CN105956303A - Machine tool assembling joint surface design method capable of taking deformation error compensation as target - Google Patents

Abstract

The invention discloses a machine tool assembling joint surface design method capable of taking deformation error compensation as a target. The method comprises the following steps: in CAD (Computer Aided Design) software, according to statics analysis requirements and the structural characteristics of a machine tool, carrying out structure simplification; then, carrying out parameterized assembling on a machine tool entity model; importing the entity model subjected to the simplification and the parameterized assembling into finite element analysis software Ansys Workbench; creating a container capable of carrying out automatic analysis and extraction deformation by utilizing the software, carrying out orthogonalization of a processing spatial position and mesh generation, and meanwhile, finishing statics analysis and result storage; then, through data learning, obtaining the practical deformation amount of a machine tool assembling nominal joint surface under a situation that whole processing space deformation is considered; flipping the obtained practical deformation amount, adding the flipped amount and the nominal joint surface to serve as an actively designed machine tool assembling joint surface; and on the basis of a Monte Carlo method and a least square method, finishing profile tolerance design of the assembling joint surface with the combination of a data regression technology.

Description

A kind of with offset the distortion inaccuracy lathe as target be fitted to each other face method for designing

Technical field

The invention belongs to machine-building and design field, relate to a kind of lathe based on data study and be fitted to each other face method for designing, bag Contain Workbench and automatically analyze and extracted deformation technology, faying face active designs technology and faying face profile tolerance tolerance design skill Art, be specifically related to a kind of with offset the distortion inaccuracy lathe as target be fitted to each other face method for designing.

Background technology

Digit Control Machine Tool can produce obvious additional geometry deformation under the effects such as cutting force, clamping force, gravity and inertia force, thus destroys The each ingredient of lathe original mutual alignment relation, produces additional mismachining tolerance, and this is referred to as deforming (power) error.Research refers to Go out distortion inaccuracy and account for the 12% of all errors of lathe, be the third-largest error source causing lathe space error.The top lathe system in the world The precision machine tool 7CN positioning precision that exceedes making factory three well essence machine production reaches 0.5um, points out to be assembled by scraping on its official website Faying face effectively reduces distortion inaccuracy；Internationally famous machine tool plant DIXI points out that its full accuracy product DIXI270 is by more than 500 Hour scraping also effectively reduce deformation by mistake, it is ensured that the error between single shaft precision and axle and axle；China's famous lathe main frame Factory, Kunming lathe, Qin Chuan lathe etc., machine part drawing all can mark (-) or (+) as linearity, flatness Additional requirement reduces the distortion inaccuracy impact on mismachining tolerance.

It is fitted to each other face offsets the important means that geometry deformation is Machine Manufacture enterprise to the destruction of machining accuracy, so by scraping grinding And the face that is fitted to each other by lathe is scraped grinding and is planted shape why, scraping the most wherein, in the problems such as scraping is how many, heavy dependence experience, Lack the technical scheme of scientific and reasonable quantification.Essentially it is a lack of the parts profile shape design in the case of considering part deformation Method.

In the research few in number to Machine Tool Joint Surface active designs, the Sun of Dalian University of Science & Engineering etc. utilizes multi-form approximating method, right Rolling guide joint has carried out " predeformation " design and has offset the error that feed system causes because of plastic deformation, but not from complete machine tool Angle carries out being fitted to each other face active designs；The method is for machine tool slide faying face simultaneously, inapplicable to Fixed Joints in Machine Tools； And being only applicable to build two-dimensional silhouette with polynomial fitting method, being fitted to each other that facial contour do not has for common lathe three-dimensional can Row.

Summary of the invention

Present invention aims to current Machine Manufacture main engine plants shortage lathe scientific and reasonable, quantification and be fitted to each other facial contour Method for designing, heavy dependence pincers worker experience, and the present situation of the defect for two-dimensional shapes it is only capable of for the design of current faying face, carry Go out a kind of with offset the distortion inaccuracy lathe as target be fitted to each other face method for designing, the method is based on data learning style, it is considered to whole The faying face profile of individual processing space deformation and profile tolerance tolerance thereof, it is possible to be substantially reduced distortion inaccuracy.

For reaching above-mentioned purpose, the technical solution used in the present invention is:

A kind of with offset the distortion inaccuracy lathe as target be fitted to each other face method for designing, comprise the following steps:

1) require according to statics Analysis in CAD software and machine tool structure feature carries out lathe Solid Model Simplification；

2) parametric assembly lathe physical model in CAD software, obtains parameterized entire assembly model；

3) utilize Ansys Workbench software by machine tooling orthogonal space, obtain orthogonalized processing space point；

4) parameterized entire assembly model and orthogonalized processing space point are imported in Ansys Workbench container；

5) based on the processing space deformation data obtained in Ansys Workbench container, utilize data learning style, complete dress Join the design of faying face；

6) based on monte carlo method and method of least square, combine data regression method, complete to be fitted to each other surface profile tolerance Design.

Described step 1) in lathe Solid Model Simplification comprise the following steps:

(1) size < corners and fillets of 20mm is deleted；

(2) feature that not impact analysis calculates is deleted, including screw hole, bolt hole and oil filler point；

(3) drive system simplifies, including: reduction box is internal uses equivalent mass to replace carrying out structure simplification, and screw structure uses letter Single cylindrical replaces；

(4) main axle structure simplifies, and when carrying out complete machine analysis, main shaft considers as overall structure, uses equivalent entity model and equivalence Quality replaces main shaft；

(5) quality point is used to replace complete machine structure rigidity is not produced the supplementary structure of impact, including driving motor and counterweight；

(6) for ensureing element quality, in the model size < plane characteristic of 20mm is deleted or modified；

(7) when number of parts is more, the screw coupling member that material is identical is merged；

(8) amendment adjusts pad moulded dimension, it is ensured that the accurate assembling of each parts.

Described step 2) in parametric assembly refer to for can along with machine tool motion shaft position change and the lathe of position change occur Big part, its rigging position of parametrization, for other parts not changed along with machine tool motion shaft position, not its assembling position of parametrization Put；Wherein can along with machine tool motion shaft position change and occur the Big Part in Machine Tool of position change include saddle, main spindle box, workbench and Supporting plate.

Described step 3) in refer to machine tooling orthogonal space to select the stroke of each axle of lathe, cutting force, clamping workpiece weight Measure as orthogonal test factor, commonly use the stroke range of processing, cutting force scope, often processing workpiece weight according to lathe, at Ansys Workbench arranges the bound of above-mentioned factor, utilizes the automatic orthogonalization function of Ansys Workbench, complete orthogonal examination Test design a little, obtain orthogonalized processing space point.

Described step 4) particularly as follows: parameterized entire assembly model and orthogonalized processing space point are imported Ansys Workbench In container, and carry out contact setting, stress and strain model and material parameter setting, wherein for complete machine tool structure, the list of stress and strain model Elemental size is between 10mm to 100mm, and carries out grid at the dangerous spot of, force-bearing situation complexity complicated in structure and care Refinement.

Described step 5) in data learning style include two, specific as follows:

Data learning style one: when each axle of lathe is positioned at fixed position, if the face that is fitted to each other is S, appoints and takes 1 P₀∈ S, Its coordinate is (x₀,y₀,z₀)；With P₀For the center of circle, justify on faying face S for radius with r, will be contained in the some P in this circle territory_i Deformation measure weighted arithmetic mean, using this weighted arithmetic mean value as P₀The deflection δ of point₀, wherein weighted arithmetic mean value Weights ρ be P_iWith P₀The inverse of distance, P_iDeflection be present in the destination file of Ansys Workbench container, P_i Coordinate be (x_i,y_i,z_i), i=1～t, t be in this circle territory total number a little；Wherein δ₀As follows with the computing formula of ρ:

${\mathrm{\&delta;}}_0=\left\{\begin{array}{cc}\frac{\underset{i=1}{\overset{t}{\&Sigma;}}{\mathrm{\&rho;}}_i{\mathrm{\&delta;}}_i}{\underset{i=1}{\overset{t}{\&Sigma;}}{\mathrm{\&rho;}}_i}& d_i\&NotEqual;0\\ {\mathrm{\&delta;}}_i& d_i=0\end{array}\right.$

${\mathrm{\&rho;}}_i=\left\{\begin{array}{cc}\frac{1}{d_i}& 0<d_i\&le;r\\ 0& d_i>r\end{array}\right.$

$d_i=\sqrt{{(x_i-x_0)}^2+{(y_i-y_0)}^2+{(z_i-z_0)}^2}$

In formula, d_iIt is a P_iWith a P₀Distance, δ_iFor P_iDeflection；

Data learning style two: set total number of orthogonalized processing space point as n, when each axle of lathe is positioned at processing space point k During corresponding position, k=1～n, P₀The deflection obtained by data learning style one isEach axle of lathe is positioned at n add P during work spatial point₀The all deflections obtained are averaged, as P₀Deflection in whole processing spaceWhereinPass through Following formula calculates:

${\mathrm{\&delta;}}_0^{*}=\frac{\underset{k=1}{\overset{n}{\&Sigma;}}{\mathrm{\&delta;}}_0^k}{n}.$

Described step 5) in be fitted to each other face design particularly as follows:

A) definition space plane F_nFor assembling name faying face, F_nBorder true by the geometric parameter of initial practical set faying face Fixed, F_nBy scatterplot P of gridding_ij(P_ijx,P_ijy,P_ijz) composition；

B) definition face F_dFor the faying face of design, F_dBy pointComposition,It is obtained by the following formula:

$P_{ijx}^{*}=P_{ijx}-{\mathrm{\&delta;}}_x$

$P_{ijy}^{*}=P_{ijy}-{\mathrm{\&delta;}}_y$

$P_{ijz}^{*}=P_{ijz}-{\mathrm{\&delta;}}_z$

Wherein δ_x、δ_y、δ_zIt is respectively P_ijIts deflection in three directions of x, y, z obtained by data learning style two, F_d It is the result being fitted to each other face design.

Described step 6) in be fitted to each other surface profile tolerance design particularly as follows:

If faying face F_dThe maximum angle deviation allowed isThe faying face profile tolerance tolerance needing design is T, if m is [-1,1] Between random number, utilize below equation can generate meet faying face profile tolerance tolerance T require faying face F_d', if F_dOn ' Point is Computing formula as follows:

$P_{ijx}=P_{ijx}^{*}+\frac{T}{2}m$

$P_{ijy}=P_{ijy}^{*}+\frac{T}{2}m$

$P_{ijz}=P_{ijz}^{*}+\frac{T}{2}m$

Then least square fitting is utilized to obtain faying face F_d' least square plane, if the normal vector of this least square plane isThen faying face F_d' angular deviation that causes is ω (ω_x,ω_y,ω_z), the computing formula of ω is as follows:

Repeat n above-mentioned computing, obtain the n group angular deviation that n group faying face profile tolerance tolerance is corresponding, utilize data regression method Obtain the mapping relations between faying face profile tolerance tolerance T and angular deviation ω, by these mapping relations, in given angle deviation beTime corresponding faying face profile tolerance tolerance be design be fitted to each other surface profile tolerance.

Relative to prior art, the invention have the benefit that

What the present invention proposed is fitted to each other face method for designing with the counteracting distortion inaccuracy lathe as target, ties by simplifying not impact analysis The machine tool structure of fruit, decreases finite element analysis computation amount；Meanwhile, based on parametric assembly method, combine Ansys Workbench Software building Ansys Workbench automatization solves and preserves container, greatly improves analysis efficiency；And building two On the basis of kind of data learning style, it is achieved that lathe three-dimensional is fitted to each other the reversible deformation design of the scientific quantitative analysis in face, it is to avoid appoint The predicament of meaning three-dimension curved surface function representation；By whole processing space is deformed be fitted to each other facial contour and profile tolerance tolerance sets Meter, reduces manufacture difficulty, it is ensured that technological feasibility.The present invention is fitted to each other by designing lathe based on data learning style Face, has broken away from the difficult problem that 3 D deformation toroidal function is expressed, and the most respond well, main transformer shape Direction distortion error reduction rate reaches 95.3%.

Accompanying drawing explanation

Fig. 1 is certain model precision horizontal jig boring machine, and wherein the left side is pictorial diagram, and the right is machine tool structure sketch；

Fig. 2 is parametric assembly schematic diagram, and wherein (a), (b), (c) are respectively X, Y, Z-direction mounting distance parametrization schematic diagram；

Fig. 3 is for being fitted to each other face selection figure；

Fig. 4 is the sectional drawing that Ansys Workbench automatically analyzed and extracted data capsule；

Fig. 5 is the design result figure of the lathe bed column faying face of active designs；

Fig. 6 is without anti-deformation d esign practical distortion cloud atlas；

Practical distortion cloud atlas when Fig. 7 is for there being anti-deformation d esign；

Wherein: 1 is column, 2 is saddle, and 3 is main spindle box, and 4 is workbench, and 5 is supporting plate, and 6 is lathe bed, and 7 is X-axis, 8 For X-axis distance parameter, 9 is Y-axis, and 10 is Y-axis distance parameter, and 11 is Z axis, and 12 is Z axis distance parameter, 13 is saddle main spindle box faying face, and 14 is workpiece workbench faying face, and 15 is column saddle faying face, and 16 combine for lathe bed column Face, 17 is lathe bed workbench faying face.

Detailed description of the invention

As a example by the lathe bed column faying face active designs of certain type precision jig boring machine, in conjunction with accompanying drawing, the present invention is described in detail.

One, Ansys Workbench automatically analyzes and extracts processing space deformation

(1) machine tool structure simplifies

1) corners and fillets of part small-medium size (size < 20mm) is deleted；

2) the little feature that not impact analysis such as installing bolt hole, screw hole, oil filler point calculates is deleted；

3) drive mechanisms such as drive system simplifies: remove the gear of big frame inside configuration, turbine and worm.In complete machine is analyzed only Retain shell body, main shaft and three parts of power transmission shaft, carry out the qualitative character of equivalent prototype structure by changing its density of material attribute. Delete all motor models, replace with quality point equivalence in analyzing model；

4) leading screw cylinder model is replaced；

5) due to the restriction of minimum grid size in stress and strain model, boss or the groove of undersized (< 20mm) is deleted or modified Feature；

Shown in lathe realistic model such as Fig. 1 (a), shown in this lathe such as Fig. 1 (b) after having simplified.

(2) lathe parameter makeup is joined

When assembling machine part in CAD software, above-mentioned automatically extract processing space deformation requirements to be able to realize, left to saddle Side on the downside of the distance (Fig. 2 (a)) on the left of column, main spindle box on front side of the distance (Fig. 2 (b)) on the downside of saddle, supporting plate to vertical Distance (Fig. 2 (c)) on front side of post has carried out parameterized treatment.

(3) orthogonalization processing space

The rigging position of Selection parameter, cutting force, clamping workpiece weight are as orthogonal design factor, according to lathe actual capabilities feelings Condition, sets the bound of these factors in ANSYS Workbench, utilizes the automatic orthogonalization of Ansys Workbench Function, completes the design of orthogonal test point.

By the X, Y, Z axis position of lathe, cutting force, workpiece weight is as the factor of orthogonal test, the scope of its value such as table Shown in 1.

Table 1 orthogonal test factor span

(4) Ansys Workbench automatically analyzes and extracts deformation data

1) cad file that parametric assembly processes directly is imported to by the Geometry module of Ansys Workbench In Ansys Workbench container.By Boolean calculation, correlated parts is merged, it is possible to reduce number of parts in model, reduce knot The setting in conjunction face, reduces FEM calculation amount.In Ansys Workbench material parameter storehouse, select and set up in analysis model The material parameter needed.In the present embodiment, faying face is broadly divided into two classes, and a class is bolt connection or laminating bonding fixing Faying face, a class is the face that is slidably connected that guide rail connects.Bolt is connected by analysis and laminating bonding fixing connection faying face sets It is set to Bonded (bonding), and the face of being slidably connected that guide rail connects is set to No separation (slide and do not separate).

2) stress and strain model

Owing to complete machine tool model geometric size is relatively big, amount of parts is more and structure is complicated, can arrange size of mesh opening and be 50mm, carries out mesh refinement in critical concern regions such as structure complexity and spindle noses.

3) boundary condition applies

In Ansys Workbench software, Analysis on Static Stiffness boundary condition is set, including: gravity, cutting force, foundation bolt are solid Conclude a contract or treaty bundle etc..Carry out static construction analysis, select as shown in Figure 3 be fitted to each other face, the analysis result of these faying faces is carried out Preserve.

Above-mentioned be provided with after, just built faying face whole processing space deflection automatic calculation with preserve container (Ansys Workbench container).As shown in Figure 4.

Two, the face that is fitted to each other based on data study design

(1) data learning style definition

1) definition learning style one: when each axle of lathe is positioned at fixed position, if the face that is fitted to each other is S, appoints and takes 1 P₀∈ S, Its coordinate is (x₀,y₀,z₀)；With P₀For the center of circle, justify on faying face S with r for radius；Will be contained in the point in this circle territory P_iThe deformation of (i=1～t, t are in this circle territory total number a little) measures weighted arithmetic mean；Using this weighted arithmetic mean value as P₀The deflection of point, wherein the weights of weighted arithmetic mean value are P_iWith P₀The inverse of distance, P_iDeflection be present in Ansys In the destination file of Workbench container, P_iCoordinate be (x_i,y_i,z_i)。

${\mathrm{\&delta;}}_0=\left\{\begin{array}{cc}\frac{\underset{i=1}{\overset{t}{\&Sigma;}}{\mathrm{\&rho;}}_i{\mathrm{\&delta;}}_i}{\underset{i=1}{\overset{t}{\&Sigma;}}{\mathrm{\&rho;}}_i}& d_i\&NotEqual;0\\ {\mathrm{\&delta;}}_i& d_i=0\end{array}\right.$

${\mathrm{\&rho;}}_i=\left\{\begin{array}{cc}\frac{1}{d_i}& 0<d_i\&le;r\\ 0& d_i>r\end{array}\right.$

$d_i=\sqrt{{(x_i-x_0)}^2+{(y_i-y_0)}^2+{(z_i-z_0)}^2}$

In formula, d_iIt is a P_iWith a P₀Distance.δ₀For P₀Deflection, δ_iFor P_iDeflection

2) definition learning style two: set the orthogonal experiment total number of point that processing space orthogonalization produces as n, P₀Pass through learning style One goes the deformation data learning orthogonal test point k (k=1～n) to obtain deflectionBy this n orthogonal test point by learning To deflection be averaged as a P₀Deflection in whole processing spaceWhereinFor being positioned at orthogonal when each axle of lathe The P obtained by learning style one during corresponding for testing site k position₀Deflection δ₀

${\mathrm{\&delta;}}_0^{*}=\frac{\underset{k=1}{\overset{n}{\&Sigma;}}{\mathrm{\&delta;}}_0^k}{n}$

Three, face design it is fitted to each other

1) definition space plane F_nFor assembling name faying face, F_nBorder true by the geometric parameter of initial practical set faying face Fixed, by scatterplot P of gridding_ij(P_ijx,P_ijy,P_ijz) composition.

2) definition face F_aFor the actual faying face after deformation, F_aBy some P '_ij(P′_ijx,P′_ijy,P′_ijz) composition, P '_ijIt 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、δ_zIt is respectively P_ijBy the deflection in three directions of x, y, z that learning style two obtains.

3) definition face F_dFor the faying face of design, F_dBy pointComposition,It is obtained by the following formula:

$P_{ijx}^{*}=P_{ijx}-{\mathrm{\&delta;}}_x$

$P_{ijy}^{*}=P_{ijy}-{\mathrm{\&delta;}}_y$

$P_{ijz}^{*}=P_{ijz}-{\mathrm{\&delta;}}_z$

By F_dAs the result being fitted to each other face design.Designed F_dThe distortion inaccuracy even offsetting dress processing space can be weakened, Expection can improve the machine finish of lathe 12%.

Use and in this paper above-mentioned be fitted to each other face method for designing, it is considered to whole processing space stand under load deforms, by learning style two Deformation data obtain and formula be designed, obtain the design result of Fig. 5.

Four, surface profile tolerance design it is fitted to each other

Profile tolerance tolerance design method: set faying face F_dThe maximum angle deviation allowed isNeed the faying face profile tolerance of design Tolerance is T, if m is the random number between [-1,1].Utilize below equation to generate and meet the faying face F that tolerance T requires_d', If F_dPoint on ' is

$P_{ijx}=P_{ijx}^{*}+\frac{T}{2}m$

$P_{ijy}=P_{ijy}^{*}+\frac{T}{2}m$

$P_{ijz}=P_{ijz}^{*}+\frac{T}{2}m$

Least square fitting is utilized to obtain F_d' least square plane, if the normal vector of this least square plane isSo F_d’ Angular deviation ω (the ω caused_x,ω_y,ω_z) it is:

It is provided with n Monte Carlo computing, i.e. has the n group angular deviation that n group profile tolerance tolerance causes.Utilize the data regression method can To obtain the mapping relations between profile tolerance tolerance T and angular deviation ω.In given angle deviationAfterwards, profile tolerance can be completed The design of tolerance.

Concrete, utilize the faying face profile tolerance tolerance design method that the present invention proposes, utilize linear regression method, use matlab Carry out numerical computations, by maximum angle deviation beIt is set as 1 × 10^-7Rad, obtaining faying face profile tolerance tolerance T is 0.010mm.

In order to verify the effectiveness of design effect and this method, take 1100 in parametric assembly position, 850,1130 (x₀/y₀/z₀) time, The design result is calculated with Ansys Workbench again, and when contrasting with or without faying face anti-deformation d esign, lathe bed column Angular deviation that faying face causes and flatness, as shown in table 2.

Table 2 lathe bed column deformation of joint surface angular deviation

It is found that use this method (having anti-deformation d esign) angular deviation to be afterwards greatly reduced to negligible from table 2 Degree, in main transformer shape Y-direction, angular deviation reduces 95.3%, and the flatness in combination with face also becomes more preferable.

The lathe bed column deformation of joint surface result data matlab that Ansys Workbench calculates is shown, can be obvious When finding without anti-deformation d esign, as shown in Figure 6, the deformation of joint surface amount order of magnitude is 10^-6, after having anti-deformation d esign, such as Fig. 7 Shown in, the order of magnitude is reduced to 10^-7, resistance to deformation effect is the most notable.

Claims (8)

1. one kind is fitted to each other face method for designing to offset the distortion inaccuracy lathe as target, it is characterised in that comprise the following steps:

1) require according to statics Analysis in CAD software and machine tool structure feature carries out lathe Solid Model Simplification；

2) parametric assembly lathe physical model in CAD software, obtains parameterized entire assembly model；

3) utilize Ansys Workbench software by machine tooling orthogonal space, obtain orthogonalized processing space point；

4) parameterized entire assembly model and orthogonalized processing space point are imported in Ansys Workbench container；

5) based on the processing space deformation data obtained in Ansys Workbench container, utilize data learning style, complete dress Join the design of faying face；

6) based on monte carlo method and method of least square, combine data regression method, complete to be fitted to each other surface profile tolerance Design.

The most according to claim 1 with offset the distortion inaccuracy lathe as target be fitted to each other face method for designing, its feature exists In, described step 1) in lathe Solid Model Simplification comprise the following steps:

(1) size < corners and fillets of 20mm is deleted；

(2) feature that not impact analysis calculates is deleted, including screw hole, bolt hole and oil filler point；

(3) drive system simplifies, including: reduction box is internal uses equivalent mass to replace carrying out structure simplification, and screw structure uses letter Single cylindrical replaces；

(4) main axle structure simplifies, and when carrying out complete machine analysis, main shaft considers as overall structure, uses equivalent entity model and equivalence Quality replaces main shaft；

(5) quality point is used to replace complete machine structure rigidity is not produced the supplementary structure of impact, including driving motor and counterweight；

(6) for ensureing element quality, in the model size < plane characteristic of 20mm is deleted or modified；

(7) when number of parts is more, the screw coupling member that material is identical is merged；

(8) amendment adjusts pad moulded dimension, it is ensured that the accurate assembling of each parts.

The most according to claim 1 with offset the distortion inaccuracy lathe as target be fitted to each other face method for designing, its feature exists In, described step 2) in parametric assembly refer to, for changing along with machine tool motion shaft position, the lathe of position change occur Big part, its rigging position of parametrization, for other parts not changed along with machine tool motion shaft position, not its assembling position of parametrization Put；Wherein can along with machine tool motion shaft position change and occur the Big Part in Machine Tool of position change include saddle, main spindle box, workbench and Supporting plate.

The most according to claim 1 with offset the distortion inaccuracy lathe as target be fitted to each other face method for designing, its feature exists In, described step 3) in refer to select the stroke of each axle of lathe, cutting force, clamping workpiece weight by machine tooling orthogonal space Measure as orthogonal test factor, commonly use the stroke range of processing, cutting force scope, often processing workpiece weight according to lathe, at Ansys Workbench arranges the bound of above-mentioned factor, utilizes the automatic orthogonalization function of Ansys Workbench, complete orthogonal examination Test design a little, obtain orthogonalized processing space point.

The most according to claim 1 with offset the distortion inaccuracy lathe as target be fitted to each other face method for designing, its feature exists In, described step 4) particularly as follows: parameterized entire assembly model and orthogonalized processing space point are imported Ansys Workbench In container, and carry out contact setting, stress and strain model and material parameter setting, wherein for complete machine tool structure, the list of stress and strain model Elemental size is between 10mm to 100mm, and carries out grid at the dangerous spot of, force-bearing situation complexity complicated in structure and care Refinement.

The most according to claim 1 with offset the distortion inaccuracy lathe as target be fitted to each other face method for designing, its feature exists In, described step 5) in data learning style include two, specific as follows:

Data learning style one: when each axle of lathe is positioned at fixed position, if the face that is fitted to each other is S, appoints and takes 1 P₀∈ S, Its coordinate is (x₀,y₀,z₀)；With P₀For the center of circle, justify on faying face S for radius with r, will be contained in the some P in this circle territory_i Deformation measure weighted arithmetic mean, using this weighted arithmetic mean value as P₀The deflection δ of point₀, wherein weighted arithmetic mean value Weights ρ be P_iWith P₀The inverse of distance, P_iDeflection be present in the destination file of Ansys Workbench container, P_i Coordinate be (x_i,y_i,z_i), i=1～t, t be in this circle territory total number a little；Wherein δ₀As follows with the computing formula of ρ:

${\mathrm{\&delta;}}_0=\left\{\begin{array}{cc}\frac{\underset{i=1}{\overset{t}{\&Sigma;}}{\mathrm{\&rho;}}_i{\mathrm{\&delta;}}_i}{\underset{i=1}{\overset{t}{\&Sigma;}}{\mathrm{\&rho;}}_i}& d_i\&NotEqual;0\\ {\mathrm{\&delta;}}_i& d_i=0\end{array}\right.$

${\mathrm{\&rho;}}_i=\left\{\begin{array}{cc}\frac{1}{d_i}& 0<d_i\&le;r\\ 0& d_i>r\end{array}\right.$

$d_i=\sqrt{{(x_i-x_0)}^2+{(y_i-y_0)}^2+{(z_i-z_0)}^2}$

In formula, d_iIt is a P_iWith a P₀Distance, δ_iFor P_iDeflection；

Data learning style two: set total number of orthogonalized processing space point as n, when each axle of lathe is positioned at processing space point k During corresponding position, k=1～n, P₀The deflection obtained by data learning style one isEach axle of lathe is positioned at n add P during work spatial point₀The all deflections obtained are averaged, as P₀Deflection in whole processing spaceWhereinPass through Following formula calculates:

${\mathrm{\&delta;}}_0^{*}=\frac{\underset{k=1}{\overset{n}{\&Sigma;}}{\mathrm{\&delta;}}_0^k}{n}.$

The most according to claim 6 with offset the distortion inaccuracy lathe as target be fitted to each other face method for designing, its feature exists In, described step 5) in be fitted to each other face design particularly as follows:

A) definition space plane F_nFor assembling name faying face, F_nBorder true by the geometric parameter of initial practical set faying face Fixed, F_nBy scatterplot P of gridding_ij(P_ijx,P_ijy,P_ijz) composition；

B) definition face F_dFor the faying face of design, F_dBy pointComposition,It is obtained by the following formula:

$P_{ijx}^{*}=P_{ijx}-{\mathrm{\&delta;}}_x$

$P_{ijy}^{*}=P_{ijy}-{\mathrm{\&delta;}}_y$

$P_{ijz}^{*}=P_{ijz}-{\mathrm{\&delta;}}_z$

Wherein δ_x、δ_y、δ_zIt is respectively P_ijIts deflection in three directions of x, y, z obtained by data learning style two, F_d It is the result being fitted to each other face design.

The most according to claim 7 with offset the distortion inaccuracy lathe as target be fitted to each other face method for designing, its feature exists In, described step 6) in be fitted to each other surface profile tolerance design particularly as follows:

If faying face F_dThe maximum angle deviation allowed isThe faying face profile tolerance tolerance needing design is T, if m is [-1,1] Between random number, utilize below equation can generate meet faying face profile tolerance tolerance T require faying face F_d', if F_dOn ' Point is Computing formula as follows:

$P_{ijx}=P_{ijx}^{*}+\frac{T}{2}m$ $P_{ijy}=P_{ijy}^{*}+\frac{T}{2}m$

$P_{ijz}=P_{ijz}^{*}+\frac{T}{2}m$

Then least square fitting is utilized to obtain faying face F_d' least square plane, if the normal vector of this least square plane isThen faying face F_d' angular deviation that causes is ω (ω_x,ω_y,ω_z), the computing formula of ω is as follows:

Repeat n above-mentioned computing, obtain the n group angular deviation that n group faying face profile tolerance tolerance is corresponding, utilize data regression method Obtain the mapping relations between faying face profile tolerance tolerance T and angular deviation ω, by these mapping relations, in given angle deviation beTime corresponding faying face profile tolerance tolerance be design be fitted to each other surface profile tolerance.