Flywheel energy storage rotor topology optimization method and system

Abstract

The invention relates to the technical field of flywheel energy storage systems, in particular to a flywheel energy storage rotor topology optimization method and system. Finite element analysis software is used for conducting simulation analysis on the initial shape of the rotor, stress distribution data are obtained, a stress threshold value is set according to the stress distribution data, a small internal stress area of the rotor is recognized, and an optimizable area is marked; according to the method, void parameters are designed in an optimizable area based on an SDEA optimization SwinTransform algorithm, so that rotor structure optimization is realized; and performing performance verification on the optimized rotor structure to obtain optimization effect data. A unified coding mode is adopted to represent cavity parameters, a comprehensive fitness function containing structural strength, mass reduction and process realizability is established, and the effects of remarkably reducing rotor mass, remarkably improving stress distribution and remarkably improving limit rotating speed and energy storage density are achieved through balance optimization under multi-target constraint. The problems of large rotor mass, heavy magnetic bearing burden, stress concentration and the like in the traditional uniform disc design are effectively solved.

Description

Technical Field

[0001] The present invention relates to the technical field of flywheel energy storage systems, and in particular to a flywheel energy storage rotor topology optimization method and system. Background Art

[0002] At present, the rotor of the flywheel energy storage system usually adopts a uniform disc design. This design scheme is mature in technology and simple in process, but it has many limitations, such as: The rotor mass is large: The uniform design increases the overall mass of the flywheel rotor, which limits the energy storage density of the system and makes it difficult to meet some application scenarios with high requirements for energy storage density. The magnetic bearing is burdened: The larger rotor mass increases the bearing pressure of the magnetic bearing, which requires higher bearing performance, thereby increasing the complexity of system design and maintenance, increasing costs and being unfavorable to the stable operation of the system. Stress concentration: When rotating at high speed, the uniform disc design causes stress to be concentrated in a specific area of ​​the rotor, resulting in a large maximum stress on the rotor, increasing the risk of failure, and severely limiting the maximum speed of the system, hindering the further improvement of the performance of the flywheel energy storage system. In summary, these factors of the existing uniform disc design jointly restrict the performance improvement of the flywheel energy storage system in terms of high speed, high energy storage density, and manufacturing and maintenance costs. It is necessary to propose a new rotor design scheme to overcome these limitations. Summary of the invention

[0003] In view of the problems existing in the prior art, the present invention is proposed.

[0004] Therefore, the problem to be solved by the present invention is how to make the rotor mass larger: the uniform design increases the overall mass of the flywheel rotor, which limits the energy storage density of the system and makes it difficult to meet some application scenarios with high requirements for energy storage density. Heavy burden on magnetic bearings: The larger rotor mass increases the bearing pressure of the magnetic bearings, and correspondingly requires higher bearing performance, which in turn increases the complexity of system design and maintenance, increases costs and is not conducive to stable operation of the system. Stress concentration: When rotating at high speed, the uniform disc design causes stress to be concentrated in a specific area of ​​the rotor, resulting in a larger maximum stress on the rotor, increasing the risk of failure, and severely limiting the maximum speed of the system, hindering further improvement in the performance of the flywheel energy storage system. In summary, these factors of the existing uniform disc design jointly restrict the performance improvement of the flywheel energy storage system in terms of high speed, high energy storage density, and manufacturing and maintenance costs. It is necessary to propose a new rotor design scheme to overcome these limitations.

[0005] In order to solve the above technical problems, the present invention provides the following technical solutions:

[0006] In a first aspect, an embodiment of the present invention provides a method for topological optimization of a flywheel energy storage rotor, which includes using finite element analysis software to simulate and analyze the initial shape of the rotor to obtain rotor stress distribution data;

[0007] Based on the stress distribution data, identifying areas with lower stress in the rotor, and setting stress thresholds to mark areas that can be optimized;

[0008] In the optimizable region, cavity parameters are designed based on SDEA optimization of the SwinTransformer algorithm to achieve rotor structure optimization;

[0009] The performance of the optimized rotor structure is verified to obtain optimization effect data.

[0010] As a preferred solution of the flywheel energy storage rotor topology optimization method described in the present invention, the simulation analysis of the initial shape of the rotor using finite element analysis software includes: setting rotor material parameters, including density and Poisson's ratio; establishing a rotor geometric model and performing meshing; applying axial and radial displacement constraints; setting rotation speed operating conditions; and calculating the stress distribution in each region of the rotor under the set operating conditions.

[0011] As a preferred solution of the flywheel energy storage rotor topology optimization method described in the present invention, the identification of the region with low stress in the rotor includes: setting a stress threshold σ0; extracting the radial stress σ of each node of the rotor r and the tangential stress σ θ ; Calculate the node comprehensive stress σ i =max(σ r,i ,σ θ,i );When the node comprehensive stress σ i When it is less than the stress threshold σ0, the area where the node is located is marked as an optimizable area.

[0012] As a preferred solution of the flywheel energy storage rotor topology optimization method described in the present invention, the design of void parameters based on SDEA optimization of SwinTransformer algorithm includes: using a unified coding method to represent void parameters; establishing a comprehensive fitness function including structural strength, mass reduction and process feasibility; using the SwinTransformer model to predict the void fitness value; and optimizing and iterating the model parameters through the SDEA algorithm.

[0013] As a preferred solution of the flywheel energy storage rotor topology optimization method described in the present invention, the unified encoding method of the cavity parameters includes: a circular cavity is represented by a vector [x, y, r, 0], where (x, y) is the center coordinate, r is the radius, and 0 represents a circle; a square cavity is represented by a vector [x, y, a, θ, 1], where (x, y) is the center coordinate, a is the side length, θ is the rotation angle, and 1 represents a square.

[0014] As a preferred solution of the flywheel energy storage rotor topology optimization method described in the present invention, the comprehensive fitness function includes: a structural strength fitness function f1, which is used to evaluate the stress distribution under the void configuration; a mass reduction fitness function f2, which is used to evaluate the material removal effect; a process feasibility fitness function f3, which is used to evaluate the feasibility of void parameters; and a comprehensive fitness function F=f1×f2×f3.

[0015] As a preferred solution of the flywheel energy storage rotor topology optimization method described in the present invention, the optimization iteration process includes: evaluating individuals in the population according to fitness values; calculating the social influence of individuals on other individuals; updating echolocation; checking whether the updated position is within a reasonable range; repeating iterations until the maximum number of iterations is reached;

[0016] The optimization effect data include: mass reduction amount and reduction percentage; stress distribution change and reduction percentage; limit speed increase value and increase percentage; energy storage density increase value and increase percentage.

[0017] In a second aspect, an embodiment of the present invention provides a flywheel energy storage rotor topology optimization system, which includes a stress analysis module, which uses finite element analysis software to simulate and analyze the initial shape of the rotor to obtain rotor stress distribution data;

[0018] A region identification module, which identifies regions with lower stress in the rotor according to the stress distribution data, and sets stress thresholds to mark regions that can be optimized;

[0019] An optimization design module, which optimizes the SwinTransformer algorithm based on SDEA to design cavity parameters in the optimizable area;

[0020] The performance verification module verifies the performance of the optimized rotor structure and obtains optimization effect data.

[0021] In a third aspect, an embodiment of the present invention provides a computer device, including a memory and a processor, wherein the memory stores a computer program, wherein: when the computer program instructions are executed by the processor, the steps of a flywheel energy storage rotor topology optimization method as described in the first aspect of the present invention are implemented.

[0022] In a fourth aspect, an embodiment of the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein: when the computer program instructions are executed by a processor, the steps of a flywheel energy storage rotor topology optimization method as described in the first aspect of the present invention are implemented.

[0023] The beneficial effects of the present invention are as follows: the present invention can reasonably design the cavity under the conditions of satisfying the structural strength and process feasibility. The total volume of material removal can be calculated by using the relevant algorithm, and then the amount of rotor mass reduction can be obtained, thereby achieving the effect of reducing the rotor mass, which has positive significance for improving the efficiency of the entire system, reducing energy consumption, etc.

[0024] During the cavity configuration process, the maximum stress value of the rotor under the current cavity configuration is calculated through structural strength assessment, and the stress condition is judged based on the structural strength fitness function. The optimized rotor can effectively improve stress distribution and reduce stress concentration. Compared with traditional designs, it can reduce the risk of rotor failure caused by stress concentration, enable the rotor to run stably at a higher speed, and improve the performance and reliability of the rotor.

[0025] When designing voids, the process feasibility evaluation function is used to determine whether each void parameter is within the size, shape and position range allowed by the process, ensuring the feasibility of the design in the actual manufacturing process, avoiding the problem of design being unfeasible due to process limitations, and improving the practicality and manufacturability of the design.

[0026] The present invention comprehensively considers multiple factors such as structural strength requirements, process feasibility and mass reduction, and achieves balanced optimization through a comprehensive fitness function. When designing the cavity, a unified coding method is used to represent the cavity parameters, and an advanced algorithm (socially guided dolphin echolocation algorithm optimization SwinTransformer) is used to optimize the cavity selection and calculation, which improves the flexibility and accuracy of the design. This comprehensive optimization method can ensure that the designed cavity configuration meets various practical requirements while achieving the best performance improvement effect, and the algorithm and parameters can be adjusted according to the actual situation to obtain a better design solution. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0028] Figure 1 A flow chart of the flywheel energy storage rotor topology optimization method;

[0029] Figure 2 A diagram of the computer equipment for the flywheel energy storage rotor topology optimization method; DETAILED DESCRIPTION

[0030] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the accompanying drawings.

[0031] In the following description, many specific details are set forth to facilitate a full understanding of the present invention, but the present invention may also be implemented in other ways different from those described herein, and those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0032] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The term "in one embodiment" that appears in different places in this specification does not necessarily refer to the same embodiment, nor is it a separate or selective embodiment that is mutually exclusive with other embodiments.

[0033] Example 1

[0034] Reference Figure 1-2 , which is the first embodiment of the present invention, and provides a flywheel energy storage rotor topology optimization method, comprising:

[0035] S100: Use finite element analysis software to simulate and analyze the initial shape of the rotor and obtain rotor stress distribution data;

[0036] In the embodiment of the present application, the specific steps of using ANSYS software for simulation analysis include:

[0037] Geometric modeling: Use the DesignModeler module in ANSYS Workbench to create a 3D model of the rotor. First, establish the basic shape of the disk based on the given parameters (diameter D, thickness H), and then add detailed features such as the center axis hole. When modeling, pay attention to maintaining the symmetry of the geometric structure and avoiding unnecessary geometric singularities.

[0038] Material property definition: Define the high-strength steel material properties in the Engineering Data module. Usually the following parameters need to be entered: Density (ρ): 7850kg / m 3 ; Elastic modulus (E): 210GPa; Poisson's ratio (ν): 0.3; Yield strength: 800MPa; Tensile strength: 950MPa.

[0039] Meshing: Meshing is performed in the Mechanical module, including: selecting the hexahedral element type (SOLID186); setting the global mesh size to H / 20 (H is the rotor thickness); locally refining the mesh in the stress concentration area; performing a mesh quality check to ensure that the minimum element quality is >0.7; and controlling the total number of elements within the range of 50,000 to 100,000.

[0040] In an optional embodiment, the meshing strategy can also use adaptive meshing technology. The system first uses a coarser mesh for preliminary analysis, and then automatically refines the mesh in the high stress area according to the stress gradient until the convergence accuracy requirement is met. This method can reduce the consumption of computing resources while ensuring the calculation accuracy.

[0041] It should be noted that meshing directly affects the calculation accuracy and efficiency. A mesh that is too dense will lead to a sharp increase in the amount of calculation, while a mesh that is too sparse may cause inaccurate results. Therefore, it is necessary to verify the independence of the mesh, that is, to gradually increase the mesh density and observe whether the result changes converge, so as to determine the appropriate mesh density.

[0042] Boundary condition settings: Axial constraint: apply axial displacement constraint on the inner surface of the shaft hole; Radial constraint: apply radial displacement constraint at key nodes; Load condition: apply rotation speed ω (usually 10000rpm).

[0043] In an optional embodiment, the boundary conditions can also consider the influence of temperature load. Since high-speed rotation will generate heat, temperature field analysis can be added to the model, and the temperature load can be set as follows: ambient temperature: 20°C; operating temperature: 20-80°C linear distribution; heat transfer coefficient: 15W / (m 2 〃K).

[0044] S101: Use finite element analysis software to simulate and analyze the initial shape of the rotor, including: setting rotor material parameters, including density and Poisson's ratio; establishing a rotor geometry model and performing meshing; applying axial and radial displacement constraints; setting rotation speed operating conditions; and calculating the stress distribution in each region of the rotor under the set operating conditions.

[0045] Specifically, finite element analysis includes four main steps: material property setting, geometric modeling, meshing and stress calculation. Material property setting mainly includes density ρ = 7850kg / m 3, elastic modulus E = 210GPa, Poisson's ratio ν = 0.3 and other basic parameters. Geometric modeling mainly includes the solid modeling of a disk with a rotor outer diameter of D = 1m and a thickness of H = 0.2m. The meshing mainly uses hexahedral units with a mesh size of H / 20, and the mesh is encrypted in the stress concentration area. Stress calculation includes calculating the radial stress σr and tangential stress σθ distribution under the set speed ω = 10000rpm.

[0046] In an optional embodiment, the finite element analysis process can also consider the influence of temperature load. By introducing a temperature field distribution of 20-80°C in the model and setting the heat transfer coefficient to 15W / (m 2 〃K), simulating the thermal-stress coupling effect of the rotor during high-speed rotation. In addition, the anisotropic properties of the material can be considered, and different elastic moduli and strength parameters can be set in different directions to more accurately reflect the actual working conditions.

[0047] In an optional embodiment, meshing can also use adaptive meshing technology. First, a coarser mesh is used for preliminary calculations, and then the local mesh density is automatically adjusted according to the stress gradient. The mesh density is increased in the stress concentration area and appropriately increased in the stress gentle area, which ensures both the calculation accuracy and the calculation efficiency. The system iterates the calculation and considers that the meshing reaches the convergence standard when the maximum difference between two adjacent calculation results is less than 1%.

[0048] It should be noted that finite element analysis is the basis for subsequent optimization design, and its accuracy directly affects the reliability of the optimization results. The degree of simplification of the model should be moderate, reflecting the main force characteristics while avoiding excessive complexity. Mesh quality control is the key to ensuring calculation accuracy, and mesh independence verification is required to ensure the reliability of the results. In addition, when setting boundary conditions, it is necessary to fully consider the actual installation and use status to ensure that the calculation results have practical engineering significance.

[0049] S200: Identify the area with less stress in the rotor based on the stress distribution data, and set the stress threshold to mark the area that can be optimized;

[0050] Based on the stress distribution data, identify the areas with less stress in the rotor and set stress thresholds to mark the areas that can be optimized;

[0051] In the embodiment of the present application, the stress region identification process includes the following specific steps: stress data extraction: exporting node stress data from the ANSYS result file; establishing a stress data matrix including node coordinates and corresponding radial stress σ r , tangential stress σ θ ; Data formatting to ensure data continuity and integrity.

[0052] Stress threshold determination: The reference stress threshold σ0 is set to 100 MPa; the safety factor K = 1.5 is considered; the actual stress threshold is calculated as:

[0053] σ t resold=σ0 / K

[0054] Specifically, stress data processing includes three steps: stress extraction, data filtering, and area marking. Stress extraction mainly obtains the radial stress σ of each node. r and the tangential stress σ θ value; data filtering is mainly to remove abnormal data points and the comprehensive stress value of the calculation node; the region marking is to mark the region with stress value lower than the threshold value σ0=100MPa as the potential optimization region. The comprehensive stress is determined by taking the maximum value of radial stress and tangential stress, that is, σ i =max(σ r,i ,σ θ,i ).

[0055] In an optional embodiment, the stress threshold can be determined by an adaptive method. By calculating the average stress level and stress standard deviation of the entire rotor, the threshold is set to the average stress minus 0.5 times the standard deviation, but at the same time the minimum threshold is set to 50 MPa to ensure the rationality of the optimization area. This method can better adapt to the stress distribution characteristics under different working conditions.

[0056] In an optional embodiment, the region marking process can also consider the influence of stress gradient. When marking potential optimization regions, not only the absolute value of stress is considered, but also the severity of stress change. By calculating the stress gradient between adjacent nodes, the region with smaller gradient is preferentially marked as the potential optimization region, so as to avoid optimization in the stress sudden change region.

[0057] It should be noted that the region identification process needs to fully consider the integrity and local characteristics of the structure. Since the rotor is an integral structure, the local stress distribution will affect each other, so the continuity and integrity of the region should be considered when identifying the low stress area. In addition, considering the actual processing and manufacturing needs, the identified area cannot be too scattered or too small, and appropriate regional merging and shape optimization are required to ensure the feasibility of subsequent optimization design.

[0058] S201: Identifying the region with low stress in the rotor includes: setting a stress threshold σ0; extracting the radial stress σ of each node of the rotor r and the tangential stress σ θ ; Calculate the node comprehensive stress σ i =max(σ r,i ,σ θ,i );When the node comprehensive stress σ iWhen it is less than the stress threshold σ0, the area where the node is located is marked as an optimizable area.

[0059] S300: Design cavity parameters based on SDEA optimization of SwinTransformer algorithm in the optimizable area to achieve rotor structure optimization;

[0060] S301: Designing void parameters based on SDEA-optimized SwinTransformer algorithm includes: using a unified coding method to represent void parameters; establishing a comprehensive fitness function that includes structural strength, mass reduction, and process feasibility; using the SwinTransformer model to predict the void fitness value; and optimizing and iterating the model parameters through the SDEA algorithm.

[0061] S302: The unified encoding method of the hole parameters includes: a circular hole is represented by a vector [x, y, r, 0], where (x, y) is the coordinate of the center of the circle, r is the radius, and 0 represents a circle; a square hole is represented by a vector [x, y, a, θ, 1], where (x, y) is the center coordinate, a is the side length, θ is the rotation angle, and 1 represents a square.

[0062] S303: The comprehensive fitness function includes: structural strength fitness function f1, which is used to evaluate the stress distribution under the void configuration; mass reduction fitness function f2, which is used to evaluate the material removal effect; process feasibility fitness function f3, which is used to evaluate the feasibility of void parameters; comprehensive fitness function F = f1×f2×f3.

[0063] Specifically, the hole parameter design includes three steps: parameter encoding, prediction model construction, and optimization iteration. The parameter encoding adopts a unified format. The circular hole is represented by [x, y, r, 0], where (x, y) is the coordinate of the center of the circle, r is the radius, and 0 represents a circle; the square hole is represented by [x, y, a, θ, 1], where (x, y) is the center coordinate, a is the side length, θ is the rotation angle, and 1 represents a square. The prediction model adopts the SwinTransformer architecture, which includes an input layer, an encoding layer, an attention layer, and an output layer. The optimization iteration adopts the SDEA algorithm, which includes two stages: social influence update and echolocation update.

[0064] In an optional embodiment, the cavity shape can also be an ellipse or a polygon. An elliptical cavity is represented by [x, y, a, b, θ, 2], where a and b are the lengths of the major and minor axes, respectively; a regular polygon is represented by [x, y, r, n, θ, 3], where r is the radius of the circumscribed circle and n is the number of sides. By increasing the diversity of cavity shapes, a more flexible optimization scheme can be obtained to better adapt to different stress distribution characteristics. Each shape is equipped with corresponding constraints to ensure that the generated cavity meets the process and strength requirements.

[0065] In an optional embodiment, the model training process can also adopt a transfer learning strategy. First, the model is pre-trained on a general structure optimization dataset, and then fine-tuned using specific data of the flywheel rotor. 100,000 general structure samples are used in the pre-training phase, and 1,000 flywheel rotor samples are used in the fine-tuning phase. This method can significantly improve the convergence speed and optimization effect of the model.

[0066] In an optional embodiment, the social influence parameter α and echolocation parameter β of the SDEA algorithm can be adjusted dynamically. In the early stage of optimization, a larger echolocation parameter β is set to enhance the global search capability; as the iteration proceeds, the social influence parameter α is gradually increased to enhance local optimization. For example, α = 0.3, β = 0.7 are initially set, and adjusted every 100 iterations until α = 0.7, β = 0.3.

[0067] It should be noted that cavity parameter optimization is a multi-objective optimization problem, which requires considering multiple objectives such as structural strength, mass reduction and process feasibility at the same time. During the optimization process, attention should be paid to maintaining the symmetry of the structure to avoid generating unreasonable cavity distribution. At the same time, since the rotor is a high-speed rotating component, the optimization result must ensure dynamic balance. In addition, the minimum spacing between cavities and the minimum distance between edges must meet the process requirements, usually not less than 2 times the processing accuracy.

[0068] S400: Verify the performance of the optimized rotor structure and obtain optimization effect data.

[0069] S401: The optimization iteration process includes: evaluating individuals in the population according to fitness values; calculating the social influence of individuals on other individuals; updating echolocation; checking whether the updated position is within a reasonable range; repeating iterations until the maximum number of iterations is reached;

[0070] The optimization effect data include: mass reduction amount and reduction percentage; stress distribution change and reduction percentage; limit speed increase value and increase percentage; energy storage density increase value and increase percentage.

[0071] Specifically, performance verification includes four steps: stress verification, mass calculation, speed analysis, and energy storage evaluation. Stress verification mainly checks the maximum stress and uniformity of stress distribution to ensure that the maximum stress does not exceed 70% of the allowable stress of the material. Mass calculation includes statistics on material weight reduction and weight reduction percentage, requiring the overall weight reduction to be no less than 20%. Speed ​​analysis mainly calculates the critical speed and the maximum operating speed to ensure that the operating speed is far away from the critical speeds of each order. Energy storage evaluation includes the calculation of energy storage density per unit mass and prototype comparison.

[0072] In an optional embodiment, the performance verification may also include fatigue analysis. The fatigue life of the rotor is evaluated using the stress-life method (SN curve), and the material SN curve data comes from experimental tests. The cumulative damage value is calculated using the Miner linear cumulative damage theory, and the service life of the rotor under various operating conditions is estimated. The design life is usually required to be no less than 10 years or 100,000 start-stop cycles.

[0073] In an optional embodiment, dynamic characteristics analysis can also be performed. This includes modal analysis and transient dynamics analysis, calculating the rotor's natural frequency and vibration mode, and evaluating the vibration response at different speeds. Special attention is paid to the resonance risk within the operating speed range to ensure that the vibration amplitude meets the ISO 1940 standard requirements, which usually requires the dynamic unbalance to be lower than level G2.5.

[0074] In an optional embodiment, the performance verification can be extended to extreme operating condition analysis, including overspeed test (1.2 times rated speed), emergency stop condition and temperature cycle condition, etc. The stress level and deformation of the rotor under these extreme conditions are evaluated by finite element simulation to ensure sufficient safety margin.

[0075] It should be noted that performance verification is the final test of the optimized design and requires a comprehensive evaluation from multiple dimensions. The verification process should focus on the reliability and integrity of the data, and key performance indicators should have clear quantitative requirements. The evaluation of the optimization effect should be combined with actual application needs, and should not only meet technical indicators but also consider economic efficiency. In addition, the verification results should form a complete technical document to provide a basis for subsequent engineering applications and continuous optimization.

[0076] Furthermore, this embodiment also provides a flywheel energy storage rotor topology optimization system, comprising:

[0077] Stress analysis module, using finite element analysis software to simulate and analyze the initial shape of the rotor to obtain rotor stress distribution data;

[0078] A region identification module, which identifies regions with lower stress in the rotor according to the stress distribution data, and sets stress thresholds to mark regions that can be optimized;

[0079] An optimization design module, which optimizes the SwinTransformer algorithm based on SDEA to design cavity parameters in the optimizable area;

[0080] The performance verification module verifies the performance of the optimized rotor structure and obtains optimization effect data.

[0081] In summary, by using finite element analysis software for fine meshing and multi-condition simulation analysis, high-precision calculation of the rotor's global stress distribution is achieved. This provides a more accurate stress field distribution feature and provides a reliable data basis for subsequent optimization. This method avoids the blindness of stress prediction in traditional empirical design and makes the cavity layout more targeted.

[0082] By setting a reasonable stress threshold standard and combining the comprehensive judgment mechanism of radial stress and tangential stress, the automatic identification of the optimizable area is realized. This method overcomes the subjectivity and uncertainty of traditional manual experience judgment and improves the accuracy and efficiency of area identification. In particular, the continuity and machinability of the optimized area are guaranteed through the area merging algorithm.

[0083] By combining deep learning with swarm intelligence optimization, the void parameter optimization problem is innovatively solved. This method can not only adaptively learn the optimization strategy under different working conditions, but also avoid local optimality through social guidance mechanism, achieving a better solution than traditional optimization methods. This is an unexpected technical effect.

[0084] By establishing a comprehensive fitness function that includes structural strength, mass reduction, and process feasibility, a balanced optimization under multi-objective constraints is achieved. This method overcomes the limitations of traditional single-objective optimization and maximizes the goal of mass reduction while ensuring structural safety.

[0085] By setting up a strict performance verification system, including stress verification, mass calculation, speed analysis and energy storage evaluation, the reliability of the optimization results is guaranteed. In particular, through dynamic characteristics analysis and extreme working condition verification, the safety and stability of the optimization results under various working conditions are ensured.

[0086] Example 2

[0087] Reference Figure 1 - Figure 2 , which is the second embodiment of the present invention, and this embodiment provides a flywheel energy storage rotor topology optimization method. In order to verify the beneficial effects of the present invention, scientific demonstration is carried out through economic benefit calculation and simulation experiments.

[0088] The professional finite element analysis software ANSYS is used to calculate the initial shape of the proposed rotor. High-strength steel material is selected, with a diameter of D meters and a thickness of H meters for simulation analysis. During the simulation, the rotor's rotation speed and other working conditions are set, and the speed is set to ω revolutions per minute. The software calculates the stress distribution of each area of ​​the rotor under this working condition. The stress calculation formula is based on the basic stress calculation formula in elastic mechanics, and the calculation accurately presents the stress numerical distribution of each area of ​​the entire rotor.

[0089] In ANSYS software, after creating a new analysis project, first enter the material property setting module. For the selected high-strength steel material, enter its density ρ, Poisson's ratio ν and other material parameters. These parameters will serve as the basic data for subsequent stress calculations.

[0090] Create a disc-shaped rotor geometry model and set the dimensions according to the given diameter D meters and thickness H meters. Use the ANSYS Mechanical interface to draw a disc solid model that meets the requirements using operations such as stretching and rotation.

[0091] Mesh the created geometric model. Discretize the continuous geometric entity into a finite set of units and nodes, select the appropriate meshing method - hexahedral mesh, and set the mesh size parameters according to the complexity of the model and the calculation accuracy requirements. Use the default medium mesh density for preliminary division, and then gradually refine the mesh to observe the changes in the calculation results. When the results tend to be stable, it means that the mesh density has reached the convergence requirement.

[0092] According to the actual installation and working conditions of the rotor, reasonable constraints are imposed on the rotor model. The rotor is installed on the supporting structures at both ends through the shaft. Axial and radial displacement constraints are imposed at the position corresponding to the rotor center hole (where it contacts the actual installation shaft) to limit its movement in these directions and simulate the actual fixed support state. This is achieved by selecting the corresponding nodes or faces in the constraint definition interface of the ANSYS software and then using the fixed constraint Fixed Support.

[0093] Set the rotor's rotation speed condition. Set the speed to ω rpm. ANSYS simulates the rotor's rotation state by applying a rotational angular velocity load. Select the entire rotor model, add a rotational angular velocity load in the load definition interface, and set its value to the radians per second corresponding to ω (radians per second = ω × 2π / 60).

[0094] ANSYS calculates each unit and node based on the set finite element model, material properties, boundary conditions and load conditions, and gradually iterates to solve the stress distribution of each area of ​​the entire rotor under the set working conditions. The software will calculate the radial stress σ according to the input ρ (density), ν (Poisson's ratio), ω (speed) and r (radius) value corresponding to each node, combined with constants A and C. r and the tangential stress σ θ The calculation formula is used to calculate the stress value at each node.

[0095] Determine constants A and C. The constants A and C in the stress calculation formula for a disk rotating under a plane stress state need to be determined based on specific boundary conditions. Common boundary conditions include stress conditions at the inner boundary (center hole) and outer boundary (outer diameter edge of the disk). For example, if the radial stress σ at the center hole of the disk (radius r = r0, assuming the center hole radius is r0 meters) is known r The value of (r0) r is a certain value (for example, the stress boundary value at that position obtained based on experience or actual measurement), and the tangential stress σ at the edge of the outer diameter of the disk (radius r = D / 2) θ The value of (D / 2) is another known boundary value. Substituting these two boundary conditions into the above stress calculation formula, we can get a set of two-variable linear equations about A and C:

[0096]

[0097] By solving this set of equations, the specific values ​​of constants A and C can be determined, and accurate parameter values ​​can be used in the entire stress calculation process.

[0098] Analyze radial stress σ r and the tangential stress σ θ The cloud map will use different colors to intuitively represent the stress distribution. The darker the color (or according to the software's default color scale, such as from blue to red to represent stress from small to large, etc.), the higher the stress value. In this way, you can clearly see the stress concentration area and the stress value range of each part of the entire rotor, providing data basis for identifying areas with lower stress in subsequent topology optimization design.

[0099] Based on the above stress analysis results, the areas with lower stress in the rotor are identified. These areas are potential areas where holes can be arranged for optimization. A stress threshold σ0 (in Pascals) is set. When the calculated stress value σ of a certain area is lower than σ0, the area is marked as an optimizable area.

[0100] The stress threshold σ0 is set based on the rotor material properties of high-strength steel. The initial value is set to σ0 = 100 × 10 6 .

[0101] Check the stress cloud map. Based on the stress distribution represented by the color on the cloud map, first determine the range of the stress value region intuitively, and extract the radial stress σ corresponding to each node of the rotor. r and the tangential stress σ θ Numeric value.

[0102] For the extracted stress data of each node, point-by-point comparison and area marking are performed in the following manner.

[0103] Traverse the stress values ​​corresponding to all nodes, consider radial stress and tangential stress at the same time, and take the larger value of the two to comprehensively judge the stress level at the node. Suppose the stress value of a node is σ i , whose radial stress is σ r,i , the tangential stress is σ θ,i , then the node comprehensive stress σ i It can be determined as follows:

[0104] σ i =max(σ r,i ,σ θ,i )

[0105] The comprehensive stress value σ of each node i is compared with the set stress threshold σ0. If σ i <σ0, the small area where the node is located is marked as a potential optimizable area. Continue to perform such judgment and marking operations on all nodes, and finally mark the areas where all nodes that meet the condition that the stress value is lower than the threshold are located.

[0106] After point-by-point marking, multiple scattered small areas marked as potentially optimizable are obtained. These small areas need to be integrated to determine the final large potential optimizable area that is continuous and convenient for subsequent hole design.

[0107] Check the adjacent relationship between each marked small area. If two small areas are adjacent in space and their stress levels are both below the threshold, the adjacency can be determined by judging the coordinate continuity of the regional boundary nodes and other geometric relationships. For example, if two areas share some boundary nodes, they can be considered as adjacent areas and merged. Repeat the operation of merging adjacent areas until they can no longer be merged, and finally form several continuous, relatively large areas with low stress. These areas are potential areas for the subsequent topology optimization design that can be considered to arrange holes.

[0108] In the identified areas with lower stress, the shape, size, distribution position and other parameters of the voids are reasonably designed according to comprehensive factors such as structural strength requirements and process feasibility. The socially guided dolphin echolocation algorithm (SDEA) is used to optimize the combined optimization algorithm of SwinTransformer key parameters to achieve optimal selection and calculation of voids.

[0109] From the stress analysis and region identification steps, the geometric information, boundary conditions, and material properties (density ρ) of the region with low stress are obtained. The parameters related to the structural strength requirements are determined, including the maximum allowable stress value σ allow and the ultimate stress value σ limit, process feasibility parameters, including minimum processing size and processing accuracy.

[0110] Hole representation and encoding.

[0111] A unified encoding method is used for the shape, size and position of the hole. A circular hole is represented by a vector [x, y, r, 0], where (x, y) is the coordinate of the center of the circle, r is the radius, and the last digit 0 indicates a circle. A square hole is represented by a vector [x, y, a, θ, 1], where (x, y) is the center coordinate, a is the side length, θ is the rotation angle, and the last digit 1 indicates a square. For polygonal holes, a vertex coordinate sequence and an identifier indicating the polygon type can be used to encode. Convert the encoded information of the hole into a tensor form suitable for processing by SwinTransformer. Discretize all possible hole parameter ranges into fixed-length vectors, and then combine these vectors into batch tensors as input to SwinTransformer.

[0112] Formulate a comprehensive fitness function.

[0113] Structural strength assessment. Calculate the stress distribution of the rotor under the current cavity configuration. For a given cavity parameter, obtain the maximum stress value σ of the rotor. The structural strength fitness function f1 is defined as:

[0114] max

[0115]

[0116] Mass reduction assessment. Calculate the total volume V of material removed for the current cavity configuration 去 . Let the area of ​​a single hole be S i , there are n cavities in total, and the rotor thickness is H, then Rotor mass reduction Δm=ρV 去 .

[0117] -The mass reduction fitness function f2 is defined as:

[0118]

[0119] Where Δm target is the desired mass reduction target value.

[0120] Process feasibility evaluation. Check whether the current cavity configuration meets the process feasibility conditions. For each cavity parameter, determine whether it is within the size, shape and position range allowed by the process. The process feasibility fitness function f3 is defined as:

[0121]

[0122] Comprehensive fitness function. The comprehensive fitness function F is:

[0123] F=f1×f2×f3

[0124] SwinTransformer implements prediction. All individual hole parameter encoding vectors are combined into a total dataset D, whose size is M (i.e., it contains M different hole configuration encoding samples). The dataset D is divided into training set D according to 80% for training, 10% for validation, and 10% for testing. train , validation set D val and the test set D test . Assume that the number of training set samples is M train , the number of samples in the validation set is M val , the number of test set samples is M test , satisfying M train +M val +M test =M.

[0125] Each sample (i.e., the individual hole parameter encoding vector) is normalized to a distribution with a mean of 0 and a standard deviation of 1. For each element x in the input encoding vector ij (represents the jth eigenvalue of the i-th sample), calculate its mean μ j and standard deviation σ j :

[0126]

[0127] Then normalize each element:

[0128]

[0129] Build the SwinTransformer architecture.

[0130] The input two-dimensional code vector is divided into non-overlapping patches. Assume that the size of each patch is P×P. After the patch division, the dimension of the input feature map will change from [patches size ,num features ] becomes [batc size ,num patces ,C], where num patces is the number of small blocks after division, and the calculation formula is C is the embedding dimension of each small block after linear mapping.

[0131] Among them, batc size is the batch size, which can be the number of samples selected for each training, num patces is the number of features after hole parameter encoding

[0132] The block merging operation is performed to change the resolution and number of channels of the feature map to gradually extract higher-level features. Each block merging will merge adjacent 2×2 small blocks into a large block, and the number of channels will double. The feature map dimension of the current layer [batc size ,num patces ,C], after block merging, the new feature map dimension becomes By analogy, such operations are performed between different layers to construct feature representations of different resolutions.

[0133] Set the window size to W×W, divide the feature map into multiple windows, and apply the multi-head attention mechanism in each window. For the multi-head attention mechanism, there are H heads, and the calculation process of each head is as follows:

[0134] (1) First, perform a linear transformation on the input feature vector to obtain the Query vector Q, Key vector K, and Value vector V. Assume that the input feature dimension is [batc size ,num patces ,C], through three different linear transformation matrices:

[0135] W Q ∈R C×C / H , W K ∈R C×C / H , W V ∈R C×C / H get:

[0136] Q=XW Q ,K=XW K ,V=XW V

[0137] Where X is the input feature, and the dimensions of the output Q, K, and V are:

[0138] [batc size ,num patces ,C / H].

[0139] (2) Calculate the dot product between Query and Key to obtain the attention score A, and then perform scaling and normalization.

[0140]

[0141] The attention score is then normalized by the Softmax function:

[0142] A softmax =Softmax(A)

[0143] (3) Multiply the normalized attention score by the Value vector to obtain the attention output of each head.

[0144] O=A softmax V

[0145] Finally, the attention outputs of all heads are concatenated and a linear transformation is performed to obtain the final window attention output, the dimension is still [batc size ,num patces ,C].

[0146] (4) Add a feedforward network after each attention layer, which includes two fully connected layers with an activation function ReLU in the middle to further perform nonlinear transformation on the features. Assume that the output dimension of the first fully connected layer of the feedforward network is 4C (the multiple of the intermediate expansion dimension can be adjusted according to actual conditions), then the calculation after the feedforward network is:

[0147] FFN(X)=ReLU(XW1+b1)W2+b2

[0148] where W1∈R C×4C , b1 is the weight and bias of the first fully connected layer, W2∈R 4C×C , b2 is the weight and bias of the second fully connected layer, and the output dimension is the same as the input feature dimension, which is [batc size ,num patces ,C].

[0149] Select the loss function. According to the nature of the prediction target, use the mean square error loss function (MSE) or the cross entropy loss function (MSE). If the fitness value is a continuous value, the mean square error loss function is usually used, and its calculation formula is as follows:

[0150]

[0151] in is the fitness value of the i-th sample predicted by the model, F i is the actual fitness value of the sample.

[0152] If the fitness value is processed into a classification form, the cross entropy loss function is selected, and its calculation formula is:

[0153]

[0154] where y i is the true category label of the sample (0 or 1), is the probability that the sample belongs to category 1 as predicted by the model.

[0155] Set the number of training epochs, which indicates the number of times the entire training set will be reused to train the model.

[0156] Use an optimizer (such as Adam) to perform back propagation based on the calculated loss value and update the model parameters. Update the key parameters of SwinTransformer through the back propagation algorithm to minimize the loss function. Using the Adam optimizer, its update formula is as follows:

[0157]

[0158] Where m t and v t are the first-order moment estimate and the second-order moment estimate of the gradient, β1 and β2 are related hyperparameters, and g t is the current gradient, α is the learning rate, ∈ is a small constant to prevent the denominator from being zero, and θ t are model parameters.

[0159] Select sample data from the training set in batches, and set the batch size to batchc size , input the batch data into the SwinTransformer model for forward propagation calculation, and obtain the model's predicted fitness value for the batch of samples. According to the selected loss function, calculate the loss value L between the predicted value and the actual value MSE or L CE .

[0160] After training is completed, prepare the test set data, whose data format is consistent with the training set and validation set. Input the test set data batch by batch into the trained SwinTransformer model, and obtain the fitness value prediction result for each test set sample through the forward propagation calculation of the model.

[0161] The SwinTransformer model was optimized using the Socially Guided Dolphin Echolocation Algorithm (SDEA).

[0162] Fitness value predicted by SwinTransformer Each individual in the population is evaluated. The higher the fitness value, the better the hole configuration corresponding to the individual.

[0163] For each individual i, calculate the social influence of other individuals. The update calculation formula on dimension j is as follows:

[0164]

[0165] This formula means that individual i adjusts its position on dimension j according to the difference in fitness between other individuals and itself and the difference in position, and moves towards a better individual.

[0166] Update individual positions, the formula is as follows:

[0167]

[0168] For each individual i, the echolocation update in dimension j is:

[0169]

[0170] where r ij is a random number in the range [-1,1]. This update simulates the behavior of dolphins exploring their surroundings through echolocation, adding randomness to the search process.

[0171] Update individual positions, the formula is as follows:

[0172]

[0173] Check whether the updated individual position exceeds the reasonable value range of the void parameter, including whether the void shape coding is within the legal category range, whether the size parameter is within the size range that can be achieved by the process, whether the position parameter is within the area with low stress, etc. If it exceeds the range, adjust the individual position to the boundary value and update the individual position matrix X in the population.

[0174] Repeat 3.4 and 3.5 until the maximum number of iterations I is reached max After the iteration is completed, the individual with the highest fitness value F in the population is selected as the optimal solution. The hole parameter encoding corresponding to the optimal solution is the result of the hole optimization selection.

[0175] Decode the hole parameter encoding vector of the optimal solution to obtain the actual shape, size and position parameters of the hole. Output detailed parameter information of the hole, including the shape, size (such as the radius of a circle, the side length of a square, etc.) and position (coordinates) of each hole. At the same time, calculate and output the final total volume of material removal V 去 , rotor mass reduction Δm, and rotor stress distribution after optimization are used to evaluate the optimization effect and verify whether the design requirements are met.

[0176] For the rotor with a hollow structure after topological optimization design, finite element analysis software is used again to perform stress analysis, modal analysis, etc., to test the maximum stress, overall quality, and changes in performance indicators such as the limiting speed under the same working conditions (such as the same speed ω rpm, etc.). At the same time, the rotor with a traditional uniform disc design (same material and initial size) is compared to verify the optimization effect of the present invention, confirming that the optimized rotor has a lower maximum stress and significantly reduced overall quality compared to the uniformly designed rotor, allowing a higher speed and higher energy storage density, and effectively reducing the load or unloading pressure of the magnetic bearing.

[0177] Mass reduction analysis

[0178] For the initial design (assuming a uniform disc rotor), the disc mass is calculated according to the formula:

[0179]

[0180] Where ρ is the material density (kg / m3), D is the disk diameter (m), H is the disk thickness (m), and V initial is the initial disk volume.

[0181] According to the optimized cavity configuration, calculate the total volume of material removal V 去 , let the area of ​​a single hole be S i (square meters), there are n holes in total, then

[0182] Optimized rotor quality

[0183] Calculate the mass reduction Δm=m initial -m optimized .

[0184] Calculate the mass reduction percentage:

[0185]

[0186] By comparing the mass before and after optimization and the mass reduction percentage, analyze the effect of the optimization algorithm in reducing mass. If the mass reduction percentage does not reach the expected target, repeat step 3 for correction.

[0187] Analysis of stress distribution changes

[0188] Use ANSYS to calculate the stress distribution of the initial design rotor at the operating speed. Analyze the characteristics of the initial design stress distribution and determine the maximum stress value and The same method is used to calculate the stress distribution of the optimized rotor (with cavities) at the same operating speed, and the radial stress σ r,optimized and the tangential stress σ θ,optimized The distribution of .

[0189] Determine the maximum stress value of the optimized stress distribution.

[0190] Calculate the stress reduction value:

[0191]

[0192] Calculate the stress reduction percentage:

[0193]

[0194] By comparing the stress reduction value and the stress reduction percentage, the effect of the optimization algorithm on the improvement of stress distribution can be evaluated. If the stress performance in key areas (such as near the rotor surface or center) and the maximum stress reduction percentage do not meet the expected goals, repeat step 3 for correction.

[0195] Limiting speed analysis

[0196] Initial design, the limit speed n can be calculated by the critical speed calculation formula initial (rev / min). The optimized rotor also uses the corresponding method to calculate its limiting speed n optimized .

[0197] Calculate the speed increase value Δn = n optimized -n initial And the speed increase percentage:

[0198]

[0199] Speed ​​increase percentage

[0200] Energy storage density analysis

[0201] The energy storage density calculation formula is: Where I is the moment of inertia, ω is the rotational speed, and m is the rotor mass.

[0202] Initial design calculation of its energy storage density E initial , moment of inertia I initial According to the disk rotation inertia formula Calculate the energy storage density E of the optimized rotor optimized , where the moment of inertia I optimized Need to recalculate based on the optimized rotor geometry.

[0203] Calculate the energy storage density increase ΔE=E optimized -E initial And the percentage increase in energy storage density:

[0204]

[0205] By comparing the changes in energy storage density, the impact of the optimization algorithm on the performance of the energy storage system can be evaluated. If the energy storage density improvement does not reach the expected goal, repeat the steps of "in the identified area with lower stress, according to the structural strength requirements, process feasibility and other comprehensive factors, reasonably design the shape, size and distribution position of the cavity and other parameters, and use the social-oriented dolphin echolocation algorithm (SDEA) to optimize the combined optimization algorithm of SwinTransformer key parameters to achieve cavity optimization selection and calculation" for correction. The rotor of the flywheel energy storage system is initially designed as a uniform disk made of high-strength steel with a density of kg / m3, Poisson's ratio (V=0.3), diameter (D=1) meter, and thickness (H=0.2) meter.

[0206] Stress analysis: Using ANSYS software, the rotor speed was set at rpm, and simulation analysis was performed to obtain the stress distribution. It was found that the stress value in the annular area with a radius (r) of 0.2-0.4 meters from the center of the disk was continuously lower than the set stress threshold Pascal, and the area was marked as an optimizable area.

[0207] Topology optimization design: In this optimizable area, a circular cavity is designed with a cavity diameter (d = 0.1) meters, and a total of 10 cavities are evenly arranged. According to the above material removal and mass reduction calculation formula, the area of ​​a single cavity is square meters, the total volume of material removal is cubic meters, and the rotor mass reduction is kilograms.

[0208] Performance verification: The optimized rotor was simulated and analyzed again, and it was found that its maximum stress was reduced to 60% of the original, and the overall mass was reduced by about 20%. By comparing the energy storage density calculation formula, it was found that the energy storage density was increased by about 30%, and the maximum speed was increased from the original 12,000 rpm to 15,000 rpm. At the same time, the bearing pressure of the magnetic bearing was also significantly reduced, verifying the effectiveness and advantages of the topological optimization design of the present invention.

[0209] Through the above detailed comparative analysis of performance indicators, the effectiveness of the optimization algorithm can be fully verified, and whether the optimized cavity configuration meets the design requirements and expected goals in terms of mass reduction, stress distribution improvement, limit speed increase, and energy storage density improvement can be evaluated. At the same time, the optimization algorithm can be further improved and adjusted based on the analysis results to obtain a better design solution.

[0210] Example 3

[0211] This embodiment also provides a computer device, which is applicable to a flywheel energy storage rotor topology optimization method, including a memory and a processor; the memory is used to store computer executable instructions, and the processor is used to execute computer executable instructions to implement a forced oscillation detection and positioning method for a distribution network as proposed in the above embodiment.

[0212] This embodiment further provides a storage medium on which a computer program is stored. When the program is executed by a processor, a forced oscillation detection and positioning method for a distribution network is implemented as proposed in the above embodiment.

[0213] The computer device may be a terminal, and the computer device includes a processor, a memory, a communication interface, a display screen and an input device connected via a system bus. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The communication interface of the computer device is used to communicate with an external terminal in a wired or wireless manner, and the wireless manner can be achieved through WIFI, an operator network, NFC (near field communication) or other technologies. The display screen of the computer device may be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer device may be a touch layer covering the display screen, or a key, trackball or touchpad provided on the housing of the computer device, or an external keyboard, touchpad or mouse, etc.

[0214] If the function is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions for a computer device (which can be a personal computer, server, or network device, etc.) to perform all or part of the steps of the methods of each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc. Various media that can store program codes.

[0215] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as an ordered list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by an instruction execution system, device or apparatus (such as a computer-based system, a system including a processor, or other system that can fetch instructions from an instruction execution system, device or apparatus and execute instructions), or in conjunction with such instruction execution systems, devices or apparatuses. For the purposes of this specification, "computer-readable medium" can be any device that can contain, store, communicate, propagate or transmit a program for use by an instruction execution system, device or apparatus, or in conjunction with such instruction execution systems, devices or apparatuses.

[0216] More specific examples of computer-readable media (a non-exhaustive list) include the following: an electrical connection with one or more wires (electronic device), a portable computer disk case (magnetic device), a random access memory (RAM), a read-only memory (ROM), an erasable and programmable read-only memory (EPROM or flash memory), an optical fiber device, and a portable compact disk read-only memory (CDROM). In addition, the computer-readable medium may even be a paper or other suitable medium on which the program is printed, since the program may be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, deciphering, or processing in another suitable manner as necessary, and then stored in a computer memory.

[0217] It should be understood that the various parts of the present invention can be implemented by hardware, software, firmware or a combination thereof. In the above embodiments, a plurality of steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented by hardware, as in another embodiment, any one of the following technologies known in the art or a combination thereof can be used to implement: a discrete logic circuit having a logic gate circuit for implementing a logic function for a data signal, a dedicated integrated circuit having a suitable combination of logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc.

[0218] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A flywheel energy storage rotor topology optimization method, characterized in that: include, Use finite element analysis software to simulate and analyze the initial shape of the rotor and obtain rotor stress distribution data; Based on the stress distribution data, identifying areas with lower stress in the rotor, and setting stress thresholds to mark areas that can be optimized; In the optimizable region, cavity parameters are designed based on SDEA optimization of the SwinTransformer algorithm to achieve rotor structure optimization; The performance of the optimized rotor structure is verified to obtain optimization effect data.

2. A flywheel energy storage rotor topology optimization method as claimed in claim 1, characterized in that: The simulation analysis of the initial shape of the rotor using finite element analysis software includes: setting rotor material parameters, including density and Poisson's ratio; establishing a rotor geometric model and performing meshing; applying axial and radial displacement constraints; setting rotation speed operating conditions; and calculating the stress distribution in each region of the rotor under the set operating conditions.

3. A flywheel energy storage rotor topology optimization method as claimed in claim 2, characterized in that: The method of identifying the region with low stress in the rotor includes: setting a stress threshold σ0; extracting the radial stress σ of each node of the rotor r and the tangential stress σ θ ; Calculate the node comprehensive stress σ i =max(σ r,i ,σ θ,i );When the node comprehensive stress σ i When it is less than the stress threshold σ0, the area where the node is located is marked as an optimizable area.

4. A flywheel energy storage rotor topology optimization method as claimed in claim 3, characterized in that: The method for designing void parameters based on SDEA-based optimization of the SwinTransformer algorithm includes: expressing void parameters in a unified coding method; establishing a comprehensive fitness function including structural strength, mass reduction and process feasibility; predicting the void fitness value using the SwinTransformer model; and iterating and optimizing the model parameters through the SDEA algorithm.

5. A flywheel energy storage rotor topology optimization method as claimed in claim 4, characterized in that: The unified encoding method of the hole parameters includes: a circular hole is represented by a vector [x, y, r, 0], where (x, y) is the coordinate of the center of the circle, r is the radius, and 0 represents a circle; a square hole is represented by a vector [x, y, a, θ, 1], where (x, y) is the center coordinate, a is the side length, θ is the rotation angle, and 1 represents a square.

6. A flywheel energy storage rotor topology optimization method as claimed in claim 5, characterized in that: The comprehensive fitness function includes: a structural strength fitness function f1, which is used to evaluate the stress distribution under the void configuration; a mass reduction fitness function f2, which is used to evaluate the material removal effect; a process feasibility fitness function f3, which is used to evaluate the feasibility of the void parameters; and a comprehensive fitness function F=f1×f2×f3.

7. A flywheel energy storage rotor topology optimization method as claimed in claim 6, characterized in that: The optimization iteration process includes: evaluating individuals in the population according to fitness values; calculating the social influence of individuals on other individuals; updating echolocation; checking whether the updated position is within a reasonable range; and repeating iterations until the maximum number of iterations is reached; The optimization effect data include: mass reduction amount and reduction percentage; stress distribution change and reduction percentage; limit speed increase value and increase percentage; energy storage density increase value and increase percentage.

8. A flywheel energy storage rotor topology optimization system, based on the flywheel energy storage rotor topology optimization method according to any one of claims 1 to 7, characterized in that: It also includes a stress analysis module, which uses finite element analysis software to simulate and analyze the initial shape of the rotor and obtain rotor stress distribution data; A region identification module, which identifies regions with lower stress in the rotor according to the stress distribution data, and sets stress thresholds to mark regions that can be optimized; An optimization design module, which optimizes the SwinTransformer algorithm based on SDEA to design cavity parameters in the optimizable area; The performance verification module verifies the performance of the optimized rotor structure and obtains optimization effect data.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the flywheel energy storage rotor topology optimization method according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the flywheel energy storage rotor topology optimization method according to any one of claims 1 to 7 are implemented.

Images