A method and device for generating stiffness matrix of each connection structure unit in whole machine assembly
By using the connection factor and punishment factor correction coefficient matrix in the finite element analysis of aviation structures, and using the augmented Lagrangian theory to deal with multi-point constraints, the sparsity damage caused by multi-point constraints is solved, and the simulation analysis efficiency is improved , meets the needs of refined simulation of aviation equipment.
Patent Information
- Application Number
- CN202411085061.2
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-08
- Publication Date
- 2024-11-12
- Estimated Expiration
- 2044-08-08
AI Technical Summary
In the finite element analysis of aviation structures, multi-point constraints lead to the sparseness failure of the overall stiffness matrix of the structural structure, affecting the efficiency of system of equations, especially when large-scale matrix processing is processed, existing solvers cannot complete effective solutions.
By determining the connection factor and punishment factor, the coefficient matrix constructed by the degrees of freedom of the active node and the slave node is corrected, and the multi-point constraint relationship is handled by the augmented Lagrangian theory to maintain the sparseness of the matrix, avoiding the stiffness of the matrix, and then Improve simulation analysis efficiency.
It effectively maintains the sparsity of the structural stiffness matrix, improves the efficiency of finite element simulation calculation, and can effectively solve it in large-scale matrices to meet the needs of refined simulation analysis of aviation equipment.
Smart Images

Figure CN118607332B_ABST
Abstract
Description
Technical Field
[0001] The present application belongs to the technical field of structural simulation, and in particular relates to a method and device for generating a stiffness matrix of each connecting structural unit in the assembly of a whole machine. Background Art
[0002] Aviation structures are complex and have many parts. Various connection structures are used in the assembly of the whole machine, including rivets, bolts, joints, etc. In traditional finite element modeling, these connection structures are processed into common node connections or a small number of rods and beams. However, as the development of my country's aviation equipment has entered a high-speed and refined process, the structural design and optimization process involves the fine finite element modeling and simulation analysis of the whole machine. The simulation of the connection structure can no longer be ignored or greatly simplified as before. Instead, large-scale MPC, RBE2, RBE3 and other multi-point constraints are used to achieve the connection between structural components and the distribution of aerodynamic loads.
[0003] In the process of finite element analysis, the overall stiffness matrix of the structure will be obtained. Since the nodes are processed according to the principle of proximity during the numbering process, the overall stiffness matrix will show the characteristics of sparsity, that is, the non-zero elements of the matrix are concentrated in the strip area centered on the main diagonal, such as Figure 1 Specifically, for a simple structure, without considering multi-point constraints, the overall stiffness matrix of the structure is as follows: Figure 2 The 12×12 matrix shown is for intuitive description. Figure 2 In , non-zero terms are represented as the value 1, and zero terms are left blank. Therefore, when solving the equilibrium equations, large sparse matrix solving algorithms can be used to greatly improve the efficiency of solving the linear equation system.
[0004] However, in the process of dealing with multi-point constraints, traditional structural analysis software mostly uses linear elimination theory, which eliminates the degrees of freedom of the slave point from the solution set by expressing the displacement of the slave point in the form of a relational expression of the active point, thereby achieving the condensation and reduction of the overall stiffness matrix of the structure and completing the solution of the system of equations. However, in the process of linear elimination, as the order of the matrix is reduced, several non-zero terms will be added to the non-diagonal terms of the overall stiffness matrix of the structure. When there are large-scale multi-point constraints in the structure, the sharp increase in the number of non-zero terms will seriously damage the sparsity of the overall stiffness matrix of the structure, making it impossible to use large sparse matrix solution algorithms, affecting the efficiency of solving the system of equations. For example, Figure 2 The 12×12 matrix is processed as Figure 3The dense matrix of 8×8 is shown in the figure. At the same time, the process of using active points to represent the displacement of the slave points involves the inversion calculation of large matrices, which will greatly prolong the time of finite element analysis, especially when the matrix reaches hundreds of millions, the current solver cannot complete the corresponding solution. Therefore, a large-scale multi-point constraint processing algorithm is urgently needed to solve the problems encountered in the current autonomous aviation structure analysis software, so as to meet the needs of refined simulation analysis of aviation equipment. Summary of the invention
[0005] In order to solve at least one of the above technical problems, the present application provides a method and device for generating stiffness matrices of each connection structure unit in the whole machine assembly to improve the simulation efficiency.
[0006] In a first aspect, the present application provides a method for generating a stiffness matrix of each connection structure unit in a complete machine assembly, which is applied to a connection structure constrained by multiple nodes, and the method comprises:
[0007] Step S1, determining the distance difference between the active node and the passive node of the connection unit in each degree of freedom direction, wherein the connection unit refers to a connection structure in the whole machine assembly represented by multiple nodes, multiple degrees of freedom and a unit stiffness matrix in the finite element analysis process;
[0008] Step S2, determining a connection factor according to each distance difference exceeding half of the maximum distance difference and the number of degrees of freedom corresponding thereto, and at the same time, determining a penalty factor according to all distance differences and the total number of degrees of freedom corresponding thereto;
[0009] Step S3: The coefficient matrix A constructed by the number of degrees of freedom of the active nodes and the passive nodes is modified based on the connection factor and the penalty factor to obtain the unit stiffness matrix K.
[0010] Preferably, in step S2, the connection factor k is determined by the following formula:
[0011] ;
[0012] in, k 0 is the connection balance coefficient, Ndof is the number of degrees of freedom corresponding to each distance difference exceeding half of the maximum distance difference, L dof for each distance difference exceeding half of the maximum distance difference;
[0013] The penalty factor is determined by the following formula :
[0014] ;
[0015] in, is the penalty balance coefficient, Nd is the number of degrees of freedom corresponding to all distance differences, L d is the distance difference corresponding to each degree of freedom.
[0016] Preferably, both the connection balance coefficient and the penalty balance coefficient are set to 1.0e+5.
[0017] Preferably, in step S3, the unit stiffness matrix K is obtained by the following formula:
[0018] .
[0019] Preferably, step S3 further includes:
[0020] Step S4: assemble the stiffness matrix K of each unit into the overall stiffness matrix of the structure according to the node table, solve it through a numerical solver, and obtain the simulation analysis result.
[0021] A second aspect of the present application provides a device for generating a stiffness matrix of each connection structure unit in a complete machine assembly, which is applied to a connection structure constrained by multiple nodes, and the device comprises:
[0022] A master-slave node distance difference statistics module is used to determine the distance difference between the active node and the slave node of the connection unit in each degree of freedom direction, wherein the connection unit refers to the connection structure in the whole machine assembly represented by multiple nodes, multiple degrees of freedom and unit stiffness matrix in the finite element analysis process;
[0023] A matrix augmentation parameter calculation module is used to determine the connection factor according to each distance difference exceeding half of the maximum distance difference and the corresponding number of degrees of freedom, and at the same time, determine the penalty factor according to all distance differences and the corresponding total number of degrees of freedom;
[0024] The unit stiffness matrix generation module is used to modify the coefficient matrix A constructed by the number of degrees of freedom of the active nodes and the driven nodes based on the connection factor and the penalty factor to obtain the unit stiffness matrix K.
[0025] Preferably, the matrix augmentation parameter calculation module includes:
[0026] The connection factor calculation unit is used to determine the connection factor k by the following formula:
[0027] ;
[0028] in, k 0 is the connection balance coefficient, Ndof is the number of degrees of freedom corresponding to each distance difference exceeding half of the maximum distance difference, L doffor each distance difference exceeding half of the maximum distance difference;
[0029] The penalty factor calculation unit is used to determine the penalty factor using the following formula :
[0030] ;
[0031] in, is the penalty balance coefficient, Nd is the number of degrees of freedom corresponding to all distance differences, L d is the distance difference corresponding to each degree of freedom.
[0032] Preferably, both the connection balance coefficient and the penalty balance coefficient are set to 1.0e+5.
[0033] Preferably, in the unit stiffness matrix generation module, the unit stiffness matrix K is obtained by the following formula:
[0034] .
[0035] Preferably, the device further comprises:
[0036] The overall stiffness matrix assembly module is used to assemble the stiffness matrix K of each unit into the overall stiffness matrix of the structure according to the node table, and solve it through the numerical solver to obtain the simulation analysis results.
[0037] The present application can ensure the sparsity of the stiffness matrix during the finite element simulation calculation process, thereby improving the simulation operation efficiency. BRIEF DESCRIPTION OF THE DRAWINGS
[0038] Figure 1 It is a schematic diagram of the sparse stiffness matrix.
[0039] Figure 2 Schematic diagram of the 12×12 stiffness matrix without considering multi-point constraints.
[0040] Figure 3 For Figure 2 The stiffness matrix shown is a schematic diagram of an 8×8 dense matrix formed by linear elimination.
[0041] Figure 4 It is a flow chart of a preferred implementation method of the method for generating the stiffness matrix of each connecting structure unit in the whole machine assembly of the present application.
[0042] Figure 5 This application Figure 4 A schematic diagram of the stiffness matrix composition of the illustrated embodiment.
[0043] Figure 6 It is for Figure 2The stiffness matrix shown is a schematic diagram of a 16×16 sparse matrix formed by applying the augmented Lagrangian theory. DETAILED DESCRIPTION
[0044] In order to make the purpose, technical scheme and advantages of the implementation of this application clearer, the technical scheme in the implementation of this application will be described in more detail in combination with the drawings in the implementation of this application. In the drawings, the same or similar reference numerals throughout represent the same or similar elements or elements with the same or similar functions. The described implementation is a part of the implementation of this application, not all of the implementations. The implementation described below with reference to the drawings is exemplary and is intended to be used to explain this application, and cannot be understood as a limitation on this application. Based on the implementation in this application, all other implementations obtained by ordinary technicians in this field without making creative work are within the scope of protection of this application. The implementation of this application is described in detail below in combination with the drawings.
[0045] The first aspect of the present application provides a method for generating a stiffness matrix of each connection structure unit in a complete machine assembly, which is applied to a connection structure subject to multi-node constraints, such as Figure 4 As shown, the method mainly includes:
[0046] Step S1, determining the distance difference between the active node and the passive node of the connection unit in each degree of freedom direction, wherein the connection unit refers to a connection structure in the whole machine assembly represented by multiple nodes, multiple degrees of freedom and a unit stiffness matrix in the finite element analysis process;
[0047] Step S2, determining a connection factor according to each distance difference exceeding half of the maximum distance difference and the number of degrees of freedom corresponding thereto, and at the same time, determining a penalty factor according to all distance differences and the total number of degrees of freedom corresponding thereto;
[0048] Step S3: The coefficient matrix A constructed by the number of degrees of freedom of the active nodes and the passive nodes is modified based on the connection factor and the penalty factor to obtain the unit stiffness matrix K.
[0049] In some optional embodiments, the connection factor is used to balance the stiffness of the connection unit and the stiffness of other structural units, and its value is related to the relationship between the active point and the passive point. Accordingly, in step S2, the connection factor k is determined by the following formula:
[0050] ;
[0051] in, k 0 is the connection balance coefficient, Ndof is the number of degrees of freedom corresponding to each distance difference exceeding half of the maximum distance difference, L doffor each distance difference exceeding half of the maximum distance difference;
[0052] The penalty factor is used to avoid the matrix from being too rigid. The penalty factor can be determined by the following formula :
[0053] ;
[0054] in, is the penalty balance coefficient, Nd is the number of degrees of freedom corresponding to all distance differences, L d is the distance difference corresponding to each degree of freedom.
[0055] In some optional implementations, the connection balance coefficient and the penalty balance coefficient are both set to 1.0e+5.
[0056] In some optional implementations, in step S3, the unit stiffness matrix K is obtained by the following formula:
[0057] .
[0058] The principle and composition of the unit stiffness matrix K are described below.
[0059] The multi-node constraint relationship can be expressed as:
[0060] Au-b=0, where u is the displacement vector of the node, A is the coefficient matrix, and b is the constant vector.
[0061] For example, when there are constraints between m slave points and n active points, the multi-node constraint relationship is written in matrix form as:
[0062] .
[0063] in, R mn , R mm are represented as coefficient matrices corresponding to active points and passive points respectively, u n , u m Represent the displacement vectors of the active point and the passive point respectively.
[0064] In the background technology description, through the linear elimination theory, the degrees of freedom of the driven node need to be expressed in the form of the degrees of freedom of the active node, so the following operations need to be performed:
[0065] .
[0066] This involves the matrix Rmm If there are many multi-point constraints, this process will take a lot of time to perform the inverse operation.
[0067] In this application, according to the variational principle, the differential equations and boundary conditions of the finite element solution problem are equivalent to the variation of the corresponding functional being equal to 0, that is, the functional takes a stationary value. Therefore, the original problem of the equilibrium equation is expressed as:
[0068] .
[0069] Where K is the stiffness matrix.
[0070] According to the augmented Lagrangian theory, by introducing constraints into the functional, we can obtain:
[0071] .
[0072] For the connection structure with multiple point constraints, it is assumed to be a "connection unit" similar to the structural unit, which has nodes (consisting of master and slave nodes), degrees of freedom (consisting of master and slave degrees of freedom) and the corresponding unit stiffness matrix. For the connection unit, the first term of the above formula is 0; the second term is the Lagrange multiplier term, where λ is the Lagrange multiplier and k is the connection factor, which is the coefficient of the unit stiffness matrix of the balanced rigid element and the unit stiffness matrix of other units, so that they are basically consistent in order of magnitude, which helps to avoid the appearance of a numerically ill-conditioned stiffness matrix; the third term is the penalty function term, is a penalty factor used to avoid the matrix being too stiff.
[0073] For a single connected unit, take a variation of the functional and get:
[0074] .
[0075] in,
[0076] .
[0077] Taking the second variation of the functional, we get:
[0078] .
[0079] Referring to the functional representation of the structural unit, the unit stiffness matrix corresponding to the connection unit can be written as:
[0080] .
[0081] Among them, K L The first term is the linear stiffness matrix, K N The second term is the differential stiffness matrix.
[0082] For linear multi-point constraint relationships, the H matrix is the coefficient matrix A, and the differential stiffness matrix K in linear static analysis N = 0. The connection unit stiffness matrix at this time is expressed as:
[0083] .
[0084] in, It is a square matrix with the order of the number of degrees of freedom of the active points and the driven points, and the number of rows of kA is the number of Lagrangian degrees of freedom. Therefore, the order of the element stiffness matrix of the connection element is determined by the sum of the number of degrees of freedom of the active nodes, the driven nodes and the number of Lagrangian degrees of freedom.
[0085] In some optional implementations, step S3 further includes:
[0086] Step S4: assemble the stiffness matrix K of each unit into the overall stiffness matrix of the structure according to the node table, solve it through a numerical solver, and obtain the simulation analysis result.
[0087] Since the overall structure includes various connection units, the overall stiffness matrix can be formed by combining multiple connection unit stiffness matrices according to the node table, and then input into the numerical solver to obtain the simulation analysis results. Figure 5 As shown in the figure, the connection unit stiffness matrix is assembled into the overall stiffness matrix of the structure according to the node table. The matrix sparsity of the conventional unit degree of freedom is still retained, and the items in the connection unit stiffness matrix are placed in the rows and columns corresponding to the Lagrange multiplier degree of freedom. Therefore, the augmented Lagrangian theory is essentially a matrix expansion method. Figure 2 Given a 12×12 stiffness matrix, the stiffness matrix obtained by the augmented Lagrangian theory of this application is in the form of a 16×16 sparse matrix, such as Figure 6 shown.
[0088] This application increases the order of the overall stiffness matrix of the structure after processing by augmenting the Lagrangian theory, but retains the sparsity of the original structural stiffness matrix. The sparse matrix solving algorithm can be called to solve the equation group, which greatly improves the simulation operation efficiency.
[0089] In a second aspect, the present application provides a device for generating stiffness matrices of connection structure units in whole machine assembly corresponding to the above method, which is applied to a connection structure constrained by multiple nodes, and the device comprises:
[0090] A master-slave node distance difference statistics module is used to determine the distance difference between the active node and the slave node of the connection unit in each degree of freedom direction, wherein the connection unit refers to the connection structure in the whole machine assembly represented by multiple nodes, multiple degrees of freedom and unit stiffness matrix in the finite element analysis process;
[0091] A matrix augmentation parameter calculation module is used to determine the connection factor according to each distance difference exceeding half of the maximum distance difference and the corresponding number of degrees of freedom, and at the same time, determine the penalty factor according to all distance differences and the corresponding total number of degrees of freedom;
[0092] The unit stiffness matrix generation module is used to modify the coefficient matrix A constructed by the number of degrees of freedom of the active nodes and the driven nodes based on the connection factor and the penalty factor to obtain the unit stiffness matrix K.
[0093] In some optional implementations, the matrix augmentation parameter calculation module includes:
[0094] The connection factor calculation unit is used to determine the connection factor k by the following formula:
[0095] ;
[0096] in, k 0 is the connection balance coefficient, Ndof is the number of degrees of freedom corresponding to each distance difference exceeding half of the maximum distance difference, L dof for each distance difference exceeding half of the maximum distance difference;
[0097] The penalty factor calculation unit is used to determine the penalty factor using the following formula :
[0098] ;
[0099] in, is the penalty balance coefficient, Nd is the number of degrees of freedom corresponding to all distance differences, L d is the distance difference corresponding to each degree of freedom.
[0100] In some optional implementations, the connection balance coefficient and the penalty balance coefficient are both set to 1.0e+5.
[0101] In some optional implementations, in the unit stiffness matrix generation module, the unit stiffness matrix K is obtained by the following formula:
[0102] .
[0103] In some optional embodiments, the device further comprises:
[0104] The overall stiffness matrix assembly module is used to assemble the stiffness matrix K of each unit into the overall stiffness matrix of the structure according to the node table, and solve it through the numerical solver to obtain the simulation analysis results.
[0105] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any changes or substitutions that can be easily thought of by a person skilled in the art within the technical scope disclosed in the present application should be included in the protection scope of the present application. Therefore, the protection scope of the present application shall be based on the protection scope of the claims.
Claims
1. A method for generating stiffness matrices of connection structure units in whole machine assembly, characterized in that: Applied to a connection structure subject to multi-node constraints, the method comprises: Step S1, determining the distance difference between the active node and the passive node of the connection unit in each degree of freedom direction, wherein the connection unit refers to a connection structure in the whole machine assembly represented by multiple nodes, multiple degrees of freedom and a unit stiffness matrix in the finite element analysis process; Step S2, determining a connection factor according to each distance difference exceeding half of the maximum distance difference and the number of degrees of freedom corresponding thereto, and at the same time, determining a penalty factor according to all distance differences and the total number of degrees of freedom corresponding thereto; Step S3, the coefficient matrix A constructed by the number of degrees of freedom of the active node and the passive node is modified based on the connection factor and the penalty factor to obtain the unit stiffness matrix K; In step S2, the connection factor k is determined by the following formula: Among them, k0 is the connection balance coefficient, Ndof is the number of degrees of freedom corresponding to each distance difference exceeding half of the maximum distance difference, and L dof for each distance difference exceeding half of the maximum distance difference; The penalty factor σ is determined by the following formula: Among them, σ0 is the penalty balance coefficient, Nd is the number of degrees of freedom corresponding to all distance differences, and L d is the distance difference corresponding to each degree of freedom.
2. The method for generating the stiffness matrix of each connecting structure unit in the whole machine assembly according to claim 1, characterized in that: The connection balance coefficient and penalty balance coefficient are both set to 1.0e+5.
3. The method for generating the stiffness matrix of each connecting structure unit in the whole machine assembly according to claim 1, characterized in that: In step S3, the element stiffness matrix K is obtained by the following formula:
4. The method for generating the stiffness matrix of each connecting structure unit in the whole machine assembly according to claim 1, characterized in that: Step S3 further includes: Step S4: assemble the stiffness matrix K of each unit into the overall stiffness matrix of the structure according to the node table, solve it through a numerical solver, and obtain the simulation analysis result.
5. A device for generating stiffness matrix of each connecting structure unit in whole machine assembly, characterized in that: Applied to a connection structure subject to multi-node constraints, the device comprises: A master-slave node distance difference statistics module is used to determine the distance difference between the active node and the slave node of the connection unit in each degree of freedom direction, wherein the connection unit refers to the connection structure in the whole machine assembly represented by multiple nodes, multiple degrees of freedom and unit stiffness matrix in the finite element analysis process; A matrix augmentation parameter calculation module is used to determine the connection factor according to each distance difference exceeding half of the maximum distance difference and the corresponding number of degrees of freedom, and at the same time, determine the penalty factor according to all distance differences and the corresponding total number of degrees of freedom; The unit stiffness matrix generation module is used to modify the coefficient matrix A constructed by the number of degrees of freedom of the active node and the driven node based on the connection factor and the penalty factor to obtain the unit stiffness matrix K; Wherein, the matrix augmentation parameter calculation module includes: The connection factor calculation unit is used to determine the connection factor k by the following formula: Among them, k0 is the connection balance coefficient, Ndof is the number of degrees of freedom corresponding to each distance difference exceeding half of the maximum distance difference, and L dof for each distance difference exceeding half of the maximum distance difference; The penalty factor calculation unit is used to determine the penalty factor σ by the following formula: Among them, σ0 is the penalty balance coefficient, Nd is the number of degrees of freedom corresponding to all distance differences, and L d is the distance difference corresponding to each degree of freedom.
6. The device for generating stiffness matrix of each connecting structural unit in whole machine assembly according to claim 5, characterized in that: The connection balance coefficient and penalty balance coefficient are both set to 1.0e+5.
7. The device for generating stiffness matrix of each connecting structure unit in whole machine assembly according to claim 5, characterized in that: In the element stiffness matrix generation module, the element stiffness matrix K is obtained by the following formula:
8. The device for generating stiffness matrix of each connecting structure unit in whole machine assembly according to claim 5, characterized in that: The device also includes: The overall stiffness matrix assembly module is used to assemble the stiffness matrix K of each unit into the overall stiffness matrix of the structure according to the node table, and solve it through the numerical solver to obtain the simulation analysis results.
