and turbulent viscosity, realizing grid adaptive simulation and effectively overcoming the problem of high empirical dependence of the existing RANS-LES hybrid model on the grid.

[0054] Compared with the prior art CN 115186608 A, the grid adaptability of the grid adaptive turbulence simulation method is further broadened by introducing the Vreman dynamic coefficient. While improving the calculation accuracy, the calculation cost is greatly reduced, and the turbulence simulation process is significantly accelerated. This provides an efficient numerical simulation method for fast and high-precision simulation of multi-scale, nonlinear, unsteady and other turbulent flows in complex engineering flow problems.

[0055] The method of the present invention is concise in form and highly portable, can be well combined with the RSM turbulence model, is easy to be implanted into existing CFD codes for application and expansion, and has broad academic and engineering application prospects. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] Figure 1 It is a flow chart of a grid adaptive turbulence simulation method based on Vreman dynamic coefficient coupled RSM model of the present invention;

[0057] Figure 2 A dynamic scale-dependent adjustment function D is constructed for a grid adaptive turbulence simulation method based on a Vreman dynamic coefficient coupled RSM model according to the present invention. f Schematic diagram of

[0058] Figure 3 Schematic diagram of the turbulent vortex structure of the back-step example calculated using the LRR-RSM model;

[0059] Figure 4 The figure is a schematic diagram of the turbulent vortex structure of a back-step example calculated by a grid adaptive turbulence simulation method based on the Vreman dynamic coefficient coupled RSM model of the present invention. DETAILED DESCRIPTION

[0060] The following is a further detailed description of the specific implementation of the present invention by taking the LRR-RSM turbulence model as an example in conjunction with the accompanying drawings. The following examples are only used to illustrate the present invention, but are not used to limit the scope of the present invention.

[0061] The back step example is selected as the calculation example, and the high-quality hexahedral mesh is used to spatially discretize the calculation domain. The total number of meshes is about 410,000.

[0062] The present invention provides a grid adaptive turbulence simulation method based on a Vreman dynamic coefficient coupled RSM model, comprising the following steps:

[0063] Step 1, determine whether to apply the shielding function;

[0064] In this step, the type of flow state being simulated is considered to determine whether to use the screening function F GAS Specifically, when the flow state type is free shear flow, the shielding function is not used. In this case, the shielding function F GAS = 0; when the flow state type is near-wall flow, a shielding function is used, and the shielding function F GAS Choose one from the F in the DDES-SST model 1 Shielding function, F 2 The screening function and F from the DDES-SA model d Masking function.

[0065] In this embodiment, the F in the DDES-SA model is used. d Examples of functions include:

[0066] F GAS =F d

[0067]

[0068] Among them, κ is the Karman constant, which is 0.41; C d1 , C d2 are empirical coefficients, which are 20 and 3 respectively; U i,j is the velocity gradient tensor; ν is the viscosity coefficient of the fluid, d is the distance from the local grid to the wall; turbulent viscosity ν t It can be obtained by the LRR-RSM model as follows:

[0069]

[0070] Among them, C μ is the model constant, which is taken as 0.09; k m is the original modeled turbulent kinetic energy obtained according to the LRR-RSM model, ε mis the dissipation rate of the original model obtained according to the LRR-RSM model.

[0071] Step 2, calculate the Vreman dynamic coefficient;

[0072] In this step, based on the local velocity gradient tensor α ij , construct the Vreman dynamic coefficient C DGAS ;

[0073] The Vreman dynamic coefficient C DGAS It is obtained from the following formula:

[0074]

[0075] β ij =α mi α mj

[0076]

[0077] Where S is the strain rate, C GAS The empirical coefficient is taken as 0.6.

[0078] Step 3, determine the local dynamic grid length scale;

[0079] In this step, combined with the shielding function F in step 1 GAS , and the Vreman dynamic coefficient C in step 2 DGAS , we can get the local dynamic grid length scale Δ * ;

[0080] Local dynamic grid length scale Δ * Given by:

[0081] Δ * =C DGAS [(1-F GAS )Δ vol +F GAS Δ max ]

[0082] Δ max =max(Δ x ,Δ y ,Δ z )

[0083]

[0084] where Δ x is the length of the local hexahedral grid, Δ y is the width of the local hexahedral grid, Δ z is the height of the local hexahedral grid.

[0085] Step 4: construct a scale-dependent adjustment function;

[0086] In this step, according to the modeling method of turbulent kinetic energy in the RSM series turbulence model, the original modeled turbulent kinetic energy k is obtained by solving the trace of the Reynolds stress tensor. m In this embodiment, taking the LRR-RSM turbulence model in the RSM series model as an example, according to the modeling method of the turbulent kinetic energy in the LRR-RSM turbulence model, by solving the trace of the Reynolds stress tensor, the original modeled turbulent kinetic energy k is obtained. m .

[0087] According to the local dynamic grid length scale Δ in step 3 * Based on the turbulence energy spectrum, the actual simulated turbulent kinetic energy k is obtained by integration u , is given by:

[0088]

[0089] Among them, C k is the Kolmogorov constant coefficient, which is taken as 1.5; ε is the actual turbulence dissipation rate; κ c To solve the turbulent cutoff wave number, the dynamic grid length scale Δ in step 3 is * Determine as follows:

[0090]

[0091] Among them, π is the ratio of pi, which is 3.14;

[0092] According to the actual simulated turbulent kinetic energy k u , original modeled turbulent kinetic energy k m and the shielding function F in step 1 GAS , construct the dynamic scale-dependent regulation function D f ,like Figure 2 As shown; the scale ratio is defined as the actual simulated turbulent kinetic energy k u and the original modeled turbulent kinetic energy k m The dynamic scale-dependent regulation function D f is a function related to the scale ratio, which is obtained by the following formula:

[0093]

[0094] l GAS =(1-F GAS ) u +F GAS l m

[0095]

[0096] Among them, l u is the grid-related scale, l m is the turbulence length scale given by the LRR-RSM model, which is given by:

[0097]

[0098] Among them, ε m is the dissipation rate of the original model obtained according to the LRR-RSM model.

[0099] Step 5: Reconstruct the Reynolds stress tensor and turbulent viscosity of the RSM model:

[0100] In this step, the dynamic scale-dependent adjustment function D in step 4 is used f Reynolds stress tensor in RSM series turbulence models Regulate and obtain the reconstructed Reynolds stress tensor In this embodiment, the LRR-RSM turbulence model in the RSM series of models is taken as an example, and the Reynolds stress tensor in the RSM series of turbulence models is calculated. Regulate and obtain the reconstructed Reynolds stress tensor As shown below:

[0101]

[0102] Use the dynamic scale-dependent adjustment function D in step 4 f For the turbulent viscosity ν in the RSM series of turbulence models t By adjusting the turbulent viscosity ν, we can obtain the reconstructed turbulent viscosity ν. sfs In this embodiment, the LRR-RSM turbulence model in the RSM series model is taken as an example, and the turbulent viscosity ν in the LRR-RSM model is t By adjusting the turbulent viscosity ν, we can obtain the reconstructed turbulent viscosity ν. sfs , as shown below:

[0103]

[0104] Among them, C μ is the constant in the LRR-RSM model and is taken as 0.09.

[0105] Step 6: Use the reconstructed Reynolds stress tensor and turbulent viscosity to perform turbulence simulation:

[0106] In this step, the Reynolds stress tensor reconstructed as described in step 5 is used and turbulent viscosity ν sfs , the transport equation of the RSM series model is updated, replacing the Reynolds stress tensor in the original transport equation of the RSM series model and turbulent viscosity ν t , combined with the RSM series model, the Vreman dynamic grid adaptive turbulence simulation method based on turbulence energy spectrum coupling RSM series model is obtained. In this embodiment, taking the LRR-RSM turbulence model in the RSM series model as an example, the Reynolds stress tensor reconstructed in step 5 is used and turbulent viscosity ν sfs , the transport equation of the LRR-RSM model is updated, replacing the Reynolds stress tensor in the original transport equation of the LRR-RSM model and turbulent viscosity ν t , the new transport equation is shown as follows:

[0107]

[0108] in, is the pressure-strain related term, is the generated term of the Reynolds stress transport equation; is the generated term of the dissipation rate transport equation;

[0109]

[0110] Among them, the model coefficient C 1ε =1.44, model coefficient C 2ε =1.92;

[0111] The new transport equation is combined with the LRR-RSM model to obtain a grid adaptive turbulence simulation method based on the Vreman dynamic coefficient coupled RSM model, which is used for the numerical simulation of the back-step example.

[0112] The fully implicit coupling solution technology is used for transient calculation, and the time step meets the CFL condition in engineering computational fluid dynamics. In addition, in order to better demonstrate the advantages of the embodiments of the present invention, the LRR-RSM turbulence model is used to perform numerical simulation on the back step example, and the numerical simulation results are compared with those of the grid adaptive turbulence simulation method based on the Vreman dynamic coefficient coupled RSM model of the present invention.

[0113] Figure 3 Schematic diagram of the turbulent eddy structure of the back-step example calculated using the LRR-RSM model, colored by the turbulent viscosity ratio.

[0114] Figure 4 It is a schematic diagram of the turbulent eddy structure of a back-step example calculated by a grid adaptive turbulence simulation method based on the Vreman dynamic coefficient coupled RSM model of the present invention, and is colored by the turbulent viscosity ratio.

[0115] Figure 3 , Figure 4 Comparative analysis shows that the turbulent eddy structure analysis capability obtained by calculating the grid adaptive turbulence simulation method based on the Vreman dynamic coefficient coupled RSM model proposed in the present invention is significantly stronger than the result calculated by the LRR-RSM model. It can capture richer turbulent structures with the same number of grids and provide more accurate flow field details.

[0116] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

[0117] In summary, the grid-adaptive turbulence simulation method based on the Vreman dynamic coefficient coupled RSM model proposed in the present invention constructs the Vreman dynamic coefficient based on the local velocity gradient tensor, and then determines the local grid length scale Δ by identifying the local grid size. * , and then reconstruct the Reynolds stress tensor and turbulent viscosity by constructing scale-related functions through turbulent energy spectrum integration, realizing grid adaptive simulation, effectively overcoming the problem of high empirical dependence of existing turbulent mixing models on grids. By introducing the Vreman dynamic coefficient, the grid adaptability of the grid adaptive turbulence simulation method is further broadened, which can greatly reduce the computational cost while improving the calculation accuracy, significantly speed up the turbulence simulation process, and provide a new method for flow prediction in complex engineering fields.

Claims

1. A grid adaptive turbulence simulation method based on Vreman dynamic coefficient coupled RSM model, characterized in that: The steps include: Step 1, determine whether to apply the masking function; Step 2, calculate the Vreman dynamic coefficient; Step 3, determine the local dynamic grid length scale; Step 4: construct a scale-dependent adjustment function; Step 5, reconstruct the Reynolds stress tensor and turbulent viscosity of the RSM model; Step 6, use the reconstructed Reynolds stress tensor and turbulent viscosity to perform turbulence simulation; ① The determination of whether to apply the shielding function includes: Combined with the type of flow state being simulated, determine whether to use the shielding function F GAS Specifically, when the flow state type is free shear flow, the shielding function is not used. In this case, the shielding function F GAS = 0; when the flow state type is near-wall flow, a shielding function is used, and the shielding function F GAS Select one of the F1 shielding function and F2 shielding function from the DDES-SST model and the F from the DDES-SA model. d Masking function; ② The calculation of the Vreman dynamic coefficient includes: Based on the local velocity gradient tensor α ij , construct the Vreman dynamic coefficient C DGAS ; The Vreman dynamic coefficient C DGAS It is obtained by the following formula: β ij ＝α mi α mj Where S is the strain rate, C GAS The empirical coefficient is taken as 0.6; ③ The determination of the local dynamic grid length scale includes: Combined with the shielding function F described in step 1 GAS , and the Vreman dynamic coefficient C described in step 2 DGAS , determine the local dynamic grid length scale Δ * ; The local dynamic grid length scale Δ * Given by: D * =C DGAS [(1-F GAS )D vol +F GAS D max ] D max =max(Δ x ,D y ,D z ) Among them, Δ x is the length of the local hexahedral grid, Δ y is the width of the local hexahedral grid, Δ z is the height of the local hexahedral grid; ④ The structural scale-related adjustment functions include: According to the modeling method of turbulent kinetic energy in the RSM model, by solving the trace of the Reynolds stress tensor, the original modeled turbulent kinetic energy k is obtained. m ; According to the local dynamic grid length scale Δ * Based on the turbulence energy spectrum, the actual simulated turbulent kinetic energy k is obtained by integration u ; The actual simulated turbulent kinetic energy k u It is obtained by the following formula: Among them, C k is the Kolmogorov constant coefficient, which is taken as 1.5; ε is the actual turbulence dissipation rate; κ c is the solvable turbulence cutoff wave number, which is the dynamic grid length scale Δ obtained in step 3 * Determine as follows: Among them, π is the ratio of pi, which is 3.14; According to the actual simulated turbulent kinetic energy k u , the original modeled turbulent kinetic energy k m and the shielding function F described in step 1 GAS , construct the dynamic scale-dependent regulation function D f ; Define the scale ratio as the actual simulated turbulent kinetic energy k u and the original modeled turbulent kinetic energy k m The dynamic scale-dependent regulation function D f is a function related to the ratio of the scales, which is obtained by the following formula: the GAS =(1-F GAS )l u +F GAS the m Among them, l u is the grid-related scale, l m is the turbulence length scale given by the RSM model; ⑤ The Reynolds stress tensor and turbulent viscosity of the reconstructed RSM model include: Use the dynamic scale-dependent adjustment function D described in step 4 f The Reynolds stress tensor in the RSM model Regulate and obtain the reconstructed Reynolds stress tensor It is derived from the following formula: Adopt the dynamic scale-dependent adjustment function D described in step 4 f For the turbulent viscosity ν in the RSM model t By adjusting the turbulent viscosity ν, we can obtain the reconstructed turbulent viscosity ν. sfs , obtained from the following formula: n sfs =D f ·n t ⑥ The turbulence simulation using the reconstructed Reynolds stress tensor and turbulent viscosity includes: Using the reconstructed Reynolds stress tensor described in step 5 and turbulent viscosity ν sfs , the transport equation of the RSM model is updated, replacing the Reynolds stress tensor in the original transport equation of the RSM model and turbulent viscosity ν t , and combined with the RSM model to obtain the grid adaptive turbulence simulation method based on the Vreman dynamic coefficient coupled RSM model.

Images