Flow channel design optimization method based on feature driving

Abstract

The invention specifically relates to a runner design optimization method based on feature driving, and relates to the technical field of fluid mechanics and structure optimization, and the method comprises the following steps: S1, determining information; s2, parametric modeling is carried out on the physical domain; s3, obtaining a corresponding pseudo-density field; s4, control equation solving, state variable solving and sensitivity analysis are carried out; s5, updating the design variables; and S6, outputting an optimization result and carrying out post-processing. According to the method, different from a common density method, no obvious grid dependence exists, density projection and filtering do not need to be carried out, and the optimization process can be simplified; a clear optimization result boundary is obtained through the boundary of the geometric features, so that result post-processing and CAD reconstruction are facilitated, and subsequent processing and manufacturing are facilitated; the design variables are geometric parameters of the features and do not depend on the pseudo density of the grid units, the number of the design variables can be remarkably reduced, and efficient calculation in the optimization process is facilitated.

Description

Technical Field

[0001] This invention relates to the fields of fluid mechanics and structural optimization technology, and in particular to a feature-driven flow channel design optimization method. Background Technology

[0002] With the continuous development of modern industry, industrial products are demanding increasingly higher overall performance. Flow channels, as a fundamental design element in product structure, exist in numerous pieces of equipment. As technology advances and design requirements become more stringent, structural temperatures and heat flux densities are rising, placing new and higher demands on flow channel design. The quality of flow channel design directly impacts the product's operational stability, reliability, and lifespan. Traditional structural flow channel design relies heavily on trial and error, often resulting in regular, straight channels, leading to low design efficiency, high flow resistance, and poor cooling effects. Currently, structural optimization techniques (primarily topology optimization) can achieve automatic flow channel design and generation through computer simulation and optimization methods. Most of these methods employ density-based topology optimization, but they have the following limitations: Density methods are prone to problems such as checkerboard patterns and grid dependency, requiring additional projection filtering and other methods to facilitate the iterative process. The optimization results of the density method are prone to producing jagged boundaries, which makes the model reconstruction process complicated and cannot be directly converted into a machinable CAD model.

[0003] The design variables of the density method are pseudo-density variables on each grid cell. With many design variables, the computational cost during the optimization process is relatively high.

[0004] This invention significantly reduces design variables through parametric modeling of geometric features, resulting in clear structural boundaries and effectively solving the aforementioned problems. Summary of the Invention

[0005] The purpose of this invention is to propose a feature-driven flow channel design optimization method to solve the above-mentioned problems.

[0006] To achieve the above objectives, the present invention adopts the following technical solution: Feature-driven flow channel design optimization methods include: Step S1: Determine the physical domain, physical model, and optimization model information of the structure; Step S2: Select appropriate implicit features and perform parametric modeling of the physical domain; Step S3: Obtain the corresponding pseudo-density field through the level set function mapping of the physical domain, and interpolate the physical property parameters; Step S4: Solve the governing equations, solve the state variables, and perform sensitivity analysis; Step S5: Update the design variables using the optimization algorithm and determine whether the convergence requirement is met. If it is met, proceed to the next step; otherwise, return to step S3 to continue the optimization iteration until the convergence condition is met. Step S6: Output the optimization results and perform post-processing to obtain a flow channel optimization design scheme that meets the requirements.

[0007] Preferably, the physical domain in step S1 includes the design domain and the non-design domain, and the boundary conditions include flow-related inlet, outlet and wall conditions as well as heat-related heat source and adiabatic boundary conditions.

[0008] Preferably, the optimization model adopts the following mathematical expression: ; Minimize: ; st , ;

[0009] in, Represents the set of design variables. and Indicates the upper and lower limits of design variables. Describe the objective function. It represents all inequality constraints; the governing equations represent all flow and heat-related physical equations.

[0010] Preferably, the design variables are the geometric parameters of all features.

[0011] Preferably, the implicit features include, but are not limited to, circles, hyperellipses, closed B-splines, smooth deformable implicit curves, and B-spline bias features; the design variables are the set of geometric parameters of these features.

[0012] Preferably, the method of parametric modeling of the physical domain refers to establishing a level set function for the entire physical domain, and determining whether a point belongs to a fluid domain, a fluid boundary, or a solid domain by using the function value of any internal point.

[0013] Preferably, the parametric modeling uses function operations similar to Boolean operations to implement the "union" and "intersection" operations, including but not limited to max / min, R function, p norm, and KS function.

[0014] Preferably, the level set function of the physical domain forms a pseudo-density field corresponding to each grid cell through a mapping form; This is achieved through different forms of Heaviside functions.

[0015] Preferably, the algorithm used for sensitivity analysis in step S3 includes, but is not limited to, the adjoint method and the finite difference method.

[0016] Preferably, in step S6, when step S5 meets the convergence requirement, the optimization process will end the iteration. The output of this step includes, but is not limited to: fluid velocity field, temperature field, density field, and zero-level set curve.

[0017] In summary, due to the adoption of the above technical solution, the beneficial effects of the present invention are: 1. Unlike commonly used density methods, this invention does not have significant grid dependence and does not require density projection and filtering, which simplifies the optimization process. It obtains clear optimization result boundaries through the boundaries of geometric features, which facilitates post-processing of results and CAD reconstruction, and is beneficial for subsequent processing and manufacturing. The design variables are the geometric parameters of the features, rather than pseudo-density dependent on grid cells, which can significantly reduce the number of design variables and facilitate efficient calculation in the optimization process.

[0018] 2. This invention defines the boundary of the optimization result through the mathematical expression of implicit features (such as closed B-splines), avoiding problems such as sawtooth boundaries and checkerboard patterns in the traditional density method. The boundary of the optimization result is smooth and clear, and can be directly imported into CAD software to generate a solid model without the need for a complex boundary reconstruction process, thus improving the manufacturability of the project. After optimization, the two-dimensional heat sink obtains a complex configuration similar to bionics, which can reasonably distribute the flow and effectively reduce the temperature. Compared with the straight channel design of the traditional empirical method, the heat dissipation efficiency is significantly improved, and this design concept can be extended to three-dimensional heat sinks, manifold structures and other fields. Attached Figure Description

[0019] Further details, features, and advantages of this application are disclosed in the following description of exemplary embodiments in conjunction with the accompanying drawings, in which: Figure 1 This is a system structure diagram of the present invention; Figure 2 This is an explanation of the physical problems of the present invention; Figure 3 This is the initial feature layout of the present invention; Figure 4 This is an optimized feature layout for the present invention;

[0020] Figure 5 This is the final optimized configuration of the present invention. Detailed Implementation

[0021] Several embodiments of this application will now be described in more detail with reference to the accompanying drawings to enable those skilled in the art to implement this application. This application may be embodied in many different forms and for various purposes and should not be limited to the embodiments set forth herein. These embodiments are provided to make this application thorough and complete, and to fully convey the scope of this application to those skilled in the art. The embodiments described do not limit this application.

[0022] Unless otherwise defined, all terms used herein (including technical and scientific terms) shall have the same meaning as commonly understood by one of ordinary skill in the art to which this application pertains. It will be further understood that terms such as those defined in commonly used dictionaries shall be interpreted as having a meaning consistent with their meaning in the relevant field and / or the context of this specification, and shall not be interpreted in an idealized or overly formal sense unless expressly defined herein.

[0023] Example 1

[0024] Its specific implementation method is combined with the appendix Figure 1 To the attached Figure 5 Please provide a detailed explanation.

[0025] Appendix Figure 1 The flowchart of the feature-driven flow channel design optimization method provided in the embodiments of the present invention shows the complete steps from determining the physical domain, physical model, and optimization model information of the structure to outputting the optimization results and performing post-processing.

[0026] In this embodiment, it includes: Step S1: Determine the physical domain, physical model, optimization model, and other information of the structure; This step focuses on the optimization design of fluid-thermal coupling structures, including both shape optimization and topology optimization. It involves optimizing the distribution of fluid and solid materials within the physical domain to obtain a structural flow channel layout with optimized performance. like Figure 2 The physical domains shown include the design domain and the non-design domain. The boundary conditions include flow-related conditions such as inlet, outlet, and wall conditions, as well as heat-related conditions such as heat sources and adiabatic boundaries.

[0027] The optimization model uses the following mathematical expression; Find: ; Minimize: ; st , ; in, Represents the set of design variables. and Indicates the upper and lower limits of design variables. Describe the objective function. This represents all inequality constraints. The governing equations represent all flow and thermal-related physical equations; the design variables here are the geometric parameters of all characteristics.

[0028] objective function The energy dissipation of the flow field is taken into account. Thermal compliance with temperature field Normalized linear combination form:

[0029] Inequality constraints The constraint is based on the percentage of fluid volume, as follows: ; in, It is the physical domain level set function generated by feature modeling. The pseudo-density field obtained by mapping, It is the upper limit of the body fraction ratio; Clearly define the design domain and non-design domain to avoid ineffective optimizations (e.g., structures in non-design domains do not require adjustment) and improve optimization efficiency; for example... Figure 2 By dividing the design domain, the focus is on optimizing the flow channel layout in the heat source area.

[0030] Multiphysics coupling modeling: Simultaneously considers the flow boundary (inlet / outlet / wall) and the thermal boundary (heat source / insulation) to ensure that the optimization results meet the requirements of the flow-heat coupling condition and avoid performance imbalance caused by single physics optimization.

[0031] Standardization of mathematical models: A unified optimization mathematical expression (objective function combined with constraints) is adopted, and design variables (characteristic geometric parameters), objectives (energy dissipation and thermal compliance), and constraints (fluid volume ratio) are incorporated into the standardized framework to facilitate algorithm implementation and cross-scenario application; Step S2: Select appropriate implicit features and perform parametric modeling of the physical domain; Implicit features include, but are not limited to, mathematical expressions for circles, superellipses, closed B-splines (CBS), smoothly deformable implicit curves (SDIC), and B-spline offset features (BSOF); while design variables... That is, the set of geometric parameters of these features.

[0032] Parametric modeling of the physical domain refers to establishing the level set function of the entire physical domain. Determine whether a point belongs to the fluid domain by using the function value of any interior point. Fluid boundary Still a solid domain : ; Parametric modeling can use function operations similar to Boolean operations to implement "union" and "intersection" operations, including but not limited to mathematical expressions such as max / min, R function, p-norm, and KS function.

[0033] like Figure 3 As shown, parametric modeling of a simplified model (half) of the physical domain is performed using closed B-splines; Using geometric characteristic parameters (such as the radius of a circle and the coordinates of control points of a B-spline) as design variables to replace the pseudo-density variables of mesh elements in the density method can reduce the number of variables by more than 90% (for example, only a few dozen parameters are needed to describe complex flow channels in two-dimensional cases), significantly reducing the amount of computation. Implicit features (such as closed B-splines) can accurately describe curve boundaries through mathematical expressions, avoiding the jagged boundaries of the density method. The optimization results can be directly used for CAD modeling, improving engineering manufacturability (e.g., Figures 3 to 4 The evolution of characteristics reflects boundary smoothness). By changing characteristic parameters (such as scaling, translation, and deformation), both flow channel shape optimization (such as adjusting the bending angle) and topology optimization (such as adding or removing flow channel branches) can be achieved without distinguishing between the two optimization types, thus simplifying the design process. Step S3: Obtain the corresponding pseudo-density field through the level set function mapping of the physical domain, and interpolate the physical property parameters; In this step, the level set function of the physical domain A pseudo-density field corresponding to each grid cell is formed through mapping. This method can be implemented using different forms of the Heaviside function. The Heaviside function used in this embodiment is: ; The aforementioned property interpolation models include, but are not limited to, SIMP, RAMP, etc. By mapping the level set function to a pseudo density field through the Heaviside function, the geometric feature model can be recognized by the finite element solver (such as COMSOL, which can directly read the density field and generate a mesh), thus bridging the gap between geometric design and physical simulation. Interpolation models such as SIMP / RAMP are used to avoid abrupt changes in physical properties caused by the "checkerboard phenomenon" in the density method, and to ensure the stability of the flow field and temperature field calculation results (such as the smooth change of pseudo-density field can reduce numerical calculation errors). Step S4: Solve the governing equations, solve the state variables, and perform sensitivity analysis; The solvers used in this step include, but are not limited to, commercial software (such as COMSOL) and open-source software (such as Open-FOAM); The algorithms used in sensitivity analysis include, but are not limited to, the adjoint method and the finite difference method. With the help of commercial / open-source solvers (such as COMSOL), physical quantities such as flow field velocity and temperature distribution can be accurately calculated, providing reliable data support for optimization (e.g. Figure 5 The pseudo-density field needs to be calculated based on an accurate temperature field. The adjoint method / finite difference method can quickly calculate the sensitivity of design variables to the objective function, and identify which characteristic parameters (such as B-spline control points) can be adjusted to maximize performance, thus avoiding blind iteration. Step S5: Update the design variables using the optimization algorithm and determine whether the convergence requirement is met. If it is met, proceed to the next step; otherwise, return to step S3 to continue the optimization iteration until the convergence condition is met. The optimization algorithm uses the moving asymptote method (MMA). The convergence requirement adopted is to satisfy either of the following two conditions: The absolute value of the difference between the objective function value of the current iteration step and the previous iteration step, divided by the difference in the function value of the current iteration step, is less than a specific proportion, which is generally less than or equal to 1%. The number of iterations is greater than a certain value, which is generally greater than or equal to 20; The Moving Asymptote Method (MMA) is characterized by fast convergence and high stability. Compared with the traditional gradient descent method, it can find the optimal solution in fewer iterations (e.g., convergence in 300 iterations in this case). Meanwhile, the rate of change of the objective function (≤1%) and the number of iterations (≥20) are used as convergence criteria to ensure optimization accuracy and prevent infinite iterations caused by local extrema, thus balancing computational efficiency and result reliability. Step S6: Output the optimization results and perform post-processing to obtain a flow channel optimization design scheme that meets the requirements.

[0034] When step S5 meets the convergence requirement, the optimization process will end the iteration.

[0035] This step will output the optimized results, including but not limited to: fluid velocity field, temperature field, density field, zero level set curve, etc. The output zero-level set curve is the flow channel boundary, which can be directly imported into CAD software to generate a solid model, without the need for the complex boundary reconstruction process of the density method (such as...). Figure 4 The feature boundaries can be directly used for machining drawing ( By outputting data such as velocity field and temperature field, the optimization effect can be fully verified (e.g., in the case study, the radiator temperature decreased after optimization, proving that the heat dissipation efficiency was improved), providing a quantitative evaluation basis for the design scheme.

[0036] The feature parameterization model output in this step can be directly applied to scenarios such as 3D heat sinks and manifolds. By adjusting the feature type (such as expanding from 2D B-splines to 3D NURBS), the design concept can be quickly transferred. Modeling the physical domain using features and mapping it to a pseudo-density field for optimization design essentially involves altering the shape of the features by changing their geometric parameters, thereby driving the evolution of the overall structure. Therefore, both shape optimization and topology optimization can be performed. The material boundary of the optimized result is the geometric boundary of the feature, resulting in a clear structural boundary and avoiding issues like jagged boundaries and checkerboard patterns seen in density methods. Furthermore, features can generally be evolved using only a few geometric parameters, which serve as the design variables for optimization, thus significantly reducing the number of design variables compared to density methods.

[0037] against Figure 2 The two-dimensional heat sink configuration shown is modeled and topology optimized using closed B-spline features (CBS). Figure 3 The feature distribution diagram of the initial layout in the topology optimization design is given. Figure 4 The feature distribution map of the final optimization result is given. Figure 5 The overall pseudo-density field corresponding to the final optimization result is given.

[0038] This layout differs from the traditional straight-channel design based on empirical methods, through... Figure 4 The feature boundary can be extracted to obtain a clear and complete structural boundary, resulting in a complex configuration similar to bionics. This allows for reasonable distribution of flow, effectively reducing the temperature of the two-dimensional heat sink and thus improving heat dissipation efficiency.

[0039] Furthermore, the design concept of this method can be easily extended to topology optimization design in other fields, including but not limited to three-dimensional heat sink structures, manifold structures, and topology optimization design of fluid-thermal-solid multi-field coupling, so as to improve the specific performance of the structure.

[0040] The above formulas are derived from software simulations using a large amount of data and are selected to be close to the actual values. The influence weight factors and specific coefficient values ​​in the formulas are set by those skilled in the art based on the actual situation and can be adjusted and modified in the future.

[0041] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, as a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer program are loaded or executed on a computer, all or part of the processes or functions described in the embodiments of this application are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device.

[0042] The computer instructions may be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions may be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired or wireless means (e.g., infrared, wireless, microwave, etc.).

[0043] The computer-readable storage medium can be any available medium that a computer can access, or a data storage device such as a server or data center that includes one or more sets of available media. The available medium can be a magnetic medium, an optical medium, or a semiconductor medium. A semiconductor medium can be a solid-state drive (SSD).

[0044] It should be understood that the term "and / or" in this article is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, or B existing alone. A and B can be singular or plural. Additionally, the character " / " in this article generally indicates an "or" relationship between the preceding and following related objects, but it can also represent an "and / or" relationship. Please refer to the context for a more accurate understanding.

[0045] In the several embodiments provided in this application, it should be understood that the disclosed systems, apparatus and methods can be implemented in other ways.

[0046] For example, the device embodiments described above are merely illustrative. For instance, the division of units is only a logical functional division; in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the displayed or discussed mutual couplings, direct couplings, or communication connections may be indirect couplings or communication connections between devices or units through some interfaces, and may be electrical, mechanical, or other forms.

[0047] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0048] In addition, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0049] If the aforementioned functions are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application.

[0050] The aforementioned storage media include various media that can store program code, such as USB flash drives, portable hard drives, read-only memory, random access memory, magnetic disks, or optical disks.

[0051] The above description of the embodiments enables those skilled in the art to make or use the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A feature-driven runner design optimization method, characterized in that, The application relates to a method for optimizing a flow channel structure, and belongs to the technical field of optimization design. Step S1: determining physical domain, physical model, optimization model information of a structure; Step S2: selecting suitable implicit features and performing parameterized modeling on the physical domain; Step S3: obtaining a corresponding pseudo-density field through a level set function mapping of the physical domain, and performing interpolation on physical property parameters; Step S4: solving a control equation, solving a state variable and performing sensitivity analysis; Step S5: updating design variables through an optimization algorithm, simultaneously judging whether a convergence requirement is met, if the convergence requirement is met, the next step is performed, otherwise, the optimization iteration is continued in step S3 until the convergence condition is met; Step S6: outputting an optimization result and performing post-processing to obtain a flow channel optimization design scheme meeting a requirement.

2. The feature-driven runner design optimization method of claim 1, wherein, The physical domain in step S1 comprises a design domain and a non-design domain, and boundary conditions comprise flow-related inlet, outlet and wall conditions and heat-related heat source and adiabatic boundary conditions.

3. The feature-driven runner design optimization method of claim 1, wherein, The optimization model adopts the following mathematical expression: ； Minimize: ; s.t. , where denotes a set of design variables, and denotes upper and lower bounds on the design variables, denotes an objective function, denotes all inequality constraints; the governing equations represent all flow and heat related physical equations.

4. The feature-driven runner design optimization method of claim 3, wherein, The design variables are geometric parameters of all features.

5. The feature-driven runner design optimization method of claim 4, wherein, The implicit features comprise but are not limited to a circle, a super-ellipse, a closed B-spline, a smooth deformation implicit curve and a B-spline offset feature. The design variables are a geometric parameter set of the features.

6. The feature-driven runner design optimization method of claim 1, wherein, The parameterized modeling manner of the physical domain refers to establishing a level set function of the whole physical domain, and judging whether an arbitrary internal point belongs to a fluid domain, a fluid boundary or a solid domain through a function value of the point.

7. The feature-driven runner design optimization method of claim 6, wherein, The parameterized modeling selects a function operation mode similar to a Boolean operation to realize "and" and "intersection" operations, and comprises but is not limited to max / min, R function, p norm and KS function.

8. The feature-driven runner design optimization method of claim 1, wherein, The level set function of the physical domain forms a pseudo-density field corresponding to each grid cell through a mapping form. The mapping form is realized through different forms of Heaviside functions.

9. The feature-driven runner design optimization method of claim 1, wherein, The algorithm adopted in the sensitivity analysis in step S3 comprises but is not limited to an adjoint method and a finite difference method.

10. The feature-driven runner design optimization method of claim 1, wherein, When step S5 meets the convergence requirement in step S6, the optimization process will end iteration; The optimization result outputted in the step comprises but is not limited to a fluid velocity field, a temperature field, a density field and a zero level set curve.

Images