Cooling fan optimization design method and device, electronic equipment and storage medium

Abstract

The invention provides a cooling fan optimization design method and device, electronic equipment and a storage medium, and relates to the technical field of electronic equipment heat dissipation, and the method comprises the steps: obtaining an initial design parameter combination and an initial target function value, inputting the initial design parameter combination and the initial objective function value into a Bayesian optimizer to obtain an optimized design parameter combination; based on the optimized design parameter combination, calculating an optimized objective function value and fan noise data; judging whether the fan noise data and the optimized target function value meet preset conditions or not; and if not, continuing to optimize, and if so, outputting the optimized design parameter combination. By implementing the technical scheme provided by the invention, the technical problem that the optimization design time of the current cooling fan is too long is solved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of electronic equipment heat dissipation, and in particular to a heat dissipation fan optimization design method and device, electronic equipment and storage medium. BACKGROUND

[0002] In a CPU air-cooled heat dissipation module, a heat dissipation fan is a key component that generates airflow by rotation to carry away heat on the heat dissipation fins. In actual application, the inlet and outlet static pressure difference and air volume of the heat dissipation fan directly affect the heat dissipation efficiency, and the fan speed affects the noise and power consumption. Therefore, when optimizing the design and performance evaluation of the heat dissipation fan, the air pressure, air volume and noise need to be considered comprehensively.

[0003] The current optimization design of the heat dissipation fan mainly adjusts the installation angle and thickness of the blades. For a single blade, the installation angle and thickness at different radial heights and axial lengths have certain adjustable space. In the optimization process, engineers usually preliminarily design these parameters according to experience. In order to simplify the design process, the change curve of the blade installation angle along the axial position is often simplified as a linear function or a parabolic function. Then the design effect is verified through physical prototype production and performance test, and the test content includes the maximum pressure difference under zero air volume, the maximum air volume under zero pressure difference, and the noise level under a certain speed and other parameters. Based on the test results, the engineers adjust the design parameters according to experience and repeat the process of producing prototypes and testing.

[0004] This optimization method based on experience and repeated trial and error has obvious limitations. First of all, each design adjustment needs to produce a new prototype and conduct comprehensive performance test, which not only consumes a lot of time and cost, but also requires a lot of human resources. Secondly, even if the design parameter space is reduced to three dimensions, it is still extremely difficult to find the optimal solution that meets multiple performance indicators, and only relying on experience can easily fall into local optimization. Thirdly, this method highly depends on the personal experience and data accumulation of engineers, and when facing different air volume and pressure demand occasions, the existing experience may not be completely applicable, and a lot of experimental exploration work needs to be carried out again, which is difficult to quickly respond to market demand. SUMMARY

[0005] In order to solve the technical problem of too long optimization design time of the current heat dissipation fan, the present application provides a heat dissipation fan optimization design method, device, electronic equipment and storage medium.

[0006] In a first aspect, the present application provides a heat dissipation fan optimization design method, comprising: Obtain the initial design parameter combination and the initial objective function value, wherein the initial design parameter combination represents the blade shape of the initial cooling fan, the initial objective function value is the sum of the products of a number of airflow points and the corresponding air pressures of a number of airflow points, and the initial objective function value represents the performance of the initial cooling fan; The initial design parameter combination is iteratively optimized to obtain the optimized design parameter combination and the optimized objective function value: The initial design parameter combination and the initial objective function value are input into the Bayesian optimizer to obtain the optimized design parameter combination. Based on the optimized design parameter combination, calculate the optimized objective function value and fan noise data corresponding to the optimized design parameter combination; Determine whether the fan noise data and the optimized objective function value meet preset conditions; When the fan noise data and the optimized objective function value do not meet the preset conditions, the optimized design parameter combination and the optimized objective function value are used as the new initial design parameter combination and the new initial objective function value, and input into the Bayesian optimizer to continue the optimization operation; or the initial design parameter combination and the zero-value objective function value are re-inputted into the Bayesian optimizer to continue the optimization operation, where the zero-value objective function value is the optimized objective function value after being forcibly zeroed. When the fan noise data and the optimized objective function value meet the preset conditions, the loop operation ends and the optimized design parameter combination is output.

[0007] By adopting the above technical solution, the initial design parameter combination and objective function value are first obtained. The objective function value considers the air pressure characteristics at multiple airflow points, which can more comprehensively reflect the performance of the cooling fan. Then, the initial parameters are optimized using a Bayesian optimizer, and the objective function value and noise data are calculated based on the optimized parameters. When the preset conditions are not met, the system will choose whether to continue using the optimized parameters and objective function value or re-use the initial parameters and zero-value objective function value for the next round of optimization, depending on the actual situation. This iterative optimization mechanism can continuously improve performance while ensuring controllable noise. At the same time, by setting reasonable termination conditions, over-optimization is avoided. This optimization method based on Bayesian inference can find a near-optimal combination of design parameters in a smaller number of iterations, greatly improving the design efficiency of the cooling fan.

[0008] Optionally, the step of obtaining the initial design parameter combination specifically includes: The blades of the initial heat dissipation fan are radially divided into several layers, a first difference value of a leading edge installation angle between adjacent two layers of blades is obtained, and a second difference value of a trailing edge installation angle and the leading edge installation angle of each layer of blades is obtained; According to a corresponding relationship between a coordinate of an axial position of each layer of blades and a blade installation angle at the corresponding position, an initial parabolic function of the blade installation angle of each layer is constructed, and a first term coefficient of the initial parabolic function is determined, the initial parabolic function representing a change of the blade installation angle of each layer along an axial direction of the initial heat dissipation fan; The first difference value, the second difference value, and the first term coefficient of the parabolic function are combined as the initial design parameter combination.

[0009] By adopting the technical solution, the blades are first divided into layers along the radial direction, and the complex three-dimensional blade shape can be more accurately described by the layering method. By obtaining the difference value of the leading edge installation angle between adjacent layers and the difference value of the leading edge installation angle and the trailing edge installation angle of each layer of blades, the axial change of the blade installation angle is described by combining the parabolic function, so that the blade shape parameterization is possible. The introduction of the first term coefficient can effectively control the change rate of the blade installation angle and avoid excessive local change. This parameterization method not only retains the key features of the blade shape, but also reduces the design space to a controllable dimension, laying a foundation for subsequent optimization calculation, and making the optimization process more efficient and reliable.

[0010] Optionally, the step of obtaining the initial target function value specifically includes: According to the working scene of the initial heat dissipation fan, a plurality of working air volume points are selected; A first air pressure corresponding to each working air volume point is collected; The initial target function value is calculated according to the working air volume points and the first air pressure.

[0011] By adopting the technical solution, a plurality of working air volume points are selected according to the actual working scene of the initial heat dissipation fan, and the selection method ensures that the target function can truly reflect the performance of the fan in actual application. By collecting the air pressure data corresponding to each working air volume point and using the data to calculate the initial target function value, the optimization process is closer to the actual application requirements. This target function construction method based on the actual working scene avoids the limitation of the traditional method which only focuses on the limit working condition (such as zero air volume maximum pressure difference and zero pressure difference maximum air volume), and can more accurately guide the subsequent optimization process, improving the practicality of the optimization result.

[0012] Optionally, the step of calculating the optimized target function value and the fan noise data corresponding to the optimized design parameter combination based on the optimized design parameter combination specifically includes: According to the optimized design parameter combination, a new parabolic function of each layer of blades is established; According to the new parabolic function, the fan is modeled and performance simulated to obtain a second wind pressure corresponding to each of the working air volume points and a first noise value corresponding to each of the working air volume points, the fan noise data being a set of all the first noise values; According to each of the working air volume points and the corresponding second wind pressure, the optimized target function value is calculated.

[0013] By adopting the above technical solution, firstly, the new parabolic function of each layer of blades is established according to the optimized parameter combination, so that the accurate mathematical description of the blade shape is realized. Then, modeling and performance simulation are performed based on the parabolic functions, so that not only the wind pressure data of each working point can be obtained, but also the noise data can be obtained at the same time. Since the noise data is calculated based on complete flow field information, it is more accurate and reliable than traditional empirical estimation. Finally, the wind pressure data obtained through simulation is combined with the working air volume points to calculate the target function value. This calculation method considers the performance under multiple typical working conditions. This evaluation method based on accurate modeling and comprehensive simulation can quickly and accurately evaluate the performance and noise characteristics of the design scheme without manufacturing a physical prototype, thereby greatly improving the efficiency of the optimization process.

[0014] Optionally, when the fan noise data and the optimized target function value do not satisfy the preset condition, the optimized design parameter combination and the optimized target function value are input to the Bayesian optimizer as a new initial design parameter combination and a new initial target function value to continue the optimization operation, specifically including: When each of the first noise values is less than a preset noise value and a target function increment is less than or equal to a preset threshold, the optimized design parameter combination and the optimized target function value are input to the Bayesian optimizer as a new initial design parameter combination and a new initial target function value to continue the optimization operation, the preset noise value being a maximum allowed noise value determined according to the initial cooling fan working scene, and the target function increment being an increment of the optimized target function value relative to the initial target function value.

[0015] By adopting the technical scheme, the double judgment conditions based on the noise value and the objective function increment are set. When the first noise value is less than the preset noise value, it is indicated that the optimization result meets the noise control requirement; and when the objective function increment is less than the preset threshold, it is indicated that there is still room for improvement in the current optimization direction. In this case, the optimized parameter is taken as a new initial value for continuous optimization, which not only ensures the continuity of the optimization process, but also avoids falling into local optimization. The dynamic optimization strategy based on performance and noise can continuously improve the performance of the cooling fan until the optimal design scheme is reached under the premise of ensuring controllable noise.

[0016] Optionally, when the fan noise data and the optimized objective function value do not meet the preset condition, the initial design parameter combination and the zero value objective function value are input to the Bayesian optimizer for continuous optimization operation, specifically including: When any of the first noise values in the fan noise data is greater than or equal to the preset noise value, the optimized design parameter combination and the zero value objective function value are input to the Bayesian optimizer for continuous optimization operation. The preset noise value is the maximum allowable noise value determined according to the initial cooling fan working scene.

[0017] By adopting the technical scheme, the objective function of the optimization scheme exceeding the preset noise value is forced to be zero, and the zero value objective function value and the corresponding design parameter combination are input to the Bayesian optimizer, which can effectively avoid the continuous exploration of the optimization process in the high noise direction. This forced zero strategy directly affects the sampling probability distribution of the Bayesian optimizer, which automatically reduces the exploration tendency of the high noise parameter region in the subsequent optimization process, thereby guiding the optimization direction to develop towards the low noise region.

[0018] Optionally, when the fan noise data and the optimized objective function value meet the preset condition, the loop operation is ended, and the optimized design parameter combination is output, specifically including: When each of the first noise values is less than the preset noise value, and the objective function increment is greater than the preset threshold, the loop operation is ended, and the optimized design parameter combination is output. The preset noise value is the maximum allowable noise value determined according to the initial cooling fan working scene, and the objective function increment is the increment of the optimized objective function value relative to the initial objective function value.

[0019] By adopting the technical scheme, the termination condition of the optimization process is set: the noise control is below the preset value, and the objective function increment exceeds the preset threshold. This termination condition not only ensures that the optimization result meets the noise requirement, but also ensures that the performance improvement reaches the expected target. When these conditions are met, a satisfactory design scheme in terms of performance and noise has been found, and the optimization process is terminated in time and the result is output, avoiding resource waste caused by over-optimization. This intelligent termination mechanism improves the efficiency and practicality of the optimization process.

[0020] In a second aspect of the present application, a heat dissipation fan optimization design device is also provided, comprising: An acquisition module is configured to acquire an initial design parameter combination and an initial objective function value. An input module is configured to input the initial design parameter combination and the initial objective function value into a Bayesian optimizer to obtain an optimized design parameter combination. A calculation module is configured to calculate, based on the optimized design parameter combination, an optimized objective function value corresponding to the optimized design parameter combination and fan noise data. A judgment module is configured to judge whether the fan noise data and the optimized objective function value meet a preset condition. A processing module is configured to, when the fan noise data and the optimized objective function value do not meet the preset condition, input the optimized design parameter combination and the optimized objective function value as a new initial design parameter combination and a new initial objective function value into the Bayesian optimizer to continue optimization operation, or re-input the initial design parameter combination and a zero-value objective function value into the Bayesian optimizer to continue optimization operation; and when the fan noise data and the optimized objective function value meet the preset condition, end the loop operation and output the optimized design parameter combination.

[0021] In a third aspect of the present application, an electronic device is also provided, comprising a memory and a processor, the memory having a computer program stored thereon, and the processor implements the method steps of any one of the above aspects when executing the program.

[0022] In a fourth aspect of the present application, a computer readable storage medium is also provided, the computer readable storage medium storing instructions which, when executed, perform the method steps of any one of the above aspects.

[0023] To sum up, one or more technical solutions provided in the present application have at least the following technical effects or advantages: 1. By adopting the technical scheme, first, the initial design parameter combination and the target function value are obtained, wherein the target function value considers the wind pressure characteristics of multiple air volume points, and can more comprehensively reflect the performance of the cooling fan. Then, the initial parameters are optimized by the Bayesian optimizer, and the target function value and the noise data are calculated based on the optimized parameters. When the preset condition is not met, the system will select whether to continue using the optimized parameters and the target function value or to use the initial parameters and the zero value target function value for the next round of optimization according to the actual situation. This iterative optimization mechanism can continuously improve the performance under the premise of controllable noise. At the same time, by setting reasonable termination conditions, over-optimization is avoided. This optimization method based on Bayesian inference can find the design parameter combination close to the optimal value in a small number of iterations, greatly improving the design efficiency of the cooling fan.

[0024] 2. By adopting the technical scheme, first, the blades are processed in a radial direction. This layering method makes it possible to more accurately describe the complex three-dimensional blade shape. By obtaining the difference in leading edge installation angle between adjacent layers and the difference in leading edge and trailing edge installation angle of each layer of blades, and combining a parabolic function to describe the axial variation of the blade installation angle, the blade shape parameterization becomes possible. The introduction of the first-order coefficient can effectively control the rate of change of the blade installation angle and avoid excessive local changes. This parameterization method not only retains the key features of the blade shape, but also reduces the design space to a controllable dimension, laying a foundation for subsequent optimization calculations and making the optimization process more efficient and reliable.

[0025] 3. By adopting the technical scheme, multiple working air volume points are selected according to the actual working scene of the initial cooling fan. This selection method ensures that the target function can truly reflect the performance of the fan in actual application. By collecting the wind pressure data corresponding to each working air volume point and using these data to calculate the initial target function value, the optimization process is more close to the actual application requirements. This target function construction method based on the actual working scene avoids the limitations of traditional methods that only focus on extreme conditions (such as maximum pressure difference at zero air volume and maximum air volume at zero pressure difference), and can more accurately guide the subsequent optimization process, improving the practicality of the optimization results. BRIEF DESCRIPTION OF DRAWINGS

[0026] Figure 1 is a flow chart of a cooling fan optimization design method provided by an embodiment of the present application; Figure 2 is a flow chart of another cooling fan optimization design device provided by an embodiment of the present application; Figure 3 is a structural block diagram of a cooling fan optimization design device provided by an embodiment of the present application; Figure 4is a structural schematic diagram of an electronic device disclosed by an embodiment of the present application.

[0027] Label explanation: 400-electronic device; 401-processor; 402-communication bus; 403-user interface; 404-network interface; 405-memory. DETAILED DESCRIPTION

[0028] In order for those skilled in the art to better understand the technical solutions in the specification, the technical solutions in the specification will be clearly and completely described below in combination with the drawings in the embodiments of the specification. Obviously, the described embodiments are only some of the embodiments of the present application, not all the embodiments.

[0029] In the description of the embodiments of the present application, the words such as "for example" or "for instance" are used to represent an example, illustration or description. Any embodiment or design scheme described as "for example" or "for instance" in the embodiments of the present application should not be interpreted as more preferred or more advantageous than other embodiments or design schemes. Rather, the words such as "for example" or "for instance" are intended to present the relevant concept in a specific manner.

[0030] In the description of the embodiments of the present application, the term "a plurality of" means two or more. In addition, the terms "first", "second" are used for description purposes only, and should not be interpreted as indicating or implying relative importance or implicitly indicating the indicated technical features. Therefore, the features defined with "first", "second" can be explicitly or implicitly included one or more features. The terms "include", "contain", "have" and their variants mean "include but are not limited to", unless otherwise specifically emphasized.

[0031] The present application provides a heat dissipation fan optimization design method, referring to Figure 1 , Figure 1 is a flow chart of a heat dissipation fan optimization design method provided by an embodiment of the present application, and the method comprises: Step S101, obtaining an initial design parameter combination and an initial target function value; Wherein, the initial design parameter combination represents a set of geometric shape parameters of the initial heat dissipation fan blade, which is used for subsequent simulation test on the initial heat dissipation fan; the initial target function value refers to a quantitative evaluation index of the performance of the initial heat dissipation fan, which is composed of the sum of the products of multiple working air volume points and their corresponding air pressure, for example, the sum of the products of air volume and air pressure at air volume points of 70, 80, 90 and 100 m³ / h.

[0032] When starting the optimization design process of the heat dissipation fan, the relevant parameters and performance indicators of the initial design scheme need to be obtained first. Specifically, first, the initial heat dissipation fan is modeled and simulated, the blades are layered along the radial direction, and the installation angle difference between the layers is obtained; then, a plurality of representative air volume points are selected based on the actual working scene, and the corresponding wind pressure data of the air volume points are obtained through simulation or experimental testing; finally, the initial target function value is calculated and taken as the benchmark for subsequent optimization.

[0033] In some embodiments, the initial design parameter combination and the target function value can be obtained in various ways: alternatively, the geometric parameters can be obtained by three-dimensional scanning of the physical fan, the performance data can be measured by wind tunnel experiment, and finally the target function value can be calculated; alternatively, the geometric parameters can be obtained by modeling through CAD software, the performance data can be obtained by CFD simulation, and then the target function value can be calculated. It can be understood that other ways of obtaining initial data can also be used, such as reverse engineering or referring to existing design schemes, which are not limited here.

[0034] The optimization operation is performed on the initial design parameter combination to obtain an optimized design parameter combination and an optimized target function value, specifically including steps S102 to S106: Step S102, inputting the initial design parameter combination and the initial target function value into a Bayesian optimizer to obtain the optimized design parameter combination. The Bayesian optimizer refers to an intelligent optimization algorithm tool based on Bayesian inference, which is used to search for the optimal solution under given constraints; the optimized design parameter combination represents a new set of blade geometric parameters obtained after optimization.

[0035] After obtaining the initial data, the Bayesian optimizer is used to optimize the design parameters. Specifically, the initial design parameter combination is taken as the search starting point, and the initial target function value is taken as the reference benchmark to input into the optimizer. The optimizer searches for the possible optimal solution in the design parameter space according to the Bayesian inference principle, and outputs a new set of design parameter combinations.

[0036] In some embodiments, the Bayesian optimization process can be implemented in various ways: alternatively, the scikit-optimize library in the Python programming language can be used to implement Bayesian optimization, and appropriate search range and iteration number can be set; alternatively, MATLAB's Optimization Toolbox can be used for Bayesian optimization, and the optimization process can be controlled through custom objective function and constraint conditions. It can be understood that other optimization algorithms or tools can also be used to implement parameter optimization, which are not limited here.

[0037] Step S103, based on the optimized design parameter combination, calculate the optimized objective function value and fan noise data corresponding to the optimized design parameter combination; Wherein, the optimized objective function value represents the performance evaluation index of the new design scheme; the fan noise data refers to the noise level set under specific working conditions, which is used to evaluate the noise performance of the new design scheme.

[0038] After obtaining the optimized design parameter combination, performance evaluation of the new scheme is needed. Specifically, first, a new fan model is established according to the optimized parameters, then the fan pressure and noise data of each working point are obtained through simulation or experiment, finally the new objective function value is calculated, and these data are used for subsequent judgment and optimization.

[0039] In some embodiments, performance evaluation can be achieved in various ways: optionally, CFD software is used for flow field simulation, aerodynamic performance and noise data are extracted, and then data processing and calculation are performed; optionally, a prototype is made for wind tunnel test, various performance data are collected, and evaluation results are obtained through data analysis. It can be understood that other ways can also be used for performance evaluation, such as using simplified calculation model or empirical formula, etc., which are not limited here.

[0040] Step S104, judge whether the fan noise data and the optimized objective function value meet the preset condition; Wherein, the preset condition represents the performance and noise requirements that the optimization result needs to meet, including the preset noise value and the objective function increment threshold; the fan noise data refers to the noise value set measured at each working point; the optimized objective function value represents the comprehensive performance index of the new design scheme.

[0041] After obtaining the optimized design parameter combination and the optimized objective function value, it is needed to judge whether it meets the design requirements. Specifically, for the noise data of each working point, it is needed to compare it with the maximum allowed noise value determined according to the working scene, to ensure that the noise level of the design scheme at all working points is within the acceptable range. At the same time, the increment between the optimized objective function value and the initial value is calculated, which reflects the degree of improvement of fan performance. Only when the noise values of all working points meet the requirements and the performance improvement reaches the expected target, can it be considered that the current design scheme meets the preset condition.

[0042] In some embodiments, the judgment process can be implemented in various ways: optionally, a judgment matrix is set, the noise data and the target function value are compared with the threshold value respectively, and the judgment result is generated according to the comparison result; optionally, a comprehensive evaluation model is established, and the noise and the performance index are weighted and calculated to obtain a single evaluation result. It can be understood that other ways can also be used for conditional judgment, such as fuzzy logic judgment or multi-criteria decision method, etc., which are not limited here.

[0043] Step S105, when the fan noise data and the optimized target function value do not satisfy the preset condition, input the optimized design parameter combination and the optimized target function value as new initial design parameter combination and new initial target function value to the Bayesian optimizer to continue the optimization operation, or input the initial design parameter combination and a zero value target function value to the Bayesian optimizer to continue the optimization operation, the zero value target function value being the optimized target function value forced to zero; Wherein, the initial design parameter combination represents the parameter set at the optimization starting point; the initial target function value refers to the performance index at the optimization starting point; the new initial design parameter combination and the target function value represent the starting data for the next round of optimization.

[0044] When the optimization result does not meet the requirements, optimization needs to be continued. Specifically, the system first analyzes the specific reason why the optimization result does not meet the requirements: if it is due to excessive noise, it means that the current optimization direction may be too aggressive, and the optimization path needs to be reselected; if it is due to insufficient performance improvement, it means that the current optimization direction is feasible, but needs to be further improved. Based on the analysis result, the system will adjust the optimization strategy accordingly: either abandon the current result and start optimization again, or continue in-depth optimization based on the current result. This dynamic adjustment mechanism can ensure that the optimization process is always in the direction of controlling noise and improving performance.

[0045] In some embodiments, the optimization strategy selection can be implemented in various ways: optionally, a strategy selector is set, and the optimization path is automatically selected according to the noise exceeding standard and the target function increment; optionally, a decision tree model is established, and the next optimization scheme is determined after considering multiple factors. It can be understood that other ways can also be used to realize the optimization strategy selection, such as expert system or deep learning method, etc., which are not limited here.

[0046] Step S106, when the fan noise data and the optimized target function value satisfy the preset condition, end the loop operation and output the optimized design parameter combination.

[0047] The preset condition represents a judgment standard for optimization termination, including a noise control requirement and a performance improvement target; and the optimized design parameter combination represents a final determined fan design scheme.

[0048] When the optimization result meets all design requirements, the optimization process ends. Specifically, the system first verifies the noise control effect of the current design scheme to ensure that the noise value at all working points is lower than the maximum allowed noise value determined according to the working scene. At the same time, the system checks the degree of improvement of the objective function value to confirm that the performance improvement meets the expected target. When both aspects meet the requirements, it indicates that a design scheme that can ensure low noise operation and significantly improve performance has been found. At this time, the system will terminate the optimization loop and output the final design parameter combination as the optimization result. This double-standard-based termination mechanism can ensure that the output design scheme has practical application value.

[0049] In some embodiments, optimization termination and result output can be achieved in various ways: optionally, an automatic termination program is set to automatically stop optimization and save the result when the condition is met; optionally, a result verification and output system is established to fully verify the final scheme before outputting the design parameters. It can be understood that other ways can also be used to achieve optimization termination and result output, such as manual confirmation or distributed processing, which are not limited here.

[0050] By using the above technical solution, the initial design parameter combination and the objective function value are first obtained, wherein the objective function value considers the wind pressure characteristics of multiple air volume points, which can more comprehensively reflect the performance of the cooling fan. Then, the initial parameters are optimized by the Bayesian optimizer, and the objective function value and noise data are calculated based on the optimized parameters. When the preset condition is not met, the system will select whether to continue using the optimized parameters or to use the initial parameters for the next round of optimization according to the actual situation. This iterative optimization mechanism can continuously improve performance while ensuring controllable noise. At the same time, by setting reasonable termination conditions, over-optimization is avoided. This optimization method based on Bayesian inference can find a design parameter combination close to the optimal one in fewer iterations, greatly improving the design efficiency of the cooling fan.

[0051] The method provided by the embodiment is described in further detail.

[0052] S10101 to S10108 are more specific schemes of step S101 in the embodiment of the application.

[0053] S10101 to S10103 are specific schemes for obtaining the initial design parameter combination: S10101, radially divide the blades of the initial heat dissipation fan into several layers, obtain a first difference value of a leading edge installation angle between adjacent two layers of blades and a second difference value of the installation angle of a trailing edge of each layer of blades and a leading edge; wherein the radial layering indicates that the blades are divided into multiple cross sections along the radial direction of the fan; the first difference value of the leading edge installation angle refers to the difference of the installation angle at the leading edge of the adjacent two layers of blades, denoted as ; the second difference value indicates the difference of the installation angle of the trailing edge and the leading edge of the same layer of blades, denoted as , wherein t is the thickness of the blade.

[0054] After obtaining the fan model, the blades need to be parameterized. Specifically, first, the blades are evenly divided into several layers along the radial direction of the fan, and the present scheme adopts 5 layers, and the interval of each layer is determined according to the size of the blade and the accuracy requirement; then the installation angle of the leading edge of each layer is measured, and the angle difference value between adjacent layers is calculated; finally, the installation angle of the trailing edge and the leading edge of each layer of blades is measured, and the difference value is calculated. This layered measurement method can accurately capture the three-dimensional twisting characteristics of the blade.

[0055] In some embodiments, the blade layering analysis can be implemented in various ways: optionally, the cross section analysis tool of the CAD software is used to automatically extract the geometric parameters of each layer of blades; optionally, a program is written to automatically divide the number of layers of blades and calculate the parameters of each layer. It can be understood that other ways can also be used to realize the blade layering analysis, such as manual measurement or optical scanning, etc., which are not limited here.

[0056] S10102, according to the corresponding relationship between the coordinates of the axial position of each layer of blades and the installation angle of the blades at the corresponding position, constructing an initial parabolic function of the installation angle of each layer of blades, and determining a linear term coefficient of the initial parabolic function; wherein the initial parabolic function of the installation angle of the blades indicates a mathematical model describing the axial variation of the installation angle of the blades; the axial variation refers to the variation trend from the leading edge to the trailing edge of the blade.

[0057] Establishing a mathematical model for each layer of blades is a key step to realize parameterized design. Specifically, the corresponding relationship between the coordinates of the axial position of each layer of blades and the installation angle of the blades at the corresponding position is used to fit the axial variation law of the installation angle of the blades by a parabolic function. The parabolic function of the present scheme is: , wherein z is the relative coordinate of the axial position, and the value interval is usually (0, 1), wherein is the chordwise gradient adjustment coefficient, controlling the linear slope (inclination degree of the straight line part) of the installation angle change, determining the average twisting rate of the blade from the leading edge to the trailing edge, is the chord-wise curvature adjustment coefficient, controls the curvature (convexity and bending strength of the curve) of the installation angle change, and determines the local bending strength or weakness of the middle section of the blade. This parabolic function can more accurately describe the torsion characteristics of the blade than the linear function, and is easier to control than the high-order polynomial. The first-order coefficient of the parabolic function is determined , which is used for subsequent optimization.

[0058] In some embodiments, the construction and coefficient determination of the blade installation angle parabolic function can be achieved in various ways: Optionally, the parabolic function is determined by least squares fitting: select multiple axial position points on each layer of blades, measure the corresponding installation angle data to form coordinate point pairs; use the least squares method to establish a parabolic fitting equation and calculate the fitting error; optimize the fitting parameters through iteration until the error meets the accuracy requirement; extract the first-order coefficient value from the fitting result. Optionally, the parabolic function is constructed using the characteristic point interpolation method: select three characteristic points at the leading edge, middle and trailing edge of each layer of blades; establish an equation set based on the coordinates and installation angle values of the three points; solve the equation set to obtain the coefficients of the parabolic function; verify the fitting effect of the obtained function at other position points. It can be understood that other ways can also be used to construct and determine the coefficients of the parabolic function, such as spline interpolation, neural network fitting or other numerical methods, etc., which are not limited here.

[0059] S10103, combine the first difference value, the second difference value and the first-order coefficient of the parabolic function as the initial design parameter combination; Wherein, the initial design parameter combination represents a parameter set used for subsequent optimization, which contains key parameters describing the geometric characteristics of the blade.

[0060] After completing the parameter extraction, the parameter set required for optimization needs to be arranged. Specifically, the first difference value, the second difference value and the first-order coefficient obtained in the previous steps are combined into a complete parameter set. These parameters together determine the three-dimensional shape of the blade, and by adjusting these parameters, the optimization of the fan performance can be realized.

[0061] In some embodiments, the parameter combination can be achieved in various ways: optionally, a parameter matrix is established to store all design parameters; optionally, a parameter object is created to encapsulate all design parameters. It can be understood that other ways can also be used to realize parameter combination, such as data structure optimization or distributed storage, etc., which are not limited here.

[0062] Wherein S10104 to S10106 are specific schemes for obtaining initial target function values: S10104, according to the working scene of the initial heat dissipation fan, select several working air volume points; The working scene represents an actual application environment and operating condition of the heat dissipation fan; the working air volume point refers to an air volume value capable of representing a typical operating state of the fan; and the working air volume points represent a group of representative air volume sampling points.

[0063] Specifically, first, the actual working scene of the heat dissipation fan is analyzed, including factors such as heat dissipation demand, installation space, and environmental temperature; and then, according to the analysis result, a typical air volume point capable of reflecting the actual operating state of the fan is selected. The selection of these air volume points should cover the main working range of the fan, including not only the rated operating condition but also part-load and overload operating conditions, to ensure the comprehensiveness of performance evaluation. For example, through preliminary testing and simulation, it can be preliminarily determined that the heat dissipation fan mainly works at an air volume of 70-100 m3 / h in the CPU heat dissipation scene, and therefore, 70, 80, 90, and 100 m3 / h are selected as the air volume.

[0064] In some embodiments, the working air volume points can be selected in various ways: alternatively, based on a load demand curve, several key load points are selected as the working air volume points; alternatively, a uniform distribution method is adopted to select air volume points at equal intervals within the working range of the fan. It can be understood that other ways can also be adopted to select the working air volume points, such as based on statistical analysis or expert experience, which are not limited here.

[0065] S10105, collecting a first wind pressure corresponding to each of the working air volume points; The first wind pressure represents a static pressure value that the fan can generate at a specific working air volume point; and the collection refers to a process of obtaining wind pressure data through testing or simulation.

[0066] After the working air volume points are determined, the corresponding performance data need to be obtained. Specifically, at each selected working air volume point, the static pressure value generated by the fan is obtained through simulation calculation or experimental testing. These wind pressure data reflect the output capacity of the fan under different operating conditions and are important indicators for evaluating the performance of the fan. To ensure the accuracy of the data, the collection needs to be carried out under stable working conditions.

[0067] In some embodiments, the wind pressure data can be collected in various ways: alternatively, CFD simulation calculation is used to obtain the wind pressure values of each working point; alternatively, a wind tunnel test bench is used to obtain the wind pressure data. It can be understood that other ways can also be adopted to collect the wind pressure data, such as semi-empirical formula calculation or field testing, which are not limited here.

[0068] S10106, calculating an initial target function value according to the working air volume points and the first wind pressure; The initial target function value represents a comprehensive evaluation index of the performance of the initial fan design scheme; and the calculation refers to a process of converting the working air volume points and the corresponding wind pressure data into a single evaluation index.

[0069] After obtaining the performance data, it is necessary to establish a comprehensive evaluation index. Specifically, the air volume value of each working air volume point is multiplied by the corresponding air pressure value to obtain the aerodynamic power of the working point. Then, the aerodynamic powers of all working points are added to obtain the target function value. This calculation method considers the performance of the fan under multiple working conditions and can more comprehensively reflect the overall performance level of the fan. In general, the larger the target function value, the better the performance of the cooling fan.

[0070] In some embodiments, the target function value can be calculated in various ways: alternatively, the product sum of the air volume and air pressure of each working point is directly calculated; alternatively, a weight coefficient is introduced to calculate the contribution of different working points. It can be understood that other ways of calculating the target function value can also be used, such as introducing efficiency factors or considering other performance indicators, which are not limited here.

[0071] S10301 to S10303 are more specific solutions of step S103 in the embodiments of the application.

[0072] S10301, according to the optimized design parameter combination, a new parabolic function of each layer of blades is established; Based on the parabolic function of the present solution, z is the axial position relative coordinate, the value range is [0, 1], where 0 represents the leading edge of the blade and 1 represents the trailing edge of the blade.

[0073] is the blade leading edge installation angle, is the blade trailing edge installation angle, is equal to the opening angle between the trailing edge and the leading edge of the blade The relative position coordinates corresponding to the leading edge and the trailing edge of the blade are substituted into the parabolic function, , Therefore Thus the relationship between , and is established, according to the values of and recommended by the Bayesian optimizer, the value of can be calculated, at this time the value of is determined, that is, the new parabolic function is determined.

[0074] S10302, according to the new parabolic function, the fan is modeled and performance simulation is performed to obtain the second air pressure corresponding to each working air volume point and the first noise value corresponding to each working air volume point; Wherein, the new parabolic function represents the variation rule of the optimized blade installation angle along the axial direction; the modeling refers to constructing a three-dimensional model of the fan according to geometric parameters; the performance simulation refers to simulating the working process of the fan through computational fluid dynamics software; the second wind pressure refers to the static pressure value obtained through simulation; and the first noise value refers to the sound pressure level obtained through simulation calculation.

[0075] Specifically, based on the optimized parabolic function, the blade model is first reconstructed in the CAD software to generate a complete three-dimensional model of the fan. Then the model is imported into the CFD software, boundary conditions and mesh parameters are set, and flow field simulation calculation is performed at each working air volume point. Through post-processing, the wind pressure data of each working point is extracted, and the aerodynamic noise is calculated based on the flow field information. This numerical simulation-based method can quickly evaluate the performance of the design scheme without making a physical prototype.

[0076] In some embodiments, fan modeling and performance simulation can be implemented in various ways: alternatively, a parametric modeling software can be used to automatically generate a blade model, which is imported into a CFD software for mesh division and solution setting, flow field calculation is performed and results are extracted; alternatively, a professional fan design software can be used to complete the whole process from modeling to simulation, and a built-in special solver in the software can provide more accurate results. It can be understood that other ways can also be used to implement fan modeling and performance simulation, such as semi-empirical calculation methods or hybrid simulation methods, which are not limited here.

[0077] S10303, according to each of the working air volume points and the corresponding second wind pressure, the optimized target function value is calculated; Wherein, the working air volume point represents the characteristic working condition when the fan is actually running; the second wind pressure refers to the static pressure value obtained through simulation calculation; and the optimized target function value represents the comprehensive performance index of the new design scheme.

[0078] Specifically, the wind volume value of each working point obtained through simulation is multiplied by the corresponding second wind pressure value to obtain the aerodynamic power of the point. Then the aerodynamic powers of all working points are added to calculate the optimized target function value. This calculation method considers the performance of the fan under multiple typical working conditions and can comprehensively reflect the pros and cons of the design scheme.

[0079] In some embodiments, the target function value calculation can be implemented in various ways: alternatively, an automated data processing program can be established to read the simulation result file, extract the wind volume and wind pressure data, and complete the calculation; alternatively, an Excel template can be built, and after the wind volume and wind pressure data are input, the target function value is automatically calculated and a performance report is generated. It can be understood that other ways can also be used to implement target function value calculation, such as online calculation tools or distributed computing systems, which are not limited here.

[0080] S10501 to S10502 are more specific solutions of step S105 in the embodiments of the present application.

[0081] S10501, when each of the first noise value is less than the preset noise value, and the objective function increment is less than or equal to the preset threshold, the optimized design parameter combination and the optimized objective function value are input as new initial design parameter combination and new initial objective function value to the Bayesian optimizer to continue the optimization operation, the preset noise value is the maximum allowed noise value determined according to the initial cooling fan working scene, and the objective function increment is the increment of the optimized objective function value relative to the initial objective function value; Wherein, the first noise value represents the noise level of the optimized fan at each working point; the preset noise value refers to the maximum allowed noise value determined according to the initial cooling fan working scene; the objective function increment represents the growth of the optimized objective function value relative to the initial objective function value; and the preset threshold refers to a standard value for judging whether the performance improvement is significant.

[0082] In the optimization process, it is necessary to judge the usability of the current optimization result. Specifically, first check whether all the first noise values of the working points are less than the preset noise value to ensure that the noise control is within the allowed range, and then calculate the objective function increment and compare it with the preset threshold to judge whether the performance improvement meets the expectation. When both conditions are met, it means that the current optimization direction is feasible but there is still room for improvement, so the optimized design parameter combination and the optimized objective function value are taken as new starting points for further optimization.

[0083] It should be noted that the preset noise value mainly has two types. When the performance of the cooling fan needs to be optimized without increasing the noise, for example, in an office use scenario where the noise requirement is high, the preset noise value should not exceed the noise value of the initial cooling fan, where the noise value of the initial cooling fan is the noise value of the initial cooling fan at the conventional working speed. When the noise requirement is not high in other use scenarios, the preset noise value is calculated according to the formula: preset noise value = noise value of initial cooling fan + 10*lg(1+x%), where x is the increment of the optimized objective function value relative to the initial objective function value.

[0084] In some embodiments, the condition judgment and parameter updating can be realized in various ways: optionally, an automatic judgment program can be established to monitor the noise and performance indicators in real time; optionally, a multi-level judgment matrix can be set up to comprehensively evaluate the optimization effect; and optionally, an automatic parameter updating mechanism can be established to ensure the continuity of the optimization process. It can be understood that other ways can also be used to realize the condition judgment and parameter updating, such as fuzzy logic judgment or expert system, which are not limited here.

[0085] S10502, when any of the first noise values in the fan noise data is greater than or equal to a preset noise value, re-inputting the optimized design parameter combination and the zero-value target function value to the Bayesian optimizer to continue the optimization operation, the preset noise value being a maximum allowed noise value determined according to the initial cooling fan working scenario; Wherein, the fan noise data represents a set of all the first noise values; the first noise value represents the noise level of the optimized fan at each working point; the preset noise value refers to the maximum allowed noise value determined according to the initial cooling fan working scenario; the zero-value target function value represents the optimized target function value after being forced to zero; the optimized design parameter combination represents the current optimized fan design scheme.

[0086] When the noise exceeds the standard, the optimization strategy needs to be adjusted forcibly. Specifically, when the system detects that any of the first noise values exceeds the preset noise value, the optimized target function value is forced to be set as the zero-value target function value, and is re-inputted to the Bayesian optimizer together with the optimized design parameter combination. This forced zero strategy can effectively guide the optimizer to avoid the design scheme with high noise, and promote it to explore the optimization direction with lower noise in the new iteration.

[0087] In some embodiments, the forced zero strategy can be implemented in various ways: optionally, a noise monitoring trigger mechanism is set to automatically execute target function zeroing when noise exceeds the standard; optionally, an optimization penalty mechanism is established to record known high-noise parameter combinations; optionally, an intelligent avoidance system is implemented to avoid repeated exploration of high-noise areas in the optimization process. It can be understood that other ways can also be used to implement the forced zero strategy, such as multi-objective optimization or adaptive weight, which are not limited here.

[0088] S10601, a more specific scheme for step S106 in the embodiments of the present application.

[0089] S10601, when each of the first noise values is less than the preset noise value and the increment is greater than the preset threshold, ending the loop operation and outputting the optimized design parameter combination; When the optimization result meets the noise control and performance improvement requirements at the same time, the optimization process is completed. Specifically, the system first checks whether the first noise value of all working points is lower than the preset noise value, ensuring that the noise control meets the standard; at the same time, it verifies whether the increment exceeds the preset threshold, confirming that the performance improvement is significant. When both conditions are met, it indicates that a design scheme that can ensure low noise operation and has significant performance improvement has been found, at which time the system terminates the optimization cycle and outputs the optimized design parameter combination as the final result.

[0090] In some embodiments, optimization termination and result output can be achieved in various ways: optionally, a comprehensive evaluation system is established to automatically determine whether the termination condition is met; optionally, a result verification mechanism is set up to comprehensively check the final scheme; optionally, an optimization result archiving system is implemented to record the final determined design parameters. It can be understood that other ways can also be used to achieve optimization termination and result output, such as manual confirmation or multiple verifications, etc., which are not limited here.

[0091] In addition, the embodiment also gives an explanation of the optimization design method of the cooling fan combined with a specific scene, please refer to Figure 2 ; Take a fan of a CPU air cooling radiator as an example to detail its optimization design process.

[0092] First, calculate the initial objective function: through preliminary testing and simulation, it is found that the fan mainly works in the CPU cooling scene under the wind volume of 70-100 m 3 / h, so the performance under 70, 80, 90 and 100 m 3 / h is selected as the basis for calculating the initial objective function, and the calculation formula of the objective function is: , where , , is the wind pressure, and the wind pressure under different wind volumes needs to be obtained through simulation calculation. According to the test and simulation data of the optimization starting point physical fan, the objective function is calculated as 1855, and the regular working speed of the fan is 2000 r / min, and the noise under the corresponding speed is 31.4 dB, which is the initial noise value of this optimization. In this optimization design, the expected value of the objective function is 30% larger than the initial objective function.

[0093] Then, the initial blade modeling is performed: in the initial blade modeling process, the blades are divided into 5 layers along the radial direction, and the difference value of the leading edge installation angle between adjacent two layers is denoted as , the difference of the blade trailing edge and leading edge installation angle of each layer is recorded as, the blade axial thickness is recorded as t, then according to the coordinate of each layer blade axial position and the corresponding relationship of the blade installation angle at the corresponding position, the parabolic function is used to fit the axial variation law of the blade installation angle. The initial parabolic function is determined: , where z is the axial position coordinate. At this time, only , and three parameters can be used to completely describe the blade shape (only the optimization of the blade installation angle is considered here), and the three parameters constitute the design parameter combination.

[0094] Then the initial design parameter combination and the initial target function value are output to the Bayesian optimizer for inference optimization to obtain the optimized design parameter combination, and then a new parabolic function of each layer blade is established. Specifically: based on the parabolic function of the present scheme, where z is the relative coordinate of the axial position, the value range is [0, 1], where 0 represents the blade leading edge and 1 represents the blade trailing edge.

[0095] is the blade leading edge installation angle, is the blade trailing edge installation angle, is equal to the opening angle between the blade trailing edge and the leading edge . The relative position coordinates corresponding to the blade leading edge and trailing edge are substituted into the parabolic function , so . Thus the relationship of is established, according to the value of recommended by the Bayesian optimizer, the value of can be calculated, at this time the value of is determined, that is, the new parabolic function is determined.

[0096] Then the new parabolic function of each layer blade is used to model and simulate the performance of the blade in the commercial CFD software, to obtain the wind pressure and noise data under four different flow rates, and to calculate the corresponding target function value of the optimized design parameter combination.

[0097] Then, noise evaluation is performed. This solution employs two evaluation strategies. In scenarios with high noise requirements, such as office environments, a rigid strategy is used, requiring that no noise value in the noise data exceeds the initial noise value. In scenarios with low noise requirements, a flexible strategy is used, allowing fans with larger objective function values ​​to generate appropriately increased noise. Specifically, if the objective function at a certain design point increases by x% compared to the objective function at the optimization starting point, then the fan noise at that design point is allowed to increase by 10*lg(1+x%) dB relative to the starting point noise. When the noise data does not meet the requirements of the rigid or flexible strategy, the objective function value is reset to zero and re-inputted into the Bayesian optimizer along with the initial design parameters for optimization. When the noise data meets the requirements of the rigid or flexible strategy, an objective function increment judgment is performed. If the objective function increment requirement is not met, the optimized design parameter combination and objective function value are input into the Bayesian optimizer for optimization. When the objective function increment requirement is met, the optimized design parameter combination is output, completing the optimization.

[0098] This application also provides an optimized design device for a cooling fan, such as... Figure 3 As shown, Figure 3 This is a structural block diagram of a cooling fan optimization design device provided in an embodiment of this application. The device includes: Module 301 is used to obtain the initial design parameter combination and the initial objective function value; Input module 302 is used to input the initial design parameter combination and the initial objective function value into the Bayesian optimizer to obtain the optimized design parameter combination; Calculation module 303 is used to calculate the optimized objective function value and fan noise data corresponding to the optimized design parameter combination based on the optimized design parameter combination; The judgment module 304 is used to determine whether the fan noise data and the optimized objective function value meet the preset conditions; The processing module 305 is configured to: when the fan noise data and the optimized target function value do not satisfy the preset condition, input the optimized design parameter combination and the optimized target function value as new initial design parameter combination and new initial target function value to the Bayesian optimizer to continue the optimization operation, or re-input the initial design parameter combination and the initial target function value to the Bayesian optimizer to continue the optimization operation; and when the fan noise data and the optimized target function value satisfy the preset condition, end the loop operation and output the optimized design parameter combination. It should be noted that the apparatus provided in the above embodiments is used to implement the functions thereof, and the above functions are only used as an example for the division of the functional modules, and in actual applications, the above functions can be completed by different functional modules according to the needs, that is, the internal structure of the apparatus is divided into different functional modules to complete all or part of the above functions. In addition, the apparatus and the method embodiments provided in the above embodiments belong to the same concept, and the specific implementation process is shown in the method embodiments, which will not be repeated here.

[0099] The application further provides a computer readable storage medium, which stores instructions, and the instructions are executed to perform the method steps of any one of the above.

[0100] In an example embodiment, the computer readable storage medium can include, but is not limited to, a U disk, a read-only memory (ROM), a random access memory (RAM), a mobile hard disk, a magnetic disk or an optical disk, and various computer program storage media.

[0101] The application further discloses an electronic device. As shown in Figure 4 Figure 4 is a structural schematic diagram of an electronic device disclosed by the embodiments of the application. The electronic device 400 can include at least one processor 401, at least one communication bus 402, a user interface 403, at least one network interface 404, and a memory 405.

[0102] The communication bus 402 is configured to realize the connection and communication between the components.

[0103] The user interface 403 can include a display screen (Display) and a camera (Camera), and the optional user interface 403 can further include a standard wired interface and a wireless interface.

[0104] The network interface 404 can optionally include a standard wired interface and a wireless interface (such as a WI-FI interface). ​

[0105] The processor 401 can include one or more processing cores. The processor 401 connects various parts within the entire electronic device (such as a server) by various interfaces and lines, executes various functions of the server and processes data by running or executing instructions, programs, code sets or instruction sets stored in the memory 405, and calling data stored in the memory 405. Alternatively, the processor 401 can be implemented in at least one of a hardware form of a digital signal processing (DSP), a field-programmable gate array (FPGA), and a programmable logic array (PLA). The processor 401 can integrate a combination of one or more of a central processing unit (CPU), a graphics processing unit (GPU), and a modem. Among them, the CPU mainly processes an operating system, a user interface, and an application program; the GPU is responsible for rendering and drawing the content to be displayed on the display screen; and the modem is used for processing wireless communication. It can be understood that the above-mentioned modem can also not be integrated into the processor 401, but can be implemented by a separate chip.

[0106] The memory 405 can include a random access memory (RAM) and can also include a read-only memory (ROM). Alternatively, the memory 405 includes a non-transitory computer-readable storage medium. The memory 405 can be used to store instructions, programs, codes, code sets or instruction sets. The memory 405 can include a program storage area and a data storage area, wherein the program storage area can store instructions for implementing an operating system, instructions for at least one function (such as a touch function, a sound playing function, an image playing function, etc.), instructions for implementing the above-mentioned various method embodiments, etc.; the data storage area can store data involved in the above-mentioned various method embodiments, etc. The memory 405 can also be at least one storage device located away from the aforementioned processor 401. Referring to Figure 4 The memory 405 as a computer storage medium can include an operating system, a network communication module, a user interface module, and an application program of the method for optimizing design of a cooling fan.

[0107] In Figure 4The electronic device 400 shown, the user interface 403 is mainly used for providing the interface for the user to input, obtaining the data input by the user; and the processor 401 can be used for calling the application program of the heat dissipation fan optimization design method stored in the memory 405, and when the application program is executed by one or more processors 401, the electronic device 400 executes the method described in one or more of the above embodiments. It should be noted that, for the foregoing method embodiments, in order to simply describe, they are all expressed as a series of action combinations, but those skilled in the art should know that the application is not limited to the action sequence described, because according to the application, certain steps can be performed in other sequences or simultaneously. Secondly, those skilled in the art should know that the embodiments described in the specification all belong to preferred embodiments, and the actions and modules involved are not necessarily required by the application.

[0108] In the above embodiments, the description of each embodiment has its own focus, and the parts not described in detail in a certain embodiment can be referred to the related description of other embodiments.

[0109] In several embodiments provided in the present application, it should be understood that the disclosed device or system can be implemented in other ways. For example, the device or system embodiments described above are only schematic. The division of units is only a logical function division. In actual implementation, additional division can be made, or a plurality of units or components can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the coupling or direct coupling or communication connection between the units shown or discussed can be indirect coupling or communication connection through some service interface, device or unit, and can be electrical or other forms.

[0110] The units described as separate components can or can not be physically separate, and the components shown as units can or can not be physical units, that is, they can be located in one place, or can be distributed on a plurality of network units. Part or all of the units can be selected according to actual needs to achieve the purpose of the embodiment.

[0111] In addition, each functional unit in each embodiment of the present application can be integrated in one processing unit, or each unit can exist physically, or two or more units can be integrated in one unit. The integrated unit can be realized in the form of hardware or in the form of software functional unit.

[0112] The integrated unit, if implemented in the form of a software function unit and sold or used as an independent product, can be stored in a computer readable memory. Based on such understanding, the technical solutions of the present application essentially or say the part that contributes to the prior art or the whole or part of the technical solutions can be embodied in the form of a software product. The computer software product is stored in a memory and includes a plurality of instructions for causing a computer device (which can be a personal computer, a server or a network device, etc.) to execute all or part of the steps of the embodiments of the present application. The aforementioned memory includes: a U disk, a mobile hard disk, a magnetic disk or an optical disk, and various media that can store program codes.

[0113] The above is only exemplary embodiments of the present disclosure, and cannot limit the scope of the present disclosure. That is, any equivalent changes and modifications made in accordance with the teachings of the present disclosure are still within the scope of the present disclosure. Other embodiments of the present disclosure will be readily apparent to those skilled in the art upon considering the disclosure.

[0114] The present application is intended to cover any variations, uses or adaptive changes of the present disclosure, which follow the general principles of the present disclosure and include common knowledge or conventional technical means in the technical field not recorded in the present disclosure.

Claims

1. A method for optimizing design of a heat dissipation fan, characterized by, The method comprises: obtaining an initial design parameter combination and an initial objective function value, wherein the initial design parameter combination represents a blade shape of an initial heat dissipation fan, the initial objective function value is a sum of products of a plurality of air volume points and air pressures corresponding to the plurality of air volume points, and the initial objective function value represents performance of the initial heat dissipation fan; performing an optimization operation on the initial design parameter combination in a loop to obtain an optimized design parameter combination and an optimized objective function value: inputting the initial design parameter combination and the initial objective function value into a Bayesian optimizer to obtain the optimized design parameter combination; based on the optimized design parameter combination, calculating the optimized objective function value and fan noise data corresponding to the optimized design parameter combination; determining whether the fan noise data and the optimized objective function value meet a preset condition; when the fan noise data and the optimized objective function value do not meet the preset condition, inputting the optimized design parameter combination and the optimized objective function value into the Bayesian optimizer as new initial design parameter combination and new initial objective function value to continue performing the optimization operation, or re-inputting the initial design parameter combination and a zero-value objective function value into the Bayesian optimizer to continue performing the optimization operation, wherein the zero-value objective function value is the optimized objective function value after being forced to zero; when the fan noise data and the optimized objective function value meet the preset condition, ending the loop operation and outputting the optimized design parameter combination.

2. The method of claim 1, wherein, The step of obtaining the initial design parameter combination specifically comprises: radially dividing blades of the initial heat dissipation fan into a plurality of layers, obtaining a first difference value of a leading edge installation angle between adjacent two layers of blades and a second difference value of a trailing edge installation angle of each layer of blades; constructing an initial parabolic function of the installation angle of each layer of blades according to a corresponding relationship between coordinates of an axial position of each layer of blades and the installation angle of the blades at the corresponding position, determining a first term coefficient of the initial parabolic function, and the initial parabolic function represents a change of the installation angle of each layer of blades along an axial direction of the initial heat dissipation fan; taking the first difference value, the second difference value and the first term coefficient of the parabolic function as the initial design parameter combination.

3. The method of claim 1, wherein, The step of obtaining the initial objective function value specifically comprises: selecting a plurality of working air volume points according to a working scenario of the initial heat dissipation fan; collecting a first air pressure corresponding to each working air volume point; calculating the initial objective function value according to the working air volume points and the first air pressure.

4. The method of claim 3, wherein, The step of calculating the optimized objective function value and fan noise data corresponding to the optimized design parameter combination based on the optimized design parameter combination specifically comprises: establishing a new parabolic function of each layer of blades according to the optimized design parameter combination; According to the new parabolic function, the fan is modeled and performance simulation is performed to obtain a second wind pressure corresponding to each of the working air volume points and a first noise value corresponding to each of the working air volume points, the fan noise data being a set of all the first noise values; According to each of the working air volume points and the corresponding second wind pressure, the optimized target function value is calculated.

5. The method of claim 4, wherein, When the fan noise data and the optimized target function value do not satisfy the preset condition, the optimized design parameter combination and the optimized target function value are input to the Bayesian optimizer as new initial design parameter combination and new initial target function value to continue the optimization operation step, specifically comprising: When each of the first noise values is less than a preset noise value and a target function increment is less than or equal to a preset threshold, the optimized design parameter combination and the optimized target function value are input to the Bayesian optimizer as new initial design parameter combination and new initial target function value to continue the optimization operation, the preset noise value being a maximum allowable noise value determined according to the initial heat dissipation fan working scene, and the target function increment being an increment of the optimized target function value relative to the initial target function value.

6. The method of claim 4, wherein, When the fan noise data and the optimized target function value do not satisfy the preset condition, the initial design parameter combination and a zero value target function value are re-input to the Bayesian optimizer to continue the optimization operation step, specifically comprising: When any one of the first noise values in the fan noise data is greater than or equal to a preset noise value, the optimized design parameter combination and a zero value target function value are re-input to the Bayesian optimizer to continue the optimization operation, the preset noise value being a maximum allowable noise value determined according to the initial heat dissipation fan working scene.

7. The method of claim 5, wherein, When the fan noise data and the optimized target function value satisfy the preset condition, the loop operation is ended and the optimized design parameter combination is output, specifically comprising: When each of the first noise values is less than a preset noise value and a target function increment is greater than a preset threshold, the loop operation is ended and the optimized design parameter combination is output, the preset noise value being a maximum allowable noise value determined according to the initial heat dissipation fan working scene, and the target function increment being an increment of the optimized target function value relative to the initial target function value.

8. A design optimization method based on a heat dissipation fan, characterized by, Comprise: An acquisition module is used to acquire an initial design parameter combination and an initial target function value; An input module is used to input the initial design parameter combination and the initial target function value to a Bayesian optimizer to obtain the optimized design parameter combination; A calculation module is used to calculate the optimized target function value and fan noise data corresponding to the optimized design parameter combination based on the optimized design parameter combination; A judgment module is used to judge whether the fan noise data and the optimized target function value satisfy a preset condition; a processing module configured to, when the fan noise data and the optimized objective function value do not satisfy the preset condition, input the optimized design parameter combination and the optimized objective function value as new initial design parameter combination and new initial objective function value to the Bayesian optimizer to continue the optimization operation, or input the initial design parameter combination and a zero value objective function value to the Bayesian optimizer to continue the optimization operation; when the fan noise data and the optimized objective function value satisfy the preset condition, ending the loop operation and outputting the optimized design parameter combination.

9. An electronic device comprising a memory and a processor, said memory having stored thereon a computer program, characterized in that, The processor implements the method of any one of claims 1-7 when executing the program.

10. A computer-readable storage medium, characterized in that, The computer readable storage medium stores instructions which, when executed, perform the method of any one of claims 1-7.

Images