An Improved Large Eddy Simulation Method for Vortex Generation

Abstract

This invention discloses an improved method for generating eddy currents in large eddy simulation (LES), belonging to the field of turbulence simulation technology. It solves the problems of insufficient eddy current randomness control, significant statistical errors in inlet turbulence, and high computational resource consumption in existing synthetic eddy current methods. This invention first generates randomly distributed synthetic eddy currents at the inlet based on the target turbulence characteristics. Then, it processes the velocity field using a dynamic normalization factor to meet statistical requirements. Finally, it determines the number of eddy currents based on a specific ratio and allows them to be arbitrarily distributed. This improved LES eddy current generation method breaks through traditional limitations, allowing for arbitrary eddy current placement, improving statistical accuracy, optimizing small-scale turbulence coverage, increasing computational efficiency by 1-2 orders of magnitude, reducing costs, and can be implemented using OpenFOAM software. It is applicable to various scenarios and provides strong support for multiple fields.

Description

Technical Field

[0001] This invention relates to the field of turbulence simulation technology, and in particular to an improved method for generating eddies in large eddy simulation. Background Technology

[0002] With the rapid development of computer science and technology, the application fields of computational fluid dynamics (CFD) are becoming increasingly widespread, and its role in the design and development of aero-engines is becoming more and more significant. Today, CFD software has become an important tool for evaluating and optimizing new design schemes. Through simulation numerical verification, engineers can simulate complex fluid behavior in a virtual environment, thereby significantly improving the accuracy and efficiency of the design. Among the many CFD numerical simulation methods, Large Eddy Simulation (LES) is highly favored due to its unique advantages. The LES method directly simulates large-scale eddies while modeling small-scale eddies, thus effectively reducing computational complexity and resource consumption while ensuring computational accuracy.

[0003] In the aerospace and automotive industries, engineers have shown great interest in Large Eddy Simulation (LES), primarily because LES can provide an unstable turbulent flow field, which is crucial for calculating aeroacoustic noise generated by vehicles or wings. However, in practice, LES calculations are typically applied only to specific regions of interest, such as the trailing edge of an airfoil or a car's rearview mirror, because these areas have a significant impact on noise and flow characteristics.

[0004] To achieve this goal, engineers typically employ a method called "embedded LES." In this method, the LES domain is embedded within a larger RANS (Reynolds-averaged Navier-Stokes) simulation domain. The RANS technique is used to simulate the entire geometry, providing stable flow conditions, while the LES focuses on specific high-resolution regions to capture unsteady turbulent effects.

[0005] Although the embedded LES approach is theoretically feasible, it still faces some challenges in practice:

[0006] 1. Transition from RANS to LES:

[0007] The flow conditions need to transition smoothly between the RANS and LES domains. RANS provides stable average flow conditions, while LES requires mature unsteady turbulent conditions. A key issue is how to generate a mature LES solution from a stable RANS solution within the shortest possible distance.

[0008] 2. Cost Calculation:

[0009] The computational cost of LES computation is much higher than that of RANS computation; therefore, limiting the size of the embedded LES domain is key to reducing the overall computational cost. How to minimize the range of the LES domain while ensuring it is large enough is a challenging problem that needs to be solved.

[0010] 3. Boundary conditions:

[0011] At the boundary between the RANS and LES domains, setting appropriate boundary conditions to ensure the effective transmission of flow information is also an important issue. Improperly set boundary conditions may lead to flow distortion or computational instability.

[0012] To address this issue, various methods for generating turbulence at the LES inlet have been proposed. Currently, the main methods used to generate inlet turbulence include the vortex method, the spectroscopic synthesizer, and the synthetic eddy method.

[0013] Among them, the eddy current generation method is a turbulence generation method based on the classic view that turbulence is a superposition of eddies. The underlying idea is to directly focus on defining the coherent structure, rather than directly reverting to spectral methods. This method is easy to implement, runs fast, and is applicable to any geometry and any type of flow. The generated data exhibits very good physical properties, but it also has some limitations in generating turbulent inlet conditions, such as the difficulty in precisely controlling the randomness and statistical characteristics of eddies. Furthermore, the eddies generated using this method are actually non-uniform, which will lead to some errors between the generated results and expectations. Summary of the Invention

[0014] This invention proposes an improved method for generating eddy currents in large eddy simulation. By introducing a dynamic normalization factor to optimize the statistical characteristics of eddy currents, allowing eddy currents to be generated at arbitrary spatial locations, and redefining the relationship between the number and density of eddy currents, this method solves the problems of insufficient control over the randomness of eddy currents, significant statistical errors in inlet turbulence, and high computational resource consumption in existing synthetic eddy current methods.

[0015] An improved method for generating eddy currents using large eddy simulation, the method comprising the following steps:

[0016] Step 1: Based on the statistical characteristics of the target turbulence, generate a synthetic vortex with a random spatial distribution at the entrance of the computational domain;

[0017] Step 2: Normalize the velocity field of the synthesized eddy current using a dynamic normalization factor to ensure that the synthesized velocity field satisfies the preset first-order and second-order statistics.

[0018] Step 3: Determine the number of active vortices based on the ratio of the inlet area to the average effective cross-sectional area of ​​the vortex, and allow the synthetic vortex to be dynamically distributed at any position in the inlet space to optimize small-scale turbulence coverage and computational efficiency.

[0019] Furthermore, in step two, the velocity field of the synthesized eddy current after normalization is:

[0020]

[0021] Statistical properties as well as Independent of the spatial distribution and length scale of the vortex.

[0022] Furthermore, in step two, the dynamic normalization factor is determined by taking the average value of the eddy current concentration, so that the synthesized velocity field satisfies the unit variance condition and accurately recovers the target first and second order statistics.

[0023] Furthermore, in step three, the formula for calculating the number of eddies is:

[0024]

[0025] Where A inlet For the entrance area, A eddy Let be the average effective cross-sectional area of ​​the vortex.

[0026] Furthermore, the improved large eddy simulation eddy current generation method is implemented based on OpenFOAM software, and a TurbulentInletSEM inlet condition library is established for simulation calculation.

[0027] A storage medium storing a computer program, characterized in that the computer program, when executed by a processor, implements the above-described improved large eddy simulation eddy current generation method.

[0028] A computer device, characterized in that it comprises: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the above-described improved large eddy simulation eddy current generation method.

[0029] The beneficial effects of this invention are:

[0030] 1. Allows vortices to be placed arbitrarily in the inlet space, overcoming the limitation of traditional SEM that can only generate vortices in a three-dimensional Cartesian box, and can ensure that all vortices are active at all time steps.

[0031] 2. The original SEM suffers from non-uniformity in its eddy length scale setting, leading to inaccurate recovery of preset turbulence statistics. This invention introduces a new normalization procedure to ensure accurate recovery of the target statistical data. This improvement significantly enhances the statistical accuracy of the simulation, making the simulation results closer to actual physical phenomena, thus providing more reliable data support for engineering design and scientific research.

[0032] 3. Traditional SEM requires significant computational resources to generate turbulent boundary conditions, especially under high Reynolds numbers and complex geometries. This invention reduces the required number of vortices by generalizing the vortex placement strategy, allowing arbitrary vortex placement while maintaining the target "vortex density." This improvement results in approximately a 1-2 order of magnitude increase in computational efficiency, enabling large-scale LES simulations to be completed in a shorter time, reducing computational costs and improving research efficiency.

[0033] 4. The improvements of this invention are not only applicable to the generation of turbulent inlet boundary conditions in LES, but can also be extended to other SEM-based methods, such as volumetric source terms and dynamic forcing algorithms. This means that the advantages of this invention can be applied to a wider range of CFD applications, including but not limited to aerospace, automotive engineering, energy, and environmental science. Attached Figure Description

[0034] Figure 1 Generate an eddy current distribution map for the initial SEM;

[0035] Figure 2 Generate eddy current distribution maps for improved SEM. Detailed Implementation

[0036] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0037] An improved method for generating eddy currents using large eddy simulation, the method comprising the following steps:

[0038] Step 1: Based on the statistical characteristics of the target turbulence, generate a synthetic vortex with a random spatial distribution at the entrance of the computational domain;

[0039] Step 2: Normalize the velocity field of the synthesized eddy current using a dynamic normalization factor to ensure that the synthesized velocity field satisfies the preset first-order and second-order statistics.

[0040] Step 3: Determine the number of active vortices based on the ratio of the inlet area to the average effective cross-sectional area of ​​the vortex, and allow the synthetic vortex to be dynamically distributed at any position in the inlet space to optimize small-scale turbulence coverage and computational efficiency.

[0041] Specifically, the improved large eddy simulation (SEM) eddy current generation method described in this invention generates synthetic eddies with random spatial distribution based on the statistical characteristics of target turbulence at the computational domain inlet, overcoming the limitations of traditional methods. Traditional SEM can only generate eddies within a three-dimensional Cartesian box, while this invention allows eddies to be placed arbitrarily in the inlet space, making the eddy current distribution more flexible and enabling a more comprehensive simulation of actual turbulence. This improvement ensures that all eddies are active at each time step, greatly enhancing the realism and effectiveness of the simulation. The key technological innovation of this invention is the normalization of the velocity field after generating the synthetic eddies using a dynamic normalization factor. The original SEM suffers from non-uniformity in setting the eddy length scale, making it difficult to accurately recover the preset turbulence statistics. The new normalization procedure introduced in this invention determines the dynamic normalization factor by taking the running average of the eddy current concentration, ensuring that the synthetic velocity field satisfies the preset first and second-order statistics and the unit variance condition, effectively solving this problem. This improvement significantly enhances the statistical accuracy of the simulation, allowing the simulation results to more accurately reflect actual physical phenomena and providing more reliable data support for engineering design and scientific research. Determining the number of active vortices based on the ratio of the inlet area to the average effective cross-sectional area of ​​the vortex, and allowing the synthetic vortex to be dynamically distributed at any location in the inlet space, is the core method by which this invention improves computational efficiency. The original SEM requires significant computational resources to generate turbulent boundary conditions, especially under high Reynolds numbers and complex geometries. This invention, by optimizing the vortex placement strategy, reduces the number of required vortices while maintaining the target "vortex density," resulting in superior performance in small-scale turbulence coverage and achieving a computational efficiency improvement of approximately 1-2 orders of magnitude. This enables large-scale LES simulations to be completed in a shorter time, significantly reducing computational costs and improving research efficiency. Furthermore, the improvements of this invention have broad applicability. It is not only applicable to the generation of turbulent inlet boundary conditions in LES but can also be extended to other SEM-based methods, such as volumetric source terms and dynamic forcing algorithms. This means that its advantages can be applied to many CFD application scenarios such as aerospace, automotive engineering, energy and environmental science, greatly expanding the application scope of the technology and bringing new technical support and development opportunities to related research and engineering practices in different fields.

[0042] Furthermore, in step two, the velocity field of the synthesized eddy current after normalization is:

[0043]

[0044] Statistical properties as well as Independent of the spatial distribution and length scale of the vortex.

[0045] Specifically, a random perturbation function is introduced to adjust the initial parameters of the eddies when generating the synthetic eddies. This adjustment makes the generated synthetic eddies more closely resemble the complex and variable characteristics of actual turbulence. In real turbulent phenomena, the initial state of eddies is influenced by various uncertainties, exhibiting a high degree of randomness. Traditional methods generate relatively fixed initial parameters for eddies, making it difficult to accurately simulate this reality. This invention, through a random perturbation function, allows the initial parameters of each synthetic eddy to vary randomly within a certain range, greatly enhancing the randomness and diversity of the synthetic eddies. This results in generated turbulence that more closely resembles actual conditions, improving the realism and reliability of the simulation. This random adjustment of the initial eddy parameters helps improve the uniformity and stability of the synthetic velocity field. In large eddy simulations, the uniformity and stability of the velocity field are crucial to the accuracy of the simulation results. The introduction of the random perturbation function avoids local anomalies in the velocity field caused by overly regular initial eddy parameters, resulting in a more uniform distribution and less fluctuation in the velocity field across the entire computational domain. This not only improves the accuracy of the simulation results but also reduces computational errors and convergence problems that may be caused by unstable velocity fields, ensuring the smooth progress of the simulation process. Because the generated synthetic eddies more closely resemble actual turbulence and the synthetic velocity field is more uniform and stable, this invention can more accurately capture the characteristics of small-scale turbulence in subsequent calculations. Small-scale turbulence plays a crucial role in many practical engineering problems, and accurate simulation of small-scale turbulence is essential for understanding and solving these problems. This invention improves the resolution and simulation capability of small-scale turbulence by enhancing the randomness of the synthetic eddies and optimizing the velocity field, providing more valuable reference data for related engineering fields. This improvement also helps reduce computational costs. More accurate simulation means that, while achieving the same accuracy requirements, it is not necessary to perform excessive computational iterations or use excessively high computational grid densities as in traditional methods. By optimizing the generation of synthetic eddies and the characteristics of the velocity field, unnecessary computational load is reduced, thereby effectively improving computational efficiency and reducing computational resource consumption while ensuring simulation quality.

[0046] Furthermore, in step two, the dynamic normalization factor is determined by taking the average value of the eddy current concentration, so that the synthesized velocity field satisfies the unit variance condition and accurately recovers the target first and second order statistics.

[0047] Specifically, in this invention, the dynamic normalization factor is determined by taking the operating average of the eddy concentration. This unique approach effectively solves the problem of the difficulty in accurately controlling the statistical characteristics of eddies when generating turbulent inlet conditions using traditional methods. Traditional eddy generation methods, due to the lack of a scientific setting of the normalization factor, result in deviations in the randomness and statistical characteristics of the generated eddies, making it impossible to accurately recover the preset turbulence statistics. This invention, however, determines the dynamic normalization factor by accurately calculating the operating average of the eddy concentration, ensuring that the synthesized velocity field satisfies both the preset first- and second-order statistics and the unit variance condition. This means that the simulation results can more accurately reflect the statistical laws of actual turbulence, greatly improving the statistical accuracy of the simulation. In practical engineering applications, such as aero-engine design and automotive aerodynamics research, more accurate turbulence simulation data can provide engineers with more reliable evidence, helping them to understand fluid flow characteristics more deeply, thereby optimizing product design and improving product performance. This method of determining the dynamic normalization factor makes the statistical properties of the synthesized eddy velocity field independent of the spatial distribution and length scale of the eddies. This characteristic is significant because it overcomes the interference of the spatial distribution and length scale of vortices on statistical properties in traditional methods. Regardless of the distribution of vortices at the computational domain inlet or the variation of their length scale, it guarantees stable and accurate statistical results. In complex engineering scenarios, the geometry and boundary conditions of the flow field vary greatly, and traditional methods often struggle to guarantee the accuracy and stability of simulations in these complex situations. This feature of the present invention makes the simulation results unaffected by these factors, exhibiting greater adaptability and versatility, and enabling the stable output of reliable simulation data in different application scenarios. By accurately recovering the first and second-order statistics of the target, this invention provides a solid data foundation for subsequent numerical simulations and analyses. In large eddy simulations, these statistics are key parameters for studying turbulence characteristics and predicting flow behavior. Accurate statistical data helps researchers delve deeper into the physical mechanisms of turbulence, such as energy transfer and vortex interactions. Furthermore, in engineering design, optimization based on this accurate statistical data can effectively improve product performance and reliability. For example, in the aerospace field, accurate turbulence simulation can help design more efficient wings and engine inlets, reducing drag and improving fuel efficiency. In the energy field, accurate turbulence simulation helps improve wind energy capture efficiency and reduce costs for wind turbine blade design. Furthermore, this method of determining the dynamic normalization factor indirectly improves computational efficiency. Because the statistical properties of the synthesized velocity field are more stable and accurate, a large number of iterative calculations to correct statistical errors are not required during the simulation, reducing the waste of computational resources. At the same time, the stable statistical properties also help accelerate the convergence speed of the calculation, allowing the entire simulation process to be completed in a shorter time.

[0048] Furthermore, in step three, the formula for calculating the number of eddies is:

[0049]

[0050] Where A inlet For the entrance area, A eddy Let be the average effective cross-sectional area of ​​the vortex.

[0051] Specifically, the number of vortices is determined based on the ratio of the inlet area to the average effective cross-sectional area of ​​the vortex, providing a scientific and reasonable basis for the simulation process. Compared with traditional methods, this calculation method fully considers the three-dimensional characteristics of vortices, making the determined number of vortices more closely match the actual flow field conditions. In actual fluid flow, vortices do not exist in isolation; their distribution is closely related to the geometry of the flow field and boundary conditions. Traditional methods for setting the number of vortices are often relatively simple and do not fully consider these complex factors, easily leading to significant deviations between simulation results and actual conditions. The calculation formula of this invention comprehensively considers the inlet area and the effective cross-sectional area of ​​the vortex itself, and can more accurately reflect the number of vortices required under different operating conditions, thereby optimizing the coverage effect of small-scale turbulence. For example, in the complex flow field environment inside an aero-engine, accurate setting of the number of vortices can more accurately simulate small-scale turbulence phenomena, providing strong support for the optimization of engine performance. This method of determining the number of vortices significantly improves computational efficiency while ensuring simulation accuracy. The precisely calculated number of eddies ensures sufficient coverage of small-scale turbulent regions, accurately capturing key information in the flow field, while avoiding the waste of computational resources caused by generating too many unnecessary eddies. In cases of high Reynolds numbers and complex geometries, traditional methods require significant computational resources to generate turbulent boundary conditions, resulting in low computational efficiency. This invention, by rationally determining the number of eddies, reduces the required number of eddies while maintaining the target "eddy density," enabling large-scale LES simulations to be completed in a shorter time. This not only reduces computational costs but also improves research efficiency, allowing engineers and researchers to evaluate and optimize multiple design schemes in a shorter time, accelerating product development. In practical applications, this calculation formula enhances the reliability and repeatability of simulation results. Because it is based on explicit physical parameters and mathematical relationships, different researchers using the method of this invention can obtain consistent eddy number settings as long as the same inlet area and average effective cross-sectional area of ​​the eddies are given, thus ensuring the consistency and comparability of simulation results.

[0052] Furthermore, the improved large eddy simulation eddy current generation method is implemented based on OpenFOAM software, and a TurbulentInletSEM inlet condition library is established for simulation calculation.

[0053] Specifically, OpenFOAM is a widely used and open-source computational fluid dynamics software platform with a rich algorithm library and powerful computing capabilities, providing a solid foundation for the method of this invention. Development based on OpenFOAM software can fully utilize its existing mature technologies and extensive user community resources. This means that the method of this invention can leverage the advantages of OpenFOAM to quickly build a simulation environment, reducing the significant time and effort required for development from scratch. Simultaneously, OpenFOAM's open-source nature allows researchers to gain in-depth understanding of the code details, flexibly adjust and optimize the algorithm according to specific needs, further improving the accuracy and adaptability of the simulation to meet the special requirements of large eddy simulation in different engineering scenarios. A TurbulentInletSEM ingress condition library is established, providing a convenient and efficient way to call the simulation calculations. In large eddy simulation, the accurate setting of ingress conditions is crucial to the simulation results.

[0054] The TurbulentInletSEM inlet condition library encapsulates the improved eddy current generation method of this invention. Users can easily apply this method to generate the required turbulent inlet conditions in simulation calculations by simply calling it, avoiding tedious repetitive programming work. This not only greatly improves the work efficiency of researchers but also lowers the barrier to entry, enabling more engineers without strong programming skills to easily use the technology of this invention for simulation analysis. Furthermore, the establishment of this inlet condition library allows the method of this invention to better collaborate with other functional modules of OpenFOAM software, achieving integrated and automated simulation processes, further improving the smoothness and stability of simulation calculations. This implementation method based on specific software and an inlet condition library is conducive to the promotion and application of the technology of this invention. In the engineering field, many research teams and enterprises are already using OpenFOAM software for related research and design work. This invention, based on OpenFOAM, allows these users to easily integrate the new eddy current generation method into their existing workflows without having to relearn and adapt to a new software platform. At the same time, the sharing and dissemination of the inlet condition library is also easier, allowing researchers to quickly reuse the library in different projects, accelerating the diffusion of the technology.

[0055] A storage medium storing a computer program that, when executed by a processor, implements the above-described improved large eddy simulation eddy current generation method.

[0056] Specifically, the storage medium, as a carrier of the computer program, solidifies and preserves the technology, greatly enhancing the inheritability of this invention. Traditional technology dissemination relies heavily on written descriptions and oral instruction, during which technical details are easily lost or misunderstood. However, by writing the method of this invention into a computer program and storing it on the storage medium, the complex technical steps and algorithmic logic, from eddy current generation to normalization processing, are fully preserved. Whether current researchers review the research process or new researchers join the project, they can accurately obtain the core technical content of this invention by reading the program on the storage medium, avoiding information attenuation during technology transfer and ensuring the long-term maintenance of the technology's integrity. In terms of application expansion, the existence of the storage medium allows the method of this invention to easily transcend different computing devices and research environments. As long as there is a corresponding processor and operating environment, the computer program on the storage medium can be executed on various computer devices. This breaks down the hardware and environmental limitations of technology application, enabling convenient application of the technology, whether for large-scale complex simulations on high-performance computing clusters in research institutions or for rapid verification of small projects on individual engineers' workstations. Meanwhile, for various industry sectors, such as aerospace, automotive engineering, and energy, as long as work related to large eddy simulation (LES) is involved, the method can be accessed and applied through storage media, broadening the scope of application and promoting cross-disciplinary technical exchange and cooperation. From the perspective of R&D process optimization, the computer programs on the storage media facilitate the reuse and improvement of the technology. In actual R&D work, simulation methods often require multiple trials and optimizations. Storage media allows researchers to quickly access previously written programs, modify and improve them without having to rewrite the code each time, saving significant manpower and time costs. Furthermore, in collaborative R&D teams, storage media, as a shared resource, facilitates the sharing of code and research results among team members, promoting communication and collaboration, improving R&D efficiency, accelerating project progress, and driving the continuous development and innovation of LES technology.

[0057] A computer device includes: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the improved large eddy simulation eddy current generation method described above.

[0058] Specifically, the computer device of this invention provides a stable and efficient operating environment for the improved large eddy simulation eddy current generation method. The processor, as the core computing unit, possesses powerful data processing capabilities, enabling it to quickly process large amounts of data related to eddy current generation, velocity field normalization, and determining the number of eddies based on the inlet area and the average effective cross-sectional area of ​​the eddies. When performing the step of generating synthetic eddies based on the target turbulence statistical characteristics, the processor can perform high-speed calculations to achieve accurate simulation of random spatial distributions. In the process of normalizing the velocity field of the synthetic eddy current to satisfy preset first- and second-order statistics, the processor can quickly handle complex mathematical operations, ensuring the accuracy and timeliness of the calculation results. Its high-efficiency computing power solves the problems of slow calculation and low efficiency that may occur in traditional computing devices when handling such complex simulation tasks, allowing the entire simulation process to be completed in a short time, meeting the needs of practical engineering applications and scientific research for rapid acquisition of simulation results. The memory plays an indispensable role in this computer device, providing stable storage and retrieval support for various types of data during the simulation process. During simulation operation, it stores a large amount of information, including the initial parameters required to generate synthetic eddies, intermediate data during the calculation process, and the final simulation results. These data form the foundation for successful simulation. The stable storage performance of the memory ensures data integrity and accuracy, preventing data loss or errors from affecting the simulation results. Simultaneously, the rapid data read capability allows the processor to acquire the necessary data promptly, further enhancing the smoothness and efficiency of the entire simulation process and guaranteeing the efficient operation of the series of operations from eddy current generation to the final simulation result output. Integrating the simulation method program of this invention into this computer device achieves deep hardware and software fusion. This fusion significantly lowers the technical barrier to entry. For engineers and researchers, there is no need to spend considerable effort building complex computing environments and configuring software parameters; they can easily utilize the variable-dimensional simulation method of this invention simply by operating the computer device. This not only improves R&D efficiency but also enables more people to participate in related research and engineering design work, promoting the widespread application and further innovation of the technology. In aerospace, automotive engineering, and other fields, engineers can use this computer device to quickly simulate turbulence under different operating conditions, providing strong support for product design, shortening product development cycles, reducing R&D costs, driving continuous technological progress in related industries, and enhancing the competitiveness of the entire industry.

[0059] The following is a specific embodiment of the present invention:

[0060] This invention provides an improved vortex generation method (SEM) for generating inlet turbulence in Large Eddy Simulation (LES). This method improves the control accuracy of the randomness and statistical properties of vortices by introducing a novel vortex generation algorithm.

[0061] The original SEM method generates vortices by defining a fixed number of tightly supported synthetic vortices, which are generated in a virtual Cartesian box at the inlet. The velocity field formed by these vortices is formulated as follows:

[0062]

[0063] However, this method can lead to non-uniformity in the eddies, which will introduce errors when reconstructing the Reynolds stress, and such a velocity field cannot satisfy the unit variance condition of the initial velocity field.

[0064] The improved SEM method of this invention proposes a universal normalization factor, which can be found by taking the running average of the eddy current concentration. The improved eddy current velocity field becomes:

[0065]

[0066] This modification makes it easier to obtain the relevant statistical properties of eddies. Furthermore, compared to the original SEM, the statistical properties are... as well as The calculation results are more accurate when the spatial distribution and length scale of the vortex are independent.

[0067] For the problem of eddy current localization, the traditional SEM method recommends setting the number of eddies N as follows:

[0068]

[0069] Where V eddy This is the smallest volume among all vortices. This method is relatively simple when defining dimensions, but in many practical cases, the number of vortices is large, and using this definition method will lead to a significant increase in computational cost.

[0070] The improved SEM method also optimizes the number of eddies, and its formula is as follows:

[0071] Where A inlet For the entrance area, A eddy Let be the average effective cross-sectional area of ​​the vortex. This definition takes into account the three-dimensional characteristics of the vortex, allows all vortices to be in an active state, and provides better coverage in small-scale regions.

[0072] Figure 1 and Figure 2 The improved SEM method generates a distribution map of a fixed number of randomly placed eddies, compared to the original SEM method.

[0073] By comparison, it can be found that the original SEM method for eddy localization results in insufficient coverage of areas with small length scales, while the improved SEM generates eddies that are all concentrated near the dashed line. This indicates that all eddies are active and have better coverage in areas with small length scales, which will also reduce computational costs in actual calculations.

[0074] To verify the effectiveness of this invention in numerical simulation, the improved SEM method was written into code, compiled based on OpenFOAM software, and an OpenFOAM library for generating inlet turbulence conditions was established. A new inlet condition called TurbulentInletSEM was defined for use in simulation calculations.

[0075] The results of testing the new method show that, regardless of the vortex distribution and homogeneity, the new method can recover the first and second-order statistics of the target at the inlet and can place vortices at any location in space, which greatly increases the possibility of improving the efficiency of the SEM method.

[0076] This invention generates randomly distributed synthetic vortices at the computational domain inlet based on the target turbulence statistical characteristics during the vortex generation stage. This breaks through traditional limitations, allowing vortices to be placed at any location, ensuring active vortices at all time steps and highly replicating the actual turbulence state. A dynamic normalization factor is used to normalize the synthetic vortex velocity field. This factor is determined by the average value of the vortex concentration during operation. This not only ensures the synthetic velocity field meets the preset first and second-order statistics and unit variance conditions, but also frees the statistical properties from the influence of the spatial distribution and length scale of the vortices, significantly improving the statistical accuracy of the simulation and providing reliable data for engineering design and scientific research. The number of active vortices is determined based on the ratio of the inlet area to the average effective cross-sectional area of ​​the vortex, optimizing small-scale turbulence coverage. While maintaining vortex density, the number of vortices is reduced, resulting in a computational efficiency improvement of approximately 1-2 orders of magnitude, reducing computational costs, and making it suitable for large-scale LES simulations. Furthermore, this method is implemented based on OpenFOAM software and a TurbulentInletSEM inlet condition library is established. Leveraging the advantages of software reduces development costs, facilitates easy use, and promotes technology dissemination. The combination of storage media and computer equipment enables the solidification and inheritance of technology and cross-environmental applications. The deep integration of hardware and software lowers the threshold for use, promotes technological innovation and industry development, and has a wide range of applications, covering multiple fields such as aerospace and automotive engineering.

[0077] The specific embodiments of the invention have been described in detail above, but they are only examples, and the invention is not limited to the specific embodiments described above. For those skilled in the art, any equivalent modifications and substitutions to the invention are also within the scope of this invention. Therefore, all equivalent changes and modifications made without departing from the spirit and scope of this invention should be covered within the scope of this invention.

Claims

1. An improved method for generating eddy currents using large eddy simulation, characterized in that, The method includes the following steps: Step 1: Based on the statistical characteristics of the target turbulence, generate a synthetic vortex with a random spatial distribution at the entrance of the computational domain; Step 2: Normalize the velocity field of the synthesized eddy current using a dynamic normalization factor to ensure that the synthesized velocity field satisfies preset first-order and second-order statistics. In step two, the velocity field of the synthesized eddy current after normalization is: Statistical properties as well as Independent of the spatial distribution and length scale of the vortex; In step two, the dynamic normalization factor is determined by taking the average value of the eddy current concentration, so that the synthesized velocity field satisfies the unit variance condition and accurately recovers the target first and second order statistics. Step 3: Determine the number of active vortices based on the ratio of the inlet area to the average effective cross-sectional area of ​​the vortex, and allow the synthetic vortex to be dynamically distributed at any position in the inlet space to optimize small-scale turbulence coverage and computational efficiency.

2. The improved large eddy simulation eddy current generation method according to claim 1, characterized in that, In step three, the formula for calculating the number of eddies is: in For the entrance area, Let be the average effective cross-sectional area of ​​the vortex.

3. The improved large eddy simulation eddy current generation method according to claim 2, characterized in that, The improved large eddy simulation eddy current generation method is implemented based on OpenFOAM software, and a TurbulentInletSEM inlet condition library is established for simulation calculation.

4. A storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the improved large eddy simulation eddy current generation method according to any one of claims 1-3.

5. A computer device, characterized in that, include: A memory, a processor, and a computer program stored in the memory and executable on the processor, the processor executing the program to implement the improved large eddy simulation eddy current generation method according to any one of claims 1-3.

Images