A material distribution optimization method for blade model
By combining geometric methods and heuristic evolution algorithms, the material distribution of the blade model is optimized using the non-uniform rational B-spline basis function, and the problems of low optimization solution efficiency and poor smoothness of the design results are solved, and efficient material distribution optimization is achieved and improved The vibration characteristics of the rotating machinery.
Patent Information
- Application Number
- CN202210374472.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2022-04-11
- Publication Date
- 2024-12-06
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
In the prior art, the optimization solution efficiency of the blade model is low and the final design results are poorly smooth, making it difficult to ensure the optimization of the vibration characteristics of the rotating machinery.
Combining equal geometric methods and heuristic evolution algorithms, by constructing a non-uniform rational B-spline basis function, the material parameters of the constructed geometry are directly used as optimization control variables, avoiding repeated definition of geometry and grid division, and achieving continuous material distribution interface optimization.
Improve optimization efficiency, shorten the design cycle, and obtain a continuously changing optimized material distribution interface without further smoothing, which improves the vibration characteristics optimization effect of the blade model.
Smart Images

Figure CN114676529B_ABST
Abstract
Description
Technical Field
[0001] The invention relates to the field of structural dynamics, and in particular to a material distribution optimization method for a blade model. Background Art
[0002] In engineering, high-speed rotating machinery can be seen everywhere, and it is widely used in aircraft engines, turbines, compressors and other mechanisms. Blades are important working parts of high-speed rotating machinery, and their frequent vibration failures have become an important factor restricting the reliability of rotating machinery. Therefore, it is of great theoretical and practical significance to improve and enhance the vibration characteristics of blade models through the optimization design of early blades.
[0003] In recent years, most scholars have conducted research on the optimization design of rotating blades based on finite element simulation software. The finite element optimization method usually adopts an optimization parameterization method based on unit discretization. The final optimization result is obtained by assigning design parameters to discrete units / nodes piece by piece, and the optimization result also exists in a discrete form. When the finite element mesh is extremely fine, the continuity of the optimization result can be improved. However, such a fine mesh reduces the efficiency of the optimization solution, and it is still difficult to ensure the smoothness of the final design result.
[0004] Based on the NURBS isogeometric method, the basis function of the constructed geometry itself is used as the calculation function of the analysis and optimization, which can avoid the repeated definition of geometry and mesh re-division process in the finite element optimization analysis, thereby greatly improving the optimization efficiency. The material variables at the control points are used as the control variables of the optimization process and combined with the heuristic optimization algorithm to obtain a continuous optimal material distribution interface without further smoothing. Therefore, it is very necessary to establish such an effective material distribution optimization method that can use a heuristic optimization algorithm and is applicable to blade models of various materials. Summary of the invention
[0005] In order to solve the problems of low optimization solution efficiency and poor smoothness of the final design results existing in the above-mentioned prior art, the present invention provides a material distribution optimization method for a blade model. By combining the isogeometric method with the heuristic evolutionary algorithm, a continuously changing optimized material distribution interface can be obtained without further smoothing processing, which greatly improves the optimization efficiency and shortens the design cycle of the entire blade material optimization.
[0006] In order to achieve the above technical objectives, the present invention provides a material distribution optimization method for a blade model, comprising:
[0007] Step 1, construct a blade model and extract basic information of the blade model;
[0008] Step 2, constructing a non-uniform rational B-spline basis function based on the basic information of the blade model and the refinement strategy of the isogeometric analysis method;
[0009] Step 3, using the non-uniform rational B-spline basis function to perform interpolation calculation to obtain the material properties of the blade model;
[0010] Step 4, based on the material properties of the blade model, analyzing the vibration characteristics of the blade model to obtain an objective function value;
[0011] Step 5, obtaining a blade model fitness function value based on the objective function value and the constraint conditions of material distribution optimization;
[0012] Step 6, using the optimization algorithm program to iteratively solve, calculate and compare the fitness function value of the blade model, obtain the best design variable individual solution and the best material distribution optimized blade model, and complete the material distribution optimization of the blade model.
[0013] Optionally, the formula of the non-uniform rational B-spline basis function is:
[0014]
[0015] Among them, ω i,j is the weight information, p and q are the order information, ζ,η are the node vector information, N i,p (ζ),M i,q (η) represents the univariate B-spline function along the two node vector directions ζ and η, respectively, m and n are the number of control points in the two node vector directions, and i and j are the i-th and j-th control points in the two node vector directions.
[0016] Optionally, the process of obtaining the objective function value includes:
[0017] The strain energy and kinetic energy of the blade model are obtained based on the first-order shell theory;
[0018] Based on the strain energy and kinetic energy of the blade model and the work done by the centrifugal force, an energy functional of the blade model is obtained;
[0019] Based on the energy functional and vibration control equation of the blade model, an objective function value is obtained.
[0020] Optionally, the process of obtaining the fitness function value of the blade model includes:
[0021] Setting constraint conditions for the material distribution optimization, including: equality constraints and inequality constraints;
[0022] The blade model fitness function value is obtained based on the penalty function method, the constraint conditions of the material distribution optimization and the objective function value.
[0023] Optionally, the formula for the material property of the blade model is:
[0024] P=P me +(P ce -P me )V c
[0025]
[0026] Among them, P is the material parameter of the blade model, P me ,P ce is the material parameter of blade models with different materials, V c is the material property of the blade model, a is the ath control point, is the material property at the control point of the blade model, m and n are the number of control points in the direction of the two node vectors ζ and η, is the non-uniform rational B-spline basis function.
[0027] Optionally, the expression of the first-order shell theory is:
[0028] u(x,z)={u(x,z)w(x,z)v(x,z)} T
[0029]
[0030]
[0031] w(x,z)=w 0 (x,z)
[0032]
[0033] Where u, w, v are the deformation displacements of any point on the blade model, u 0 ,v 0 ,w 0 is the in-plane displacement of a point on the neutral plane, O-XYZ is the global coordinate system, x and z are the directions of the coordinate system, is the rotation angle of a point in the neutral plane along the x-axis and z-axis, ζ,η are the node vector information, m,n are the number of control points in the two node vector directions, p and q are the order information, and u a is the deformation displacement at the ath control point.
[0034] Optionally, the formula for the centrifugal work of the blade model is:
[0035] W = ∫∫∫ Ω N lx u ,x +N ly u,y +N lz u ,z dΩ
[0036] Where N lx ,N ly ,,N lz is the centrifugal force component along the x, y, and z directions, u ,x ,u ,y ,u ,z is the derivative of the displacement variable along the x, y, and z directions.
[0037] Optionally, the vibration control equation is:
[0038]
[0039] Where M is the mass matrix, K is the stiffness matrix, and M G is the stress matrix, M c is the centrifugal force weakening matrix, and q is the vibration displacement.
[0040] The present invention has the following technical effects:
[0041] By combining the isogeometric method with the heuristic evolutionary algorithm, the material parameters at the construction geometry control points are directly used as the control variables during optimization, and the material properties of the blade model are represented together with the basis functions used for the construction geometry, so that a continuously changing optimized material distribution interface can be obtained without further smoothing. In addition, the present invention closely combines the geometric modeling software CAD with the optimization analysis software CAE, and there is no need to perform a meshing process on the constructed blade model again, which greatly improves the optimization efficiency and shortens the design cycle of the entire blade material optimization. BRIEF DESCRIPTION OF THE DRAWINGS
[0042] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. 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 paying creative labor.
[0043] Figure 1 A flow chart of a method for optimizing material distribution of a blade model according to an embodiment of the present invention;
[0044] Figure 2 A blade model diagram of an embodiment of the present invention;
[0045] Figure 3 A distribution diagram of blade design control points according to an embodiment of the present invention;
[0046] Figure 4It is a conversion relationship diagram among the entity domain, parameter domain and parent domain in an embodiment of the present invention. DETAILED DESCRIPTION
[0047] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.
[0048] like Figure 2 As shown in the figure, an initial blade model is established. The blade dimensions are L = 0.4m, b = 0.2m, h = 0.01m, respectively, and fixed at a radius of R h = 0.4 on the hub. Assuming that the material to be optimized for the blade model is a functionally gradient material, Young's modulus E me =210GPa,E ce =390GPa, density ρ me =7850 / m 3 ,ρ ce =3960kg / m 3 , Poisson's ratio μ me =μ ce =0.3, the optimization control variable is set to the ceramic material fraction Vc at the control point. The optimization goal is to maximize the fundamental frequency of the blade vibration, and the constraint condition is that the volume fraction of the overall ceramic material of the blade model is less than or equal to 60%. The blade model x=0 is fixed at one end, and the other three directions are unconstrained and have no external loads. Consider four working states of the rotating blade model: State 1, the rotation speed is 0, and the torsion angle is 0; State 2, the rotation speed is Φz=100πrad / s, and the torsion angle is 0; State 3, the rotation speed is 0, and the torsion angle is α=30°; State 4, the rotation speed is Φz=100πrad / s, and the torsion angle is α=30°. The distribution of the design control points is as follows: Figure 3 As shown, due to the symmetry of the boundary conditions, only 1 to 8 control points are selected as the control variables for actual calculation.
[0049] like Figure 1 As shown, the present invention discloses a material distribution optimization method for a blade model, comprising:
[0050] Step 1, construct a blade model and extract basic information of the blade model;
[0051] First, build a blade model in the modeling software and extract the basic information of the blade model, including: control point information B a , weight information w i,j , node vector information ζ,η, order information p,q;
[0052] The basic information tables are shown in Table 1 and Table 2 below:
[0053] Table 1
[0054]
[0055] Table 2
[0056]
[0057] Step 2, constructing a non-uniform rational B-spline basis function based on the basic information of the blade model and the refinement strategy of the isogeometric analysis method;
[0058] Based on the basic information of the blade model obtained in step 1 and the refinement strategy of the isogeometric analysis method, a non-uniform rational B-spline (NURBS) basis function is constructed for vibration characteristic analysis and optimization research. The formula of non-uniform rational B-spline basis function is:
[0059] In the formula, ω i,j is the weight information, p and q are the order information, ζ,η are the node vector information, N i,p (ζ),M i,q (n) represents the univariate B-spline function along the two knot vector directions ζ and η, respectively, m and n are the number of control points in the two knot vector directions, and i and j are the i-th and j-th control points in the two knot vector directions.
[0060] Step 3, using the non-uniform rational B-spline basis function to perform interpolation calculation to obtain the material properties of the blade model;
[0061] According to the given design variable range variable range (V cmin ,V cmax ), determine the material parameters at the control point a of the blade model The formula is as follows:
[0062]
[0063] Where V cmin ,V cmax are the lower and upper limits of the material parameters respectively, and rand is any value between (0,1).
[0064] Material parameters at point a based on blade model control The non-uniform rational B-spline basis function is used for interpolation calculation to obtain the initial blade model material properties, and the formula is:
[0065] P=P me +(P ce-P me )V c
[0066]
[0067] Among them, P is the material parameter of the blade model, P me ,P ce is the material parameter of blade models with different materials, V c is the material property of the blade model, a is the ath control point, is the material property at the control point a of the blade model, m and n are the number of control points in the direction of the two node vectors ζ and η, is the non-uniform rational B-spline basis function.
[0068] Step 4, based on the material properties of the blade model, analyzing the vibration characteristics of the blade model to obtain an objective function value;
[0069] First, the strain energy and kinetic energy of the blade model are obtained based on the first-order shell theory. The expression of the first-order shell theory is: u(x,z)={u(x,z)w(x,z)v(x,z)} T
[0070]
[0071]
[0072] w(x,z)=w 0 (x,z)
[0073]
[0074] Where u, w, v are the deformation displacements of any point on the blade model, u 0 ,v 0 ,w 0 is the in-plane displacement of a point on the neutral plane, O-XYZ is the global coordinate system, x and z are the directions of the coordinate system, is the rotation angle of a point in the neutral plane along the x-axis and z-axis, ζ,η are the node vector information, m,n are the number of control points in the two node vector directions, p and q are the order information, and u a is the deformation displacement at the ath control point.
[0075] Then, the strain energy of the blade model is obtained:
[0076]
[0077] And the kinetic energy of the blade model:
[0078]
[0079] The velocity expression of any point on the blade model is:
[0080] Where ε and κ are the bending strain and shear strain at a point on the neutral surface, respectively; D is the material matrix of the blade model; h is the thickness of the blade model; Φ z is the rotation speed of the blade model around the z direction, Φ y is the rotation speed of the blade model around the y direction, Φ x is the rotation speed of the blade model around the x direction, u, w, v are the deformation displacements of any point of the blade model, O-XYZ is the global coordinate system, and x, y, z are the directions of the coordinate system.
[0081] Next, based on the strain energy and kinetic energy of the blade model and the work done by the centrifugal force, an energy functional of the blade model is obtained;
[0082] The expression of the centrifugal work done by the blade model is:
[0083] W = ∫∫∫ Ω N lx u ,x +N ly u ,y +N lz u ,z u ,z dΩ
[0084] Based on the obtained strain energy and kinetic energy of the blade model and the work done by centrifugal force, the energy functional of the blade model is obtained: Π=U s +T s -W
[0085] Where N lx , N ly ,,N lz is the centrifugal force component along the x, y, and z directions, u ,x ,u ,y ,u ,z is the derivative of the displacement variable along the x, y, and z directions.
[0086] Finally, based on the energy functional and vibration control equation of the blade model, the objective function value ω is obtained. 优化 ;
[0087] The vibration control equation is:
[0088]
[0089] Where M is the mass matrix, K is the stiffness matrix, and M G is the stress matrix, M cis the centrifugal force weakening matrix, q is the vibration displacement;
[0090] By solving the vibration control equation, the natural frequency ω and vibration mode of the blade model under the rotating condition can be obtained, thereby obtaining the objective function value ω to be optimized. 优化 .
[0091] Step 5, obtaining a blade model fitness function value based on the objective function value and the constraint conditions of material distribution optimization;
[0092] Set constraints for material distribution optimization, including equality constraints and inequality constraints;
[0093] Based on the penalty function method, the constraint conditions of the material distribution optimization and the objective function value, the blade model fitness function value f is obtained. 适应度 (V c ), the formula is:
[0094]
[0095] In the formula, ω 优化 is the target parameter value; λ is the penalty parameter, which is usually 10 5 ; nin and neq are the number of inequality constraints and equality constraints respectively; h r and h s are the rth and sth inequality constraints and equality constraints respectively; ε is the allowable error of the equality constraint, which is generally 10 -4 .
[0096] Step 6, using the optimization algorithm program to iteratively solve, calculate and compare the fitness function value of the blade model, obtain the best design variable individual solution and the best material distribution optimized blade model, and complete the material distribution optimization of the blade model.
[0097] Predetermine the number of iterations, set the basic parameters required by the material distribution optimization algorithm, and enter the symbiotic organism search optimization algorithm program:
[0098] Enter the mutualistic symbiosis stage. Select the two design variables in the leaf model fitness function value. Determine the co-occurrence vector of the individual solutions of two design variables Update individual solutions of design variables according to mutualistic relationships If the fitness function values of the two updated design variable individual solutions are better than the two design variable individual solutions before the update, the new design variable individual solution is adopted.
[0099] Enter the phase of partial benefit symbiosis. Select a design variable individual solution In the design variable individual solution With the help of If the updated design variable individual solution The fitness function value is better than the individual solution of the design variables before updating Then keep the updated design variable individual solution Otherwise, the original individual solutions of the design variables are retained.
[0100] Enter the parasitic stage. Select a design variable individual solution according to Generation of parasite vectors If parasite vector The fitness function value is better than the individual solution of the design variable Then keep the parasite vector Otherwise, the original individual solutions of the design variables are retained.
[0101] Save the optimal fitness function value and the optimal design variable individual solution in the three stages of mutualism, partial mutualism and parasitism, and judge whether the predetermined number of iterations has been reached. If not, return to the mutualism stage and enter the next cycle; if it has been reached, output the optimal objective function value ω 优化 And the optimal design variable individual solution, the optimized blade model is obtained, and the material distribution optimization of the blade model is completed.
[0102] The optimization analysis results are shown in Table 3, which shows the ceramic volume fraction (design variable) and the first-order natural frequency (optimization target) obtained at each design control point of the rotating blade under different conditions using this method.
[0103] Table 3
[0104]
[0105] The above shows and describes the basic principles, main features and advantages of the present invention. It should be understood by those skilled in the art that the present invention is not limited to the above embodiments. The above embodiments and descriptions are only for explaining the principles of the present invention. Without departing from the spirit and scope of the present invention, the present invention may have various changes and improvements, which fall within the scope of the present invention. The scope of protection of the present invention is defined by the attached claims and their equivalents.
Claims
1. A material distribution optimization method for a blade model, characterized in that: The following steps are involved: Step 1, construct a blade model and extract basic information of the blade model; Step 2: Based on the basic information of the blade model and the refinement strategy of the isogeometric analysis method, a non-uniform rational B-spline basis function is constructed. The formula of the non-uniform rational B-spline basis function is: Among them, ω i,j is the weight information, p and q are the order information, ζ,η are the node vector information, N i,p (ζ),M i,q (η) represents the univariate B-spline function along the two knot vector directions ζ and η, respectively, m and n are the number of control points in the two knot vector directions, and i and j are the i-th and j-th control points in the two knot vector directions; Step 3, using the non-uniform rational B-spline basis function to perform interpolation calculation to obtain the material properties of the blade model. The formula of the material properties of the blade model is: P=P me +(P ce -P me )V c Among them, P is the material parameter of the blade model, P me ,P ce is the material parameter of blade models with different materials, V c is the material property of the blade model, a is the ath control point, is the material property at the control point of the blade model, m and n are the number of control points in the direction of the two node vectors ζ and η, is the non-uniform rational B-spline basis function; Step 4: Based on the material properties of the blade model, the vibration characteristics of the blade model are analyzed to obtain an objective function value. The objective function value acquisition process includes: The strain energy and kinetic energy of the blade model are obtained based on the first-order shell theory; The expression of the first-order shell theory is: u(x,z)={u(x,z)w(x,z)v(x,z)} T w(x,z)=w0(x,z) Where u, w, v are the deformation displacements of any point on the blade model, u0, v0, w0 are the in-plane displacements of a point on the neutral plane, O-XYZ is the global coordinate system, x, z are the directions of the coordinate system, is the rotation angle of a point in the neutral plane along the x-axis and z-axis, ζ,η are the node vector information, m,n are the number of control points in the two node vector directions, p and q are the order information, and u a is the deformation displacement at the ath control point; Based on the strain energy and kinetic energy of the blade model and the work done by the centrifugal force, the energy functional of the blade model is obtained; the formula for the work done by the centrifugal force of the blade model is: W=∫∫∫ Ω N lx u ,x +N ly u ,y +N lz u ,z dΩ Where N lx ,N ly ,N lz is the centrifugal force component along the x, y, and z directions, u ,x ,u ,y ,u ,z is the derivative of the displacement variable along the x, y, and z directions; Obtaining an objective function value based on the energy functional and vibration control equation of the blade model; The vibration control equation is: Where M is the mass matrix, K is the stiffness matrix, and M G is the stress matrix, M c is the centrifugal force weakening matrix, q is the vibration displacement; Step 5: Based on the objective function value and the constraint conditions of material distribution optimization, the blade model fitness function value is obtained. The process of obtaining the blade model fitness function value includes: Setting constraint conditions for the material distribution optimization, including: equality constraints and inequality constraints; Obtaining a blade model fitness function value based on a penalty function method, the constraint conditions of the material distribution optimization, and the objective function value; Step 6, using the optimization algorithm program to iteratively solve, calculate and compare the fitness function value of the blade model, obtain the best design variable individual solution and the best material distribution optimized blade model, and complete the material distribution optimization of the blade model.
