A wind turbine front frame topology optimization method based on a proxy model

Abstract

The application discloses a wind turbine front frame topology optimization method based on a proxy model, and aims to solve the problems of mutual hindering between wind turbine lightweight and high reliability and shortening of the whole machine research and development cycle. For the purpose, the application takes the minimum compliance as the target, designs an artificial volume ratio constraint, applies different direction torques at the hub center to obtain respective topology optimization results and analyzes the result characteristics. The front frame is subjected to model reconstruction, finite element analysis is carried out based on the proxy model to determine the structure type and size. The limit strength and fatigue strength checking results of the optimized structure show that the proposed topology optimization method has feasibility and effectiveness in the lightweight design of the wind turbine front frame structure, and greatly shortens the product research and development cycle.

Description

Technical Field

[0001] This invention relates to the technical field of wind turbine topology optimization, and in particular to a method for optimizing the topology of the front rack of a wind turbine based on a surrogate model. Background Technology

[0002] Topology optimization for continuum structures refers to finding the optimal material distribution within a given region to achieve the design objective while satisfying certain constraints. In recent years, the single-unit capacity of wind turbines has continuously exceeded limits, but occasional strength failure events such as tower collapse, bolt breakage, and surface crack initiation in components have highlighted the growing contradiction between lightweight structures and high reliability. Against this backdrop, topology optimization-based structural innovation tools have been introduced into wind turbine structural design.

[0003] Although topology optimization methods have seen initial success in the structural design of wind turbine components, the following difficulties and problems still exist: Firstly, while relatively mature stiffness optimization methods are widely used, wind turbine structural components must meet strength design requirements. Despite advancements in stress-constrained and fatigue-constrained topology optimization methods, successful engineering applications are still rare. Secondly, wind turbine design conditions number in the hundreds or thousands, making topology optimization essentially a multi-objective optimization problem. The large number of finite element analyses required leads to computational difficulties. Thirdly, topology optimization results are expressed as material packing patterns, making it difficult to translate conceptual design results into final engineering designs. Fourthly, volume ratios are artificially defined parameters, and topology optimization results are expressed as force transmission paths rather than structural dimensions along those paths. Based on clear force transmission paths, extensive structural re-analysis is needed to ultimately determine the structural shape and dimensions. Summary of the Invention

[0004] The purpose of this invention is to overcome the shortcomings of the prior art and propose a topology optimization method for the front rack of a wind turbine based on a proxy model. This method solves the problem of the large amount of computing power required in the topology optimization process, which leads to excessively long analysis time. It optimizes the design cycle of the front rack of the wind turbine, simplifies the calculation time, and meets the requirements of lightweight design.

[0005] The objective of this invention is achieved through the following technical solution: a method for optimizing the topology of the front rack of a wind turbine based on a surrogate model, comprising the following steps:

[0006] S1. Determine the load conditions and material properties of the wind turbine front frame, perform topology optimization analysis, and obtain the topology optimization results;

[0007] S2. Reconstruct the geometric model of the front frame based on the topology optimization results, perform finite element analysis based on the surrogate model, and iterate until the designed structure meets the preset preliminary strength design requirements, and determine the structural type and size.

[0008] S3. Verify the structural strength based on the detailed finite element model. If the preset lightweight design requirements for the wind turbine front frame are not met, return to step S2 and repeat the iteration until the preset lightweight design requirements for the wind turbine front frame are met.

[0009] Furthermore, step S1 includes:

[0010] The load conditions and material properties of the wind turbine front frame are determined. Based on the variable density method, with the goal of minimizing compliance, artificial volume ratio constraints are designed to obtain the optimal topology optimization results under a single load.

[0011] Specifically, the design load values ​​at the center of the wind turbine hub are calculated based on wind turbine design software. The load transfer path is as follows: through the main shaft to the bearing housing, then through the flange bolts to the front frame, and finally through bolts to the rear frame to transfer the weight of the accessories it carries to the front frame. Each ultimate load condition includes forces and moments in three directions. Ultimate load condition M is selected. j_max and M j_min The working condition is the load condition, where j = x, y, z, yz.

[0012] Furthermore, step S1 includes:

[0013] The material for the front frame was determined to be QT400, with a yield strength of 220 MPa. Using wind turbine design software, the entire area occupied by the front frame was filled with this material. The design area was discretized into N elements, each corresponding to a pseudo-density value x in the range of 0 to 1. e e = 1, 2, ..., N; when x e A value of 0 indicates that the cell is deleted; when x... e A value of 1 indicates that the cell is retained. To suppress the number of cells with intermediate density values, x is established. e and unit elastic modulus E e Punishment relationship:

[0014]

[0015] In the formula, p is the penalty parameter, which is initially set to 1 and gradually increases with optimization iterations, and E0 is the elastic modulus of the solid material.

[0016] Furthermore, step S1 includes:

[0017] The force transmission paths are examined using a single load, i.e., the moments in the x, y, and z directions. The topology optimization formula, with minimum compliance as the objective and volume ratio as the constraint, is as follows:

[0018] min:c=F T U

[0019]

[0020] KU = F

[0021] 0 < ρ min ≤ρ e ≤1

[0022] In the formula, c is the compliance value; V is the optimized structural volume. The design domain volume is represented by f, which is the volume ratio that needs to be determined manually; K is the overall stiffness matrix; U is the displacement vector; F is the load vector; ρ is the load vector. min This represents the minimum unit density, with a value of 0.01.

[0023] Furthermore, step S2 includes:

[0024] The reconstruction of the geometric model of the front frame based on the topology optimization results involves selection based on the analysis of topology optimization results for different single loads. The determination of the structural type and size based on the surrogate model for finite element analysis is achieved by using the surrogate model instead of the refined model. The surrogate model reduces the size of the finite element model or simplifies the solution type by using low-order element types and simplifying the connection relationships of assembly components. While retaining the optimal force transmission path, multiple iterations of finite element analysis are performed to quantify the structure until the designed structure meets the preset preliminary strength design requirements, thereby determining the specific structural type and structural dimensions and completing the structural design. At the same time, the results of the refined model can also be used to correct the results of the surrogate model.

[0025] Furthermore, step S2 includes:

[0026] The refined model includes refined models of the tower, front frame, rear frame, bearing housing, bolts, and bearings. All refined models comprise 6.17 million nodes and 3.39 million elements. The front frame is discretized using primarily hexahedral solid elements, comprising 4.5 million nodes and 1.35 million elements. The contact relationships between bolted components are defined, and nonlinear springs are used to simulate the bearings. At the fully constrained bottom of the tower, loads in different directions are applied at the hub center, and rigid elements are used to achieve load transfer. The proxy model includes proxy models of the tower, front frame, rear frame, bearing housing, bolts, and bearings. The front frame proxy model is a low-order solid element discretization of the front frame, comprising 1.08 million nodes and 530,000 elements. The nonlinear contact between components is simplified to a binding relationship. Based on the load conditions, corresponding stress thresholds are set for the proxy model to make its contour plot approximate that of the refined model.

[0027] Furthermore, step S3 includes:

[0028] The structural strength verification based on a detailed finite element model is based on a preset material safety factor of 1.1 and an allowable stress of 200 MPa for the wind turbine front frame material. The maximum equivalent stress under different extreme conditions is calculated using a refined model. If the maximum equivalent stress under each condition is less than the allowable stress, the designed front frame structure is deemed to meet the ultimate strength design requirements. The SN curve of the front frame material QT400 is synthesized. Based on this, fatigue strength analysis of the front frame is performed using the critical plane method for multiaxial fatigue. The influence of average stress on fatigue damage is corrected using FKM, and the cumulative fatigue damage distribution of the front frame is obtained. If the cumulative fatigue damage at all points on the front frame structure is less than 1, it indicates that the fatigue strength design requirements are met. If the designed structure does not meet the preset ultimate strength design requirements and fatigue strength design requirements, the process returns to step S2 and repeats the iteration until the preset ultimate strength design requirements and fatigue strength design requirements are met.

[0029] A non-transitory computer-readable medium storing instructions, characterized in that, when the instructions are executed by a processor, the steps of the wind turbine front rack topology optimization method based on the surrogate model described above are performed.

[0030] A computing device includes a processor and a memory for storing processor-executable programs. When the processor executes the programs stored in the memory, it implements the above-described agent-based wind turbine front rack topology optimization method.

[0031] Compared with the prior art, the present invention has the following advantages and beneficial effects:

[0032] This invention takes the front frame of a wind turbine as the research object. Based on topology optimization, it proposes a structural reanalysis method based on a surrogate model, thereby realizing rapid iteration of the structural design process. Furthermore, it verifies the ultimate strength and fatigue strength of the optimized structure of the wind turbine front frame to ensure its feasibility. Compared with the traditional conventional design process, it greatly shortens the R&D cycle of the entire structural design. Attached Figure Description

[0033] Figure 1 This is a flowchart of a method for optimizing the topology of the front rack of a wind turbine based on a proxy model.

[0034] Figure 2 This is a schematic diagram of the optimal topology under a single torque.

[0035] Figure 3 This is a comparison chart of the equivalent stress distribution under the refined model and the surrogate model.

[0036] Figure 4 A comparison diagram of equivalent stress distribution of the through-hole structure under different improvement schemes.

[0037] Figure 5This is a comparison chart of the maximum equivalent stress under different extreme working conditions.

[0038] Figure 6 This is a diagram showing the cumulative fatigue damage distribution of the front frame. Detailed Implementation

[0039] The present invention will be further described below with reference to specific embodiments.

[0040] Example 1

[0041] See Figure 1 As shown, the wind turbine front rack topology optimization method based on the surrogate model provided in this embodiment includes the following steps:

[0042] S1. Determine the load conditions and material properties of the wind turbine's front frame, perform topology optimization analysis, and obtain the topology optimization results, including:

[0043] First, the load conditions and material properties of the wind turbine's front frame are determined. Based on the wind turbine design software, the design load values ​​at the center of the wind turbine hub are calculated. The load transfer path is as follows: through the main shaft to the bearing housing, then through the flange bolt connection to the front frame, and finally through bolts from the rear frame to the front frame carrying the weight of its accessories. Each ultimate load condition includes forces and moments in three directions. Ultimate load condition M is selected. j_max and M j_min The working condition is the load-bearing working condition, where j = x, y, z, yz, with M as the load condition. y_max For example, in all load conditions, the scalar value of the bending moment in the y-direction is the largest. Since the absolute value of the resultant moment in the yoz plane is small, M can be ignored in the ultimate strength check. yz_min Operating conditions; the material of the front frame is determined to be QT400, with a yield strength of 220MPa; based on wind turbine design software, the entire area occupied by the front frame is filled with the material, and the design area is discretized into N elements, each element corresponding to a pseudo-density value x in the range of 0 to 1. e e = 1, 2, ..., N; when x e A value of 0 indicates that the cell is deleted; when x... e A value of 1 indicates that the cell is retained. To suppress the number of cells with intermediate density values, x is established. e and unit elastic modulus E e Punishment relationship:

[0044]

[0045] In the formula, p is the penalty parameter, which is initially set to 1 and gradually increases with optimization iterations, and E0 is the elastic modulus of the solid material.

[0046] According to DNV 2016 specifications, wind turbine structural design involves a large number of load cases, making topology optimization essentially a multi-objective optimization problem. Generally, the results of multi-load case topology optimization are not simply a superposition of single-load case topology optimization results. While multi-load case optimization has the advantage of coordinating force transmission paths under different load conditions, it can also obscure the force transmission path under a single load. To overcome this problem, we examine the corresponding force transmission paths for each single load (i.e., the moments in the x, y, and z directions), with minimum compliance as the objective and volume ratio as a constraint, as shown in the following topology optimization formula:

[0047] min:c=F T U

[0048]

[0049] KU = F

[0050] 0 < ρ min ≤ρ e ≤1

[0051] In the formula, c is the compliance value; V is the optimized structural volume. The design domain volume is represented by f, which is the volume ratio that needs to be determined manually; K is the overall stiffness matrix; U is the displacement vector; F is the load vector; ρ is the load vector. min This represents the minimum unit density, with a value of 0.01.

[0052] Topology optimization results commonly exhibit checkerboard patterns and mesh dependency issues. This embodiment sets a minimum size to eliminate these two types of numerical instability. Using the above settings, the topology optimization results under single bending moments in different directions are as follows: Figure 2 As shown, where Figure 2 (a) is a single load M x Topology optimization results under the action, Figure 2 (b) is a single load M y Topology optimization results under the action, Figure 2 (c) is a single load M z Topology optimization results under the influence of the action;

[0053] S2. Reconstruct the geometric model of the front frame based on the topology optimization results, perform finite element analysis based on the surrogate model, iterate until the designed structure meets the preset preliminary strength design requirements, and determine the structural type and dimensions, including:

[0054] In common knowledge in this field, the bending moment M is used as a reference. y and M z Extreme operating conditions dominated by the environment are usually quite dangerous, and this section will focus on them. Figure 2 The force transmission paths are represented by (b) and (c). Figure 2(b) Taking this as an example, the rear bearing housing transmits force to the top of the tower through the side wall, forming a clear force transmission path. A large hole is formed in the side wall between the front and rear bearing housings, indicating that this area does not bear the bending moment M. y The role of topology optimization is crucial. Therefore, the reconfiguration of the front frame structure requires a trade-off based on the analysis of topology optimization results under different single loads. It is worth noting that the above topology optimization uses a relatively mature optimization model from theoretical research, namely a conceptual design scheme under the objective of maximizing stiffness and the constraint of volume ratio. However, key components of the wind turbine must meet strength design requirements. This necessitates determining the specific structural type and dimensions through finite element analysis results while retaining the optimal force transmission path, ultimately achieving the required strength. Since topology optimization results are applicable to the conceptual design stage, this process typically requires repeated finite element analyses to achieve the goal of detailed design.

[0055] Using a surrogate model for finite element analysis to determine the structural type and dimensions involves replacing the refined model with a surrogate model. The surrogate model reduces the size of the finite element model or simplifies the solution type by using lower-order element types and simplifying the connection relationships of assembly components. While retaining the optimal force transmission path, multiple iterations of finite element analysis are performed to quantify the structure until it meets the preset preliminary strength design requirements. This determines the specific structural type and dimensions, completing the structural design. At the same time, the results of the surrogate model can also be corrected using the results of the refined model.

[0056] The refined model includes refined models of the tower, front frame, rear frame, bearing housing, bolts, and bearings. All refined models comprise 6.17 million nodes and 3.39 million elements. The front frame is discretized using predominantly hexahedral solid elements, comprising 4.5 million nodes and 1.35 million elements. The contact relationships between bolted components are defined, and nonlinear springs are used to simulate the bearings. At the fully constrained tower bottom, loads in different directions are applied at the hub center, and rigid elements are used to implement load transfer. On a high-performance computing system, it takes approximately 48 hours to complete calculations for seven extreme conditions. The proxy model includes proxy models of the tower, front frame, rear frame, bearing housing, bolts, and bearings. The front frame proxy model is a low-order solid element discretization of the front frame, comprising 1.08 million nodes and 530,000 elements. This simplifies the nonlinear contact between components to a bonding relationship, reducing the total computation time to less than one hour. y_min Taking the working condition as an example, the equivalent stress distribution of the refined model and the surrogate model is as follows: Figure 3 As shown, where Figure 3 (a) is a refined model. Figure 3(b) is a surrogate model, with the stress thresholds for the contour plots set to 200 MPa and 150 MPa, respectively. Different stress thresholds need to be adjusted according to different working conditions to make the contour plots of the surrogate model approximate the refined model as closely as possible, thereby improving the reliability of the surrogate model.

[0057] Depend on Figure 3 It can be seen that although the two types of models have different levels of refinement, the stress distribution at the critical locations is consistent. In fact, the refinement process of the simulation model only approximates the real situation to a certain extent. Although the surrogate model has poor accuracy, it has advantages such as being able to quickly determine the structural form during the structural conceptual design stage. Subsequent qualitative assessments of structural improvements can be largely carried out on the surrogate model.

[0058] Taking the improvement of the wire hole structure as an example, the position of the wire hole is moved outward by 30mm, and the result is examined. y_min and M z_max Two operating conditions, calculated based on a surrogate model, show the equivalent stress distribution before and after structural improvement, for example... Figure 4 As shown, where Figure 4 (a) is the reference structure M z_max Operating conditions Figure 4 (b) is the outward movement of the thread hole by M z_max Operating conditions Figure 4 (c) is the reference structure M y_min Operating conditions Figure 4 (d) represents the outward movement of the thread hole by M. y_min Under operating conditions, moving the wire hole position outward can appropriately reduce M. z_max The stress level of this dangerous working condition, but for M y_min The stress distribution under operating conditions has little impact.

[0059] S3. Verify the structural strength based on the detailed finite element model. If the preset lightweight design requirements for the wind turbine front frame are not met, return to step S2 and repeat the iteration until the preset lightweight design requirements for the wind turbine front frame are met, including:

[0060] Based on a preset material safety factor of 1.1, the allowable stress of the wind turbine front frame material is 200 MPa. The maximum equivalent stress under different extreme conditions is calculated using a refined model, as follows: Figure 5 As shown, in all operating conditions, M y_max The maximum equivalent stress value is the largest under the working condition, and the location is near the rear cable hole. The maximum equivalent stress value under each working condition is less than the allowable strength value, and the front frame meets the ultimate strength design requirements.

[0061] Based on the SN curve of the front frame material QT400 synthesized according to the DNV 2016 certification specifications, fatigue strength analysis of the front frame was performed using the critical plane method for multiaxial fatigue. The influence of mean stress on fatigue damage was corrected using the FKM method, and the results are as follows: Figure 6 The cumulative fatigue damage distribution of the front frame, as shown in the figure, indicates that, except for locations such as the bearing housing surface and yaw contact surface, the cumulative fatigue damage at all points on the front frame structure is less than 1, suggesting that the fatigue strength design requirements are met. The above limit and fatigue strength verification results further demonstrate the feasibility of the topology optimization method in the lightweight design of the front frame structure.

[0062] Example 2

[0063] This embodiment discloses a non-transitory computer-readable medium storing instructions that, when executed by a processor, perform the steps of the wind turbine front rack topology optimization method based on the surrogate model as described in Embodiment 1.

[0064] In this embodiment, the non-transitory computer-readable medium can be a disk, optical disk, computer memory, read-only memory (ROM), random access memory (RAM), USB flash drive, portable hard drive, etc.

[0065] Example 3

[0066] This embodiment discloses a computing device, including a processor and a memory for storing processor-executable programs. When the processor executes the program stored in the memory, it implements the wind turbine front rack topology optimization method based on the proxy model described in Embodiment 1.

[0067] The computing device described in this embodiment may be a desktop computer, laptop computer, smartphone, PDA handheld terminal, tablet computer, programmable logic controller (PLC), or other terminal device with processor function.

[0068] The above-described embodiments are merely preferred embodiments of the present invention and are not intended to limit the scope of the present invention. Therefore, any changes made in accordance with the shape and principle of the present invention should be covered within the protection scope of the present invention.

Claims

1. A method for optimizing the topology of the front rack of a wind turbine based on a surrogate model, characterized in that, Includes the following steps: S1. Determine the load conditions and material properties of the wind turbine's front frame, perform topology optimization analysis, and obtain the topology optimization results, including: The load conditions and material properties of the wind turbine front frame are determined. Based on the variable density method, with the goal of minimizing compliance, artificial volume ratio constraints are designed to obtain the optimal topology optimization results under a single load. Specifically, the design load values ​​at the center of the wind turbine hub are calculated based on wind turbine design software. The load transfer path is as follows: through the main shaft to the bearing housing, then through the flange bolts to the front frame, and finally through bolts to the rear frame to transfer the weight of the accessories it carries to the front frame. Each ultimate load condition includes forces and moments in three directions. Ultimate load condition M is selected. j_max and M j_min The working condition is the load-bearing working condition, where j = x, y, z, yz; S2. Reconstruct the geometric model of the front frame based on the topology optimization results, perform finite element analysis based on the surrogate model, iterate until the designed structure meets the preset preliminary strength design requirements, and determine the structural type and dimensions, including: The reconstruction of the geometric model of the front frame based on the topology optimization results involves selection based on the analysis of topology optimization results for different single loads. The determination of the structural type and size based on the surrogate model for finite element analysis is achieved by using the surrogate model instead of the refined model. The surrogate model reduces the scale of the finite element model or simplifies the solution type by using low-order element types and simplifying the connection relationships of assembly parts. While retaining the optimal force transmission path, multiple iterations of finite element analysis are used to quantify the structure until the designed structure meets the preset preliminary strength design requirements, thereby determining the specific structural type and structural size and completing the structural design. At the same time, the results of the surrogate model can also be corrected using the results of the refined model. S3. Verify the structural strength based on the detailed finite element model. If the preset lightweight design requirements for the wind turbine front frame are not met, return to step S2 and repeat the iteration until the preset lightweight design requirements for the wind turbine front frame are met.

2. The method for optimizing the topology of a wind turbine front rack based on a surrogate model according to claim 1, characterized in that, Step S1 includes: The material for the front frame was determined to be QT400, with a yield strength of 220 MPa. Using wind turbine design software, the entire area occupied by the front frame was filled with this material. The design area was discretized into N elements, each corresponding to a pseudo-density value x in the range of 0 to 1. e e = 1, 2, ..., N; when x e A value of 0 indicates that the cell is deleted; when x... e A value of 1 indicates that the cell is retained. To suppress the number of cells with intermediate density values, x is established. e and unit elastic modulus E e Punishment relationship: In the formula, p is the penalty parameter, which is initially set to 1 and gradually increases with optimization iterations, and E0 is the elastic modulus of the solid material.

3. The method for optimizing the topology of a wind turbine front rack based on a surrogate model according to claim 2, characterized in that, Step S1 includes: The force transmission paths are examined using a single load, i.e., the moments in the x, y, and z directions. The topology optimization formula, with minimum compliance as the objective and volume ratio as the constraint, is as follows: min:c=F T U KU = F 0<x min ≤x e ≤1 In the formula, c is the compliance value; V is the optimized structural volume. The design domain volume is represented by f, which is the volume ratio that needs to be determined manually; K is the overall stiffness matrix; U is the displacement vector; F is the load vector; x min This represents the minimum unit density, with a value of 0.

01.

4. The method for optimizing the topology of a wind turbine front rack based on a surrogate model according to claim 1, characterized in that, Step S2 includes: The refined model includes refined models of the tower, front frame, rear frame, bearing housing, bolts, and bearings. All refined models comprise 6.17 million nodes and 3.39 million elements. The front frame is discretized using primarily hexahedral solid elements, comprising 4.5 million nodes and 1.35 million elements. The contact relationships between bolted components are defined, and nonlinear springs are used to simulate the bearings. At the fully constrained bottom of the tower, loads in different directions are applied at the hub center, and rigid elements are used to achieve load transfer. The proxy model includes proxy models of the tower, front frame, rear frame, bearing housing, bolts, and bearings. The front frame proxy model is a low-order solid element discretization of the front frame, comprising 1.08 million nodes and 530,000 elements. The nonlinear contact between components is simplified to a binding relationship. Based on the load conditions, corresponding stress thresholds are set for the proxy model to make its contour plot approximate that of the refined model.

5. The method for optimizing the topology of a wind turbine front rack based on a surrogate model according to claim 1, characterized in that, Step S3 includes: The structural strength was verified based on a detailed finite element model. The preset material safety factor was 1.1, and the allowable stress of the wind turbine front frame material was 200 MPa. The maximum equivalent stress under different extreme working conditions was calculated based on the refined model. If the maximum equivalent stress under each working condition was less than the allowable stress, the designed front frame structure was deemed to meet the ultimate strength design requirements. The SN curve of the front frame material QT400 was synthesized. Based on this, the fatigue strength analysis of the front frame was performed using the critical plane method of multiaxial fatigue. The influence of average stress on fatigue damage was corrected using FKM, and the cumulative fatigue damage distribution of the front frame was obtained. If the cumulative fatigue damage at all points of the front frame structure was less than 1, it indicated that the fatigue strength design requirements were met. If the designed structure did not meet the preset ultimate strength design requirements and fatigue strength design requirements, the process returned to step S2 and iterated until the preset ultimate strength design requirements and fatigue strength design requirements were met.

6. A non-transitory computer-readable medium storing instructions, characterized in that, When the instruction is executed by the processor, the steps of the wind turbine front rack topology optimization method based on the proxy model according to any one of claims 1-5 are performed.

7. A computing device, comprising a processor and a memory for storing a processor-executable program, characterized in that, When the processor executes the program stored in the memory, it implements the wind turbine front rack topology optimization method based on the proxy model as described in any one of claims 1-5.

Images