CN105956303B - It is a kind of to be fitted to each other face design method as the lathe of target to offset distortion inaccuracy - Google Patents

Abstract

The invention discloses a method for designing the joint surface of a machine tool assembly aiming at offsetting deformation errors. In CAD software, the structure is simplified according to the requirements of static analysis and the structural characteristics of the machine tool, and then the entity model of the machine tool is parametrically assembled; the simplification and parameter Import the solid model of the assembly into the finite element analysis software Ansys Workbench; use this software to create a container for automatic analysis and extraction of deformation, and orthogonalize the processing space position and grid division, and complete the static analysis and result preservation at the same time; then through Data learning obtains the actual deformation amount of the nominal joint surface of the machine tool assembly considering the deformation of the entire processing space; inverts the obtained actual deformation amount, and adds the inverted amount to the nominal joint surface as the actively designed machine tool assembly Joint surface: Based on the Monte Carlo method and the least square method, combined with data regression technology, the contour tolerance design of the assembly joint surface is completed.

Description

A Design Method of Machine Tool Assembly Joint Surface Aiming at Counteracting Deformation Error

technical field

The invention belongs to the field of mechanical manufacturing and design, and relates to a data learning-based method for designing a joint surface of a machine tool assembly, including Workbench automatic analysis and extraction deformation technology, joint surface active design technology and joint surface profile tolerance design technology, and specifically relates to a A design method for the joint surface of machine tool assembly aiming at offsetting the deformation error.

Background technique

Under the action of cutting force, clamping force, gravity and inertial force, CNC machine tools will produce obvious additional geometric deformation, thereby destroying the original mutual positional relationship of each component of the machine tool and generating additional processing errors, which is called deformation (force) error. Studies have pointed out that deformation errors account for 12% of all machine tool errors, and are the third largest error source leading to machine tool space errors. The world's top machine tool manufacturer Mitsui Seiki produced a precision machine tool 7CN with a positioning accuracy of 0.5um. It pointed out on its official website that the deformation error was effectively reduced by scraping and grinding the assembly joint surface; the internationally renowned machine tool manufacturer DIXI pointed out that its highest precision product DIXI270 passed more than The 500 hours of scraping and grinding also effectively reduced the deformation error and ensured the accuracy of the single axis and the error between the axes; my country's famous machine tool manufacturers, Kunming Machine Tool, Qinchuan Machine Tool, etc., will mark (-) on the machine tool parts drawings Or (+) as an additional requirement for straightness and flatness to reduce the influence of deformation errors on machining errors.

It is an important means for machine tool manufacturers to counteract the damage of geometric deformation to machining accuracy by scraping and grinding the assembly joint surface. Rely on experience, lack of scientific and reasonable quantitative technical solutions. Essentially, there is a lack of a design method for the contour shape of the part considering the deformation of the part.

In one of the few studies on the active design of the joint surface of the machine tool, Sun et al. of Dalian University of Technology used a multi-form fitting method to carry out a "pre-deformation" design on the joint surface of the guide rail to offset the error caused by the plastic deformation of the feed system. However, there is no active design of the assembly joint surface from the perspective of the whole machine tool; at the same time, this method is aimed at the sliding joint surface of the machine tool, and is not applicable to the fixed joint surface of the machine tool; and the polynomial fitting method is only suitable for building two-dimensional contours. 3D assembly joint surface contours are not feasible.

Contents of the invention

The purpose of the present invention is to propose a method for the lack of scientific, reasonable and quantitative method for designing the profile of the joint surface of the machine tool assembly in the current machine tool manufacturers, relying heavily on the experience of the fitter, and for the current situation that the current design of the joint surface can only address the defects of two-dimensional shapes. A method for designing the joint surface of machine tool assembly with the goal of offsetting deformation errors is provided. Based on the data learning method, the method considers the contour of the joint surface deformed in the entire machining space and its contour tolerance, which can greatly reduce the deformation error.

In order to achieve the above object, the technical scheme adopted in the present invention is:

A method for designing a joint surface of a machine tool assembly with the goal of offsetting deformation errors, comprising the following steps:

1) Simplify the solid model of the machine tool in the CAD software according to the static analysis requirements and the structural characteristics of the machine tool;

2) Parametrically assemble the solid model of the machine tool in the CAD software to obtain a parametric assembly model;

3) Use Ansys Workbench software to orthogonalize the machining space of the machine tool to obtain the orthogonalized machining space points;

4) Import the parametric assembly model and the orthogonal processing space point into the Ansys Workbench container;

5) Based on the processing space deformation data obtained in the Ansys Workbench container, use the data learning method to complete the design of the assembly joint surface;

6) Based on the Monte Carlo method and the least square method, combined with the data regression method, the tolerance design of the assembly joint surface profile is completed.

The simplification of the machine tool entity model in the step 1) includes the following steps:

(1) Delete fillets and chamfers with dimensions <20mm;

(2) Delete the features that do not affect the analysis and calculation, including screw holes, bolt holes and oil injection holes;

(3) Simplification of the transmission system, including: the equivalent mass is used inside the reduction box to simplify the structure, and the screw structure is replaced by a simple cylinder;

(4) The structure of the main shaft is simplified. When analyzing the whole machine, the main shaft is considered as the overall structure, and the equivalent solid model and equivalent mass are used instead of the main shaft;

(5) Use mass points to replace auxiliary structures that do not affect the structural rigidity of the whole machine, including drive motors and counterweights;

(6) In order to ensure the quality of the unit, delete or modify the plane features with a size <20mm in the model;

(7) When the number of parts is large, the bolted parts of the same material are combined;

(8) Modify and adjust the size of the gasket model to ensure the accurate assembly of each component.

The parametric assembly in the step 2) refers to parametrizing the assembly position of the large machine tool parts that change position with the position of the machine tool movement axis. For other parts that do not change with the position of the machine tool movement axis, do not Parameterize its assembly position; among them, the large parts of the machine tool that will change position with the position of the machine tool motion axis include slide saddle, headstock, worktable and pallet.

In said step 3), orthogonalizing the machining space of the machine tool refers to selecting the stroke of each axis of the machine tool, the cutting force, and the weight of the clamped workpiece as the orthogonal test factors. For weight, set the upper and lower limits of the above factors in Ansys Workbench, and use the automatic orthogonalization function of Ansys Workbench to complete the design of orthogonal test points and obtain orthogonal processing space points.

The step 4) is specifically: importing the parameterized assembly model and the orthogonal processing space point into the AnsysWorkbench container, and performing contact setting, grid division and material parameter setting, wherein for the machine tool structure, The cell size of mesh division is between 10mm and 100mm, and mesh refinement is carried out at complex structures, complex stress conditions and dangerous points of concern.

The data learning mode in the step 5) includes two, specifically as follows:

Data learning method 1: When each axis of the machine tool is at a fixed position, set the assembly joint surface as S, take a point P ₀ ∈ S at random, and its coordinates are (x ₀ , y ₀ , z ₀ ); take P ₀ as the center of the circle, and r is the radius and draw a circle on the joint surface S, take the weighted arithmetic mean of the deformation of the point P _i included in the circle, and take the weighted arithmetic mean as the deformation of point P ₀ δ ₀ , where the weighted arithmetic mean The weight ρ of is the reciprocal of the distance between P _i and P _0. The deformation of P _i exists in the result file of the Ansys Workbench container. The coordinates of P _i are ( _xi , y _i , z _i ), i=1～t , t is the total number of all points in the circle domain; the calculation formulas of δ ₀ and ρ are as follows:

In the formula, d _i is the distance between point P _i and point P ₀ , and δ _i is the deformation of P _i ;

Data learning method 2: Let the total number of orthogonal processing space points be n, when each axis of the machine tool is located at the position corresponding to the processing space point k, k=1~n, P ₀ is the deformation obtained through data learning method 1 The amount is Take the average of all the deformations obtained by P ₀ when each axis of the machine tool is located at n processing space points, as the deformation of P ₀ in the entire processing space in Calculated by the following formula:

Described step 5) in the design of assembling joint surface specifically is:

a) Define the spatial plane F _n as the nominal joint surface of the assembly, the boundary of F _n is determined by the geometric parameters of the initial actual assembly joint surface, and F _n is determined by the gridded scattered points P _ij (P _ijx ,P _ijy ,P _ijz ) composition;

b) Define the surface F _d as the designed joint surface, and F _d is defined by the point composition, It is obtained by the following formula:

Among them, δ _x , δ _y , and δ _z are the deformations in the three directions of x, y, and z obtained by _Pij through the second data learning method, respectively, and F _d is the result of the design of the assembly joint surface.

In the step 6), the tolerance design of the assembly joint surface profile is specifically:

Let the maximum allowable angular deviation of the joint surface F _d be The joint surface profile tolerance to be designed is T, let m be a random number between [-1,1], use the following formula to generate the joint surface F _d ' that meets the joint surface profile tolerance T requirements, let F _d ' The point on The calculation formula is as follows:

Then use the least squares method to fit the least squares plane of the combined surface F _d ', and set the normal vector of the least squares plane to be Then the angular deviation caused by the combined surface F _d ' is ω(ω _x ,ω _y ,ω _z ), and the calculation formula of ω is as follows:

Repeat the above operations n times to obtain n groups of angle deviations corresponding to n groups of joint surface profile tolerances, and use the data regression method to obtain the mapping relationship between the joint surface profile tolerance T and the angle deviation ω. Through this mapping relationship, in a given The angular deviation is The corresponding joint surface profile tolerance is the designed assembly joint surface profile tolerance.

Compared with the prior art, the beneficial effects of the present invention are:

The method for designing the joint surface of machine tool assembly with the goal of offsetting deformation errors proposed by the present invention reduces the calculation amount of finite element analysis by simplifying the machine tool structure that does not affect the analysis results; at the same time, based on the parametric assembly method, the Ansys Workbench software is combined to construct the Ansys Workbench automatically solves and saves the container, which greatly improves the analysis efficiency; and based on the construction of two data learning methods, it realizes the scientific and quantitative anti-deformation design of the three-dimensional assembly joint surface of the machine tool, avoiding the expression of arbitrary three-dimensional surface functions Dilemma; through the deformation of the entire processing space assembly joint surface contour and its contour tolerance design, the manufacturing difficulty is reduced and the process feasibility is guaranteed. The present invention designs the machine tool assembly joint surface based on the data learning method, gets rid of the difficult problem of expressing the function of the three-dimensional deformation surface, and has good effect at the same time, and the deformation error reduction rate of the main deformation direction reaches 95.3%.

Description of drawings

Figure 1 is a precision horizontal jig boring machine of a certain model, in which the left is the physical map, and the right is a schematic diagram of the machine tool structure;

Figure 2 is a schematic diagram of parametric assembly, where (a), (b), and (c) are schematic diagrams of parameterized assembly distances in the X, Y, and Z directions, respectively;

Figure 3 is a selection diagram of the assembly joint surface;

Figure 4 is a screenshot of Ansys Workbench automatically analyzing and extracting data containers;

Fig. 5 is the design result drawing of the joint surface of bed column of active design;

Fig. 6 is the actual deformation cloud diagram of no anti-deformation design;

Figure 7 is the actual deformation cloud diagram when there is an anti-deformation design;

Among them: 1 is the column, 2 is the saddle, 3 is the spindle box, 4 is the worktable, 5 is the pallet, 6 is the bed, 7 is the X axis, 8 is the X axis distance parameterization, 9 is the Y axis, 10 11 is the Z axis, 12 is the Z axis distance parameterization, 13 is the joint surface of the saddle headstock, 14 is the joint surface of the workpiece table, 15 is the joint surface of the column and the saddle, and 16 is the bed column Joint surface, 17 is bed workbench joint surface.

Detailed ways

Taking the active design of the joint surface of the bed column of a certain type of precision coordinate boring machine as an example, the present invention will be described in detail with reference to the accompanying drawings.

1. Ansys Workbench automatically analyzes and extracts processing space deformation

(1) Machine tool structure simplification

1) Delete the fillet and chamfer of the small size (size <20mm) in the part;

2) Delete small features such as mounting bolt holes, screw holes, and oil injection holes that do not affect the analysis and calculation;

3) Simplification of the transmission system: the transmission mechanisms such as gears and worm gears inside the large frame structure are removed. In the analysis of the whole machine, only the outer casing, the main shaft and the transmission shaft are kept, and the quality characteristics of the original structure are equivalent by changing their material density properties. Delete all motor models and replace them with mass point equivalents in the analysis model;

4) Replace the lead screw with a cylindrical model;

5) Due to the limitation of the minimum mesh size in mesh division, delete or modify the boss or groove features that are too small (<20mm);

The actual model of the machine tool is shown in Figure 1(a), and the machine tool after simplification is shown in Figure 1(b).

(2) Parametric assembly of machine tools

When assembling machine tool parts in CAD software, in order to achieve the above-mentioned automatic extraction of processing space deformation requirements, the distance from the left side of the saddle to the left side of the column (Figure 2(a)), the distance from the lower side of the headstock to the lower side of the saddle (Fig. 2(b)), and the distance from the front side of the pallet to the front side of the column (Fig. 2(c)) were parameterized.

(3) Orthogonal processing space

Select the parameterized assembly position, cutting force, and clamped workpiece weight as the orthogonal design factors, set the upper and lower limits of these factors in ANSYS Workbench according to the actual situation of the machine tool, and use the automatic orthogonalization function of Ansys Workbench to complete the orthogonal design. Hand in the design of test points.

The X, Y, Z axis position of the machine tool, the cutting force, and the weight of the workpiece are taken as the factors of the orthogonal test, and the range of values is shown in Table 1.

Table 1 Orthogonal test factor value range

(4) Ansys Workbench automatically analyzes and extracts deformation data

1) Import the CAD file processed by parametric assembly directly into the Ansys Workbench container through the Geometry module of Ansys Workbench. Combining related parts through Boolean operations can reduce the number of parts in the model, reduce the setting of joint surfaces, and reduce the amount of finite element calculations. In the Ansys Workbench material parameter library, select and establish the material parameters required in the analysis model. In this embodiment, the bonding surface is mainly divided into two types, one is a fixed bonding surface for bolt connection or bonding with a plastic surface, and the other is a sliding bonding surface for rail connection. In the analysis, the fixed connection joint surface of the bolt connection and the plastic surface bonding is set as Bonded (bonded), while the sliding joint surface of the guide rail connection is set as No separation (sliding without separation).

2) Mesh division

Due to the large geometric size of the machine tool model, the large number of parts and the complex structure, the mesh size can be set to 50mm, and the mesh is refined in key areas such as the complex structure and the end of the spindle.

3) Boundary conditions are applied

Set static stiffness analysis boundary conditions in Ansys Workbench software, including: gravity, cutting force, anchor bolt fixing constraints, etc. For static structural analysis, select the assembly joint surfaces shown in Figure 3, and save the analysis results of these joint surfaces.

After the above settings are completed, the container (Ansys Workbench container) for automatically solving and saving the deformation of the joint surface in the entire processing space is built. As shown in Figure 4.

2. Design of assembly joint surface based on data learning

(1) Definition of data learning method

1) Define learning method 1: when each axis of the machine tool is at a fixed position, set the assembly joint surface as S, take a point P ₀ ∈ S at random, and its coordinates are (x ₀ , y ₀ , z ₀ ); take P ₀ as the center of the circle , make a circle on the joint surface S with r as the radius; take the weighted arithmetic mean of the deformation of the point P _i (i=1～t, t is the total number of all points in the circle) contained in the circle; The weighted arithmetic mean is taken as the deformation of point P ₀ , where the weight of the weighted arithmetic mean is the reciprocal of the distance between P _i and P ₀ , the deformation of P _i exists in the result file of the Ansys Workbench container, and the value of P _i The coordinates are (x _i , y _i , z _i ).

In the formula, d _i is the distance between point P _i and point P ₀ . δ ₀ is the deformation of P ₀ , δ _i is the deformation of P _i

2) Define the second learning method: set the total number of orthogonal experimental points generated by the orthogonalization of the processing space as n, and P ₀ learn the deformation data of the orthogonal experimental point k (k=1~n) through the first learning method to obtain the deformation quantity Take the average of the deformation obtained by learning the n orthogonal test points as the deformation of point P ₀ in the entire processing space in is the deformation δ ₀ of P ₀ obtained by learning method 1 when each axis of the machine tool is located at the position corresponding to the orthogonal test point k

3. Assembly interface design

1) Define the spatial plane F _n as the nominal joint surface of the assembly. The boundary of F _n is determined by the geometric parameters of the initial actual assembly joint surface, and is composed of gridded scattered points P _ij (P _ijx ,P _ijy ,P _ijz ).

2) Define the surface F _a as the actual joint surface after deformation. F _a is composed of points P′ _ij (P′ _ijx , P′ _ijy , P′ _ijz ), and P′ _ij is obtained by the following formula:

P′ _ijx =P _ijx +δ _x

P′ _ijy =P _ijy +δ _y

P′ _ijz =P _ijz +δ _z

δ _x , δ _y , and δ _z are the deformation amounts in the x, y, and z directions obtained by P _ij through the second learning method, respectively.

3) Define the surface F _d as the combined surface of the design, and F _d is defined by the point composition, It is obtained by the following formula:

Use F _d as a result of the assembly interface design. The designed F _d can weaken or even offset the deformation error of the machining space, and it is expected to improve the machining accuracy of the machine tool by 12%.

Using the above-mentioned assembly joint surface design method proposed in this paper, considering the deformation of the entire processing space under load, the design is obtained through the acquisition of deformation data and formulas in the second learning method, and the design results in Figure 5 are obtained.

4. Tolerance design of assembly joint surface profile

Design method of contour tolerance: set the maximum allowable angular deviation of the joint surface F _d as The profile tolerance of the joint surface to be designed is T, and m is a random number between [-1,1]. The following formula can be used to generate the joint surface F _d ' that meets the requirements of the tolerance T. Let the points on F _d ' be

The least squares plane of F _d ' is obtained by fitting the least squares method, and the normal vector of the least squares plane is set as Then the angular deviation ω(ω _x ,ω _y ,ω _z ) caused by F _d ' is:

There are n Monte Carlo operations, that is, there are n sets of angular deviations caused by n sets of profile tolerances. The mapping relationship between the contour tolerance T and the angle deviation ω can be obtained by using the data regression method. Deviation at a given angle After that, the design of profile tolerance can be completed.

Specifically, using the design method of joint surface profile tolerance proposed by the present invention, using the linear regression method, using matlab to carry out numerical calculations, the maximum angle deviation is Set as 1×10 ^-7 rad, the joint surface profile tolerance T is 0.010mm.

In order to verify the design effect and the effectiveness of this method, when the parameterized assembly position is 1100, 850, 1130 (x ₀ /y ₀ /z ₀ ), the design results are calculated again with Ansys Workbench, and compared with or without combination In the anti-deformation design of the surface, the angle deviation and flatness caused by the joint surface of the bed column are shown in Table 2.

Table 2 Deformation Angle Deviation of Joint Surface of Bed Column

It can be found from Table 2 that after adopting this method (with anti-deformation design), the angle deviation is greatly reduced to a negligible level. In the Y direction of the main deformation, the angle deviation is reduced by 95.3%, and the flatness of the joint surface is also reduced. become better.

Using matlab to display the deformation result data of the joint surface of the bed column calculated by Ansys Workbench, it can be clearly found that when there is no anti-deformation design, as shown in Figure 6, the order of magnitude of deformation of the joint surface is 10 ^-6 . As shown in Figure 7, the order of magnitude is reduced to 10 ^-7 , and the anti-deformation effect is very significant.

Claims (7)

1. A method for designing the joint surface of machine tool assembly with the aim of counteracting deformation error, is characterized in that, comprises the following steps: 1) Simplify the solid model of the machine tool in the CAD software according to the static analysis requirements and the structural characteristics of the machine tool; 2) Parametrically assemble the solid model of the machine tool in the CAD software to obtain a parametric assembly model; 3) Use Ansys Workbench software to orthogonalize the machining space of the machine tool to obtain the orthogonalized machining space points; 4) Import the parametric assembly model and the orthogonal processing space point into the Ansys Workbench container; 5) Based on the processing space deformation data obtained in the Ansys Workbench container, use the data learning method to complete the design of the assembly joint surface; 6) Based on the Monte Carlo method and the least square method, combined with the data regression method, the tolerance design of the assembly joint surface profile is completed; The data learning mode in the step 5) includes two, specifically as follows: Data learning method 1: When each axis of the machine tool is at a fixed position, set the assembly joint surface as S, take a point P ₀ ∈ S at random, and its coordinates are (x ₀ , y ₀ , z ₀ ); take P ₀ as the center of the circle, and r is the radius and draw a circle on the joint surface S, take the weighted arithmetic mean of the deformation of the point P _i included in the circle, and take the weighted arithmetic mean as the deformation of point P ₀ δ ₀ , where the weighted arithmetic mean The weight ρ of is the reciprocal of the distance between P _i and P _0. The deformation of P _i exists in the result file of the Ansys Workbench container. The coordinates of P _i are ( _xi , y _i , z _i ), i=1～t , t is the total number of all points in the circle domain; the calculation formulas of δ ₀ and ρ are as follows: In the formula, d _i is the distance between point P _i and point P ₀ , δ _i is the deformation amount of P _i , and ρ _i is the weight of the deformation amount δ _i of point P _i ; Data learning method 2: Let the total number of orthogonal processing space points be n, when each axis of the machine tool is located at the position corresponding to the processing space point k, k=1~n, P ₀ is the deformation obtained through data learning method 1 The amount is Take the average of all the deformations obtained by P ₀ when each axis of the machine tool is located at n processing space points, as the deformation of P ₀ in the entire processing space in Calculated by the following formula: 2. the machine tool assembly interface design method aiming at counteracting deformation error according to claim 1, is characterized in that, in described step 1), simplification of machine tool solid model comprises the following steps: (1) Delete fillets and chamfers with dimensions <20mm; (2) Delete the features that do not affect the analysis and calculation, including screw holes, bolt holes and oil injection holes; (3) Simplification of the transmission system, including: the equivalent mass is used inside the reduction box to simplify the structure, and the screw structure is replaced by a simple cylinder; (4) The structure of the main shaft is simplified. When analyzing the whole machine, the main shaft is considered as the overall structure, and the equivalent solid model and equivalent mass are used instead of the main shaft; (5) Use mass points to replace auxiliary structures that do not affect the structural rigidity of the whole machine, including drive motors and counterweights; (6) In order to ensure the quality of the unit, delete or modify the plane features with a size <20mm in the model; (7) When the number of parts is large, the bolted parts of the same material are combined; (8) Modify and adjust the size of the gasket model to ensure the accurate assembly of each component. 3. The machine tool assembly joint surface design method aiming at counteracting deformation errors according to claim 1, characterized in that, the parameterized assembly in said step 2) refers to the For large machine tool parts whose positions change, parameterize their assembly positions, and for other components that do not change with the position of the machine tool movement axis, do not parameterize their assembly positions; among them, for large machine tool parts that change position with the change of machine tool movement axis positions Includes saddle, headstock, table and pallet. 4. The method for designing the joint surface of a machine tool assembly with the aim of offsetting deformation errors according to claim 1, wherein in said step 3), orthogonalizing the machining space of the machine tool refers to selecting the stroke, cutting and cutting of each axis of the machine tool The force and the weight of the clamped workpiece are used as the orthogonal test factors. According to the stroke range, cutting force range, and workpiece weight of the machine tool commonly used, the upper and lower limits of the above factors are set in AnsysWorkbench, and the automatic orthogonalization function of Ansys Workbench is used to complete the test. Orthogonal test points are designed to obtain orthogonal processing space points. 5. The method for designing the joint surface of a machine tool assembly aiming at offsetting deformation errors according to claim 1, wherein said step 4) is specifically: combining the parameterized assembly model and the orthogonalized processing space point Import it into the AnsysWorkbench container, and perform contact setting, mesh division and material parameter setting. For the structure of the machine tool, the unit size of the mesh division is between 10mm and 100mm, and it is used in complex structures, complex stress conditions and Mesh refinement is carried out at the dangerous points of concern. 6. The method for designing the assembly joint surface of a machine tool with the aim of offsetting deformation errors according to claim 1, wherein the design of the assembly joint surface in the step 5) is specifically: a) Define the spatial plane F _n as the nominal joint surface of the assembly, the boundary of F _n is determined by the geometric parameters of the initial actual assembly joint surface, and F _n is determined by the gridded scattered points P _ij (P _ijx ,P _ijy ,P _ijz ) composition; b) Define the surface F _d as the designed joint surface, and F _d is defined by the point composition, It is obtained by the following formula: Among them, δ _x , δ _y , and δ _z are the deformations in the three directions of x, y, and z obtained by _Pij through the second data learning method, respectively, and F _d is the result of the design of the assembly joint surface. 7. The method for designing the assembly joint surface of a machine tool with the aim of offsetting deformation errors according to claim 6, wherein the design of the profile tolerance of the assembly joint surface in the step 6) is specifically: Let the maximum allowable angular deviation of the joint surface F _d be The joint surface profile tolerance to be designed is T, let m be a random number between [-1,1], use the following formula to generate the joint surface F _d ' that meets the joint surface profile tolerance T requirements, let F _d ' The point on The calculation formula is as follows: Then use the least squares method to fit the least squares plane of the joint surface F _d ', assuming the normal vector of the least squares plane is NK, then the angle deviation caused by the joint surface F _d ' is ω(ω _x ,ω _y , ω _z ), the calculation formula of ω is as follows: Repeat the above operations n times to obtain n groups of angle deviations corresponding to n groups of joint surface profile tolerances, and use the data regression method to obtain the mapping relationship between the joint surface profile tolerance T and the angle deviation ω. Through this mapping relationship, in a given The angular deviation is The corresponding joint surface profile tolerance is the designed assembly joint surface profile tolerance.