An integral design method for tubular components

By using B-spline and NURBS curve generation methods, combined with linear equation system constraints, the overall representation and smoothness problems of tubular components in computer-aided design software were solved, achieving efficient and accurate geometric design.

CN117010097BActive Publication Date: 2026-11-17NANJING UNIV OF SCI & TECH
View PDF 2 Cites 0 Cited by

Patent Information

Application Number
CN202310563936.4
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2023-05-18
Publication Date
2026-11-17
Estimated Expiration
2043-05-18

AI Technical Summary

Technical Problem

Existing computer-aided design software cannot represent tubular components as a whole, resulting in low design efficiency and insufficient smoothness, which affects subsequent numerical simulation and processing path planning.

Method used

By employing B-spline and NURBS curve generation methods, closed curves of the inlet and outlet sections of tubular components are generated. Linear equation constraints are then used to achieve overall smoothness of the curves and surfaces, avoiding surface splicing and reducing computational complexity.

Benefits of technology

It achieves a smooth representation of tubular components, improves design efficiency and accuracy, simplifies the calculation process, and enhances the effectiveness of numerical simulation and machining path planning.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN117010097B_ABST
    Figure CN117010097B_ABST
Patent Text Reader

Abstract

This invention provides a method for the overall design of tubular components. The method includes constructing a tubular surface using B-spline surfaces based on modified basis functions, avoiding the use of spliced ​​surfaces; and forming a circle using NURBS curves based on modified basis functions, applying linear equation constraints to the first and last intervals, again avoiding curve splicing, resulting in a smooth curve overall, while simplifying program implementation and reducing computational requirements. This invention only requires providing the node vectors and degrees of the basis functions in two directions to generate an initial tubular surface. The local design of the surface is then modified by adjusting the control point mesh in Euclidean space, achieving the overall design of the tubular surface.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention belongs to the field of geometric design, and specifically relates to an overall design method for tubular components. Background Technology

[0002] Many engineering fields have significant design requirements for tubular components, especially in aircraft engine manufacturing, where the air intake is a crucial example. Air intakes play a vital role in ensuring sufficient power for aircraft engines, long-term safe and normal operation, and meeting specific aircraft performance requirements (stealth, supersonic speed, etc.), often exhibiting complex and varied shapes. In engineering design, conic sections, such as circles, hold an important position and are quite common. For example, an aircraft air intake duct with a circular cross-section at one end places high demands on the expressive power and precision of the design methodology.

[0003] In modern times, traditional geometric design has been replaced by computer-aided design (CAD) software. However, current CAD software, when representing tubular components, generally uses standard B-spline surfaces or NURBS surfaces, employing a method of splicing multiple surfaces to form the tubular surface. This method presents several problems:

[0004] 1. Tubular components cannot be represented as a whole and must be spliced ​​together, which causes many inconveniences in the design and reduces the efficiency of the design.

[0005] 2. Overall smoothness cannot be guaranteed, especially at the joints of curved surfaces. Additional constraints on the control point network are usually required to meet the smoothness requirements of the curved surface. This not only introduces extra work but also affects the accuracy of the design.

[0006] The aforementioned problems have negatively impacted subsequent numerical simulations and fabrication path planning for tubular components. Consequently, current computer-aided design software methods are not ideal for the geometric design of tubular components. Summary of the Invention

[0007] The purpose of this invention is to provide an overall design technology for tubular components, so as to solve the problem that the underlying representation of current computer-aided design software cannot represent such tubular components as a whole and cannot guarantee the overall smoothness.

[0008] The technical solution to achieve the purpose of this invention is as follows:

[0009] A method for the overall design of a tubular component includes the following steps:

[0010] Step 1: Generate closed curves for the inlet and outlet sections of the tubular component: including non-circular closed curves or circular closed curves;

[0011] Input: The degree k of the generated closed curve and the node vector U; Output: The coordinates of the control points of the closed curve;

[0012] Calculate the expressions for the 0th, ..., kth basis functions and the (nk-1), ..., n-1th basis functions, and express them as column vectors respectively;

[0013] A. Using B-splines to generate non-circular closed curves:

[0014] The column vectors of each basis function expression are horizontally concatenated to form the coefficient matrix of the homogeneous linear equation system; the system of linear equations is solved to obtain the general solution; when the coordinates of a certain control point are adjusted, the undetermined coefficients in the general solution are changed according to the change in coordinates, thereby calculating the updated coordinates of other control points, thus completing the coordinated control of control points; given the coordinates of the control points, the non-circular closed curve can be calculated according to the definition of B-spline.

[0015] B. Using NURBS curves to generate circular closed curves

[0016] The column vectors of the (nk-1), ..., (n-1)th basis function expressions are horizontally concatenated to form the coefficient matrix of the non-homogeneous linear equation system, with each basis function corresponding to a weight factor. Each basis function expression's column vector is multiplied by its corresponding weight factor to obtain a new set of vectors. These vectors are then horizontally concatenated to form the coefficient matrix, which is a full-rank matrix. Known quantities in the equations are moved to the right side as constant terms of the non-homogeneous linear equation system. Solving this system yields a unique set of weight factors ω. i Given i = nk-1, ..., n-1, the coordinates of the control points can be uniquely solved; given the coordinates of the control points, the circular closed curve can be calculated according to the definition of NURBS splines.

[0017] Step 2: Generate tubular surface:

[0018] Input: the degree k2 in the S direction and the node vector U2; Output: points on the surface;

[0019] First, convert the coordinates of the control points in step 1 into three-dimensional coordinates and fix them. For each column of control points in the T direction, excluding the inlet and outlet sections, calculate the expressions of the 0th, ..., kth basis functions and the (nk-1), ..., n-1th basis functions respectively, represent them as column vectors, and concatenate them horizontally into a matrix to obtain the coefficient matrix of multiple homogeneous linear equation systems. Solve each homogeneous linear equation system to obtain multiple general solutions. When adjusting the coordinates of a certain control point in a certain column in the T direction, change the undetermined coefficients in the general solution of the homogeneous linear equation system of that column according to the change in coordinates, thereby calculating the updated coordinates of other control points in the same column, thus completing the coordinated control of the control points. Given the coordinates of the control points, the tubular surface can be calculated according to the definition of the surface, and then points can be selected on the surface with arbitrary precision and output.

[0020] The significant advantages of this invention compared to existing technologies are:

[0021] (1) The method of this invention, when a curve forms a circle, applies a set of equations to the first and last intervals to constrain the curve, without the need for splicing, so that the overall smoothness of the curve is good and can truly represent a circle.

[0022] (2) The constraints in the method are all linear equations, which facilitates the implementation of the program, reduces the computational requirements of the computer, and is highly efficient. Attached Figure Description

[0023] Figure 1 This is a flowchart illustrating the generation of the tubular surface in this invention.

[0024] Figure 2 The result diagram of generating B-spline closed curves (as inlet or outlet sections) for this invention.

[0025] Figure 3 This is a result diagram of the intermediate curved surface generated between the inlet and outlet sections for this invention.

[0026] Figure 4 This is a diagram showing the result of forming a circle from the NURBS curves of the present invention. Detailed Implementation

[0027] The present invention will be further described below with reference to the accompanying drawings and specific embodiments.

[0028] Combination Figures 1-4 This embodiment provides an overall design method for a tubular component, including:

[0029] Step 1: Generate a closed curve (inlet or outlet section of the tubular component) using B-splines or NURBS splines.

[0030] Closed curves include two types: A) non-circular closed curves, and B) circular closed curves.

[0031] A. Generating a non-circular closed curve using B-spline

[0032] A.1. Assume that a spline curve of degree k with n control points is required, and the node vector is

[0033] U=[u0,…,u k ,…,u e ,…,u n ,…,u n+k

[0034] where u0=…=u k =0, u n =…=u n+k =1, u e is the e-th node.

[0035] When k<e<n, 0<u e <1;

[0036] The basis functions related to the interval [u k ,u k+1 ) are N i,k (u)(i=0,…,k), and the basis functions related to the interval [u n-1 ,u n ) are N i,k (u)(i=n-k-1,…,n-1)

[0037] where N i,k (u) is the i-th basis function of degree k.

[0038] The B-spline basis functions are as follows:

[0039]

[0040] where N i,p (u) is the i-th basis function of degree p.

[0041] The B-spline curve is as follows:

[0042]

[0043] where P i is the coordinate of the i-th control point (including the abscissa and ordinate).

[0044] A.2. Perform the following coordinate transformation on N i,k (u)(i=n-k-1,…,n-1)

[0045] ① When u k+1 -u k =u n -u n-1 , let u′=u+u​n-1 -u k (u∈[u k ,u k+1 ))

[0046] ②When u k+1 -u k ≠u n -u n-1 At that time, a stretching transformation is performed, making

[0047]

[0048]

[0049] Where, x i and y i These represent the x and y coordinates of the i-th control point of the curve, respectively.

[0050] A.3, because in the interval [u i ,u i+1 On the basis function N i,k (u) can be represented as

[0051] That is, this basis function is represented by a k+1 dimensional column vector:

[0052]

[0053] Among them, a i,k Let be the coefficient of the k-th term of the i-th basis function.

[0054] Taking the adjustment of the x-coordinate of any control point as an example, the constraint equation can be described as follows:

[0055]

[0056] This is a homogeneous linear system of equations with infinitely many solutions. The general solution has the following form:

[0057]

[0058] Among them, the k+1 z-dimensional vectors on the right side of the equal sign form a fundamental solution set of the homogeneous linear equation system, satisfying z = 2(k+1), that is, all vectors on both sides of the equal sign have the same dimension.

[0059] CP0,…,CP k There are k+1 undetermined coefficients, which are called control parameters. By adjusting these parameters, different solutions to the system of equations can be obtained from the basic solution set.

[0060] If you choose CP j(0≤j≤k) is used as the current control parameter, and the user attempts to move the h-th (0≤h≤k or nk-1≤h≤n-1) control point so that its x-coordinate changes from the original x... h Change to x h ′, that is, Δx=x h ′-x h Then update the control parameter CP. j For CP j ',satisfy

[0061]

[0062] Where c j,h The value in the basic solution set.

[0063] Right now

[0064]

[0065] Therefore, the x-coordinates of all control points should be updated as follows:

[0066]

[0067] Regarding the ordinates of the control points, we can obtain constraint equations that are formally identical to those of the abscissas. By performing the same steps, we can coordinately adjust the ordinates of all constrained control points.

[0068] Given the coordinates of the control points, the non-circular closed curve can be calculated according to the definition of B-spline.

[0069] Figure 2 A cubic B-spline closed curve containing 9 control points is shown.

[0070] B. Using NURBS curves to generate circular closed curves

[0071] B.1. First, define a family of rational B-splines of degree k with n control points, and define the nodal vectors.

[0072] U = [u0, ..., u k ,…,u e ,…u n-k-1 ,…,u n-1 ,u n ,…,u n+k ]

[0073] Where u0 = … = u k =0, u n-k-1 =…=u n-1 =1, u n =…=u n+k .

[0074] ue is the e-th node. When k<e<n-k-1, 0<u e <1;

[0075] X, Y, ω are used to represent the coordinate axes of the three-dimensional space. Three-dimensional homogeneous coordinates are used to represent weighted control points

[0076] D i =[ω i x i , ω i y i , ω i , when u∈[u k ,u k+1 ) and u∈[u n-1 ,u n ), when the curve is projected onto the plane of ω=1, the two segments of curves where u is located in the above two intervals respectively coincide. Given the curve, control points and corresponding weights on u∈[0,1), the curve is defined

[0077]

[0078] where u is the curve parameter, N i,k (u) is the i-th k-th degree B-spline basis function, d i is the i-th circular control point, x i and y i are the abscissa and ordinate of the control point respectively, ω i is the weight factor of the i-th control point, ω0=1, ω n-1 =1, θ is the central angle corresponding to the curve.

[0079] At this point, the defined curve has been closed end to end to form a circle.

[0080] B.2. To make the curve defined by equation (8) still a circle, the curve is written in the following form:

[0081]

[0082] Then, according to the local support property of the spline, it is obtained that:

[0083]

[0084] At this time, in order to ensure better smoothness of the circular curve, it is only necessary to make the latter part of the spline curve

[0085]

[0086] coincide with the defined circle. The condition is transformed into that u is located at [u k ,uk+1 ] and [u n-1 ,u n When the interval is within the same range, the denominators of the curves are the same, so it is only necessary to make the numerators equal.

[0087] B.3, involving [u] k ,u k+1 The basis functions of the interval are N. 0,k (u),…,N i,k (u),…N k,k (u), involving [u] n-1 ,u n The basis functions of the interval are N. n-k-1,k (u), ..., N n-1,k (u).

[0088] List the system of linear equations with constraints:

[0089]

[0090]

[0091]

[0092] Where (u+1) represents the interval from u0 to u n-1 The length of the interval, i.e., 1 = u n-1 -u0=1-0.

[0093] The condition that the coefficient of the homogeneous term is 0 is satisfied. Consider equation (14).

[0094] Since the curve, control point coordinates, and corresponding weight factors on u∈[0,1) are known, the system of equations is obtained as follows:

[0095]

[0096] Where a 0,0 ,a 0,1 ,…,a 0,k The meanings are as follows: A.3, b0, ..., b k It is a constant formed by combining the known quantities of the equation.

[0097] Based on the properties of B-spline curves, the matrix Since it is a full-rank matrix, ω can be uniquely solved. i (i = nk-1, ..., n-1).

[0098] The solution ω i Substituting (i = nk-1, ..., n-1) into equations (12) and (13), we similarly obtain information about x. i y iLinear equations (i = nk-1, ..., n-1).

[0099] At this point, equation (12) becomes:

[0100]

[0101] Where c0,…,c k It is a constant formed by combining the known quantities of the equation.

[0102] x can be solved i (i = nk-1, ..., n-1).

[0103] Equation (13) becomes:

[0104]

[0105] Where d0,…,d k It is a constant formed by combining the known quantities of the equation.

[0106] The solution y can be obtained i (i = nk-1, ..., n-1).

[0107] Given the coordinates of the control points, the complete circular closed curve can be calculated using p(u) defined in equation (8).

[0108] Step 2: Generate B-spline tubular surface (generate intermediate surface based on step 1)

[0109] 2.1 Given the degree k1 of the basis function in the direction T and the node vector

[0110]

[0111] The degree k2 of the basis function in the S direction, and the node vector

[0112]

[0113] There are a total of 4 parameters. In the node vector, n and m are the number of control points in the T direction and S direction, respectively. Here, the T direction and S direction refer to the two directions of the control point grid.

[0114] Note: The node vector U1 and the degree k1 should be consistent with the node vector U and degree k used in step 1.

[0115] 2.2 For any control point (x) on the inlet section generated in step 1 i ,y i ), convert it to (x i ,y i For any control point (x, 0) on the exit section generated in step 1,i ,y i ), convert it to (x i ,y i ,0), where l is the distance between the inlet section and the outlet section.

[0116] 2.3 For a given node vector U1 and U2, take the parameter u∈[u0,u... n+k ),v∈[u′0,u′ m+k Then, the B-spline surface is defined as follows:

[0117]

[0118] Where P is a point on the B-spline surface corresponding to parameters (u,v); P ij These are the coordinates of the control point on the control point grid, which has an index of i in the T direction and an index of j in the S direction. Let represent the i-th k1-th basis function in the direction T (i = 0, 1, ..., n-1); Let S represent the j-th k2-th basis function in the direction S (j = 0, 1, ..., m-1).

[0119] 2.4 To generate the tubular surface, a system of linear equations is applied to the B-spline surface generated above in the T direction, letting...

[0120]

[0121] Among them, if but

[0122] otherwise

[0123] Therefore, there is

[0124]

[0125] The constraint equation is quite complex. To simplify the constraint equation and strengthen the above constraints, we can make the following:

[0126]

[0127] This transforms the constraint equations into the case of generating the B-spline closed curve in step 1. Specifically, for the j-th column of control points in the control point grid direction T (with n control points), the constraint equations are:

[0128]

[0129]

[0130]

[0131] By following the exact same method as step A.3, the control point grid can be adjusted to change the shape of the tubular surface.

[0132] After adjusting the control point grid, calculate and output the points on the tubular surface with arbitrary precision according to equation (18).

[0133] The final surface effect is as follows Figure 3 As shown.

[0134] This concludes the design of the tubular curved surface. The next step is to produce the corresponding tubular components according to actual needs.

[0135] Step 3: Verify the differentiability of the circular B-spline curve.

[0136] We only need to examine the differentiability of the basis functions.

[0137] Let C(u) be the vector defined at nodes U = [0,...u]. k ,u k+1 A B-spline curve of degree k with n control points, [...,1]

[0138] The node vector U contains n+k+1 nodes.

[0139] make

[0140]

[0141] Differentiating it, we get:

[0142]

[0143]

[0144] In the definition of B-splines, the 0 / 0 type is defined as having a value of 0, therefore...

[0145]

[0146] In the above formula

[0147]

[0148] In the defined spline curve, the node multiplicity is defined as 1. Therefore, each time the spline curve is differentiated, the degree of the spline is decremented by 1. This continues until the derivative is reduced to the first degree, at which point further differentiation is not possible.

[0149] If this spline curve is composed of splines of degree 4 or higher and the node multiplicity is 1, then it is differentiable at least in degree 3.

Claims

1. A method for the overall design of a tubular component, characterized in that, Includes the following steps: Step 1: Generate closed curves for the inlet and outlet sections of the tubular component: including non-circular closed curves or circular closed curves; Input: The degree of the generated closed curve and node vectors Output: Coordinates of the control points of the closed curve; Calculate the first The basis functions and the first The expressions for the basis functions are represented as column vectors; A. Utilize Spline generation of non-circular closed curves: The column vectors of each basis function expression are horizontally concatenated to form the coefficient matrix of the homogeneous linear equation system; solving this linear equation system yields the general solution; when the coordinates of a control point are adjusted, the undetermined coefficients in the general solution are changed according to the change in coordinates, thereby calculating the updated coordinates of other control points, thus completing the coordinated control of the control points; given the coordinates of the control points, based on... The definition of a spline allows for the calculation of non-circular closed curves. B. Utilization Curve generates circular closed curve Horizontal splicing The column vectors of the basis function expressions serve as the coefficient matrix of the non-homogeneous linear equation system, with each basis function corresponding to a weight factor. Multiplying the column vector of each basis function expression by its corresponding weight factor yields a new set of vectors. These vectors are then horizontally concatenated to form a full-rank coefficient matrix. The known quantities in the equations are moved to the right-hand side and used as constant terms in the non-homogeneous linear equation system. Solving this system of equations yields a unique set of weight factors. The coordinates of the control points can be uniquely solved; given the coordinates of the control points, according to... The definition of a spline allows for the calculation of a circular closed curve; Step 2: Generate tubular surface: enter: Number of times in direction and node vectors Output: Points on the surface; First, convert the control point coordinates from step 1 to three-dimensional coordinates and fix them; for each column of control points in the T direction except for the inlet and outlet sections, calculate the first... The basis functions and the first The expressions for the basis functions are represented as column vectors and concatenated horizontally into a matrix, thus obtaining the coefficient matrix of multiple homogeneous linear equation systems. Solving each homogeneous linear equation system yields multiple general solutions. When adjusting... When the coordinates of a certain control point in a certain column are determined, the undetermined coefficients in the general solution of the homogeneous linear equation system of that column are changed according to the change in coordinates, thereby calculating the updated coordinates of other control points in the same column, thus completing the coordinated control of the control points; if the coordinates of the control points are known, the tubular surface can be calculated according to the definition of the surface, and then points can be selected on the surface with arbitrary precision and output.

2. The overall design method for tubular components according to claim 1, characterized in that, When generating non-circular closed curves, the constraint equations can be described as a homogeneous linear system of equations: in, Indicates the curve's first x-coordinates of control points or ordinate , For the first The coefficient of the term.

3. The overall design method for tubular components according to claim 1, characterized in that, When generating a circular closed curve, the non-homogeneous linear equation system with respect to the weight factors is as follows: in, For the first The coefficient of the second term, For the first Weight factors for each control point It is a constant formed by combining the known quantities of the equation.

4. The overall design method for tubular components according to claim 1, characterized in that, When generating a circular closed curve, the non-homogeneous linear equations about the x-coordinates of the control points are: in, For the first The coefficient of the second term, For the first Weight factors for each control point Indicates the curve's first The coordinates of the control points It is a constant formed by combining the known quantities of the equation.

5. The overall design method for tubular components according to claim 1, characterized in that, The multiple homogeneous linear equations in step 2 are as follows: in, It is on the control point grid. The index in the direction is ,exist The index in the direction is The coordinates of the control points; express Direction first indivual Secondary basis functions ; yes Parameters in direction; yes The parameters after coordinate transformation in the direction.

6. The overall design method for tubular components according to claim 1, characterized in that, This also includes step 3, verifying the differentiability of the circular B-spline curve: in, For defined in node vectors of Second-rate, control points spline curves, nodal vectors There is The node, For the first The coordinates of the control points For its Differentiate; In the defined spline curve, the node multiplicity is defined as 1. Thus, each time the spline curve is differentiated, the spline degree is reduced by 1; until the derivative is reached at the first degree, it is no longer possible to differentiate.

Citation Information

Patent Citations

  • Three-dimensional closed curved surface construction algorithm based on closed contour line

    CN110415352A

  • S-spline-based computer-aided and isogeometric analysis representation method and application

    CN115346021A