CN107194606A - A kind of Digit Control Machine Tool part assembles mass analysis method - Google Patents
A kind of Digit Control Machine Tool part assembles mass analysis method Download PDFInfo
- Publication number
- CN107194606A CN107194606A CN201710434211.XA CN201710434211A CN107194606A CN 107194606 A CN107194606 A CN 107194606A CN 201710434211 A CN201710434211 A CN 201710434211A CN 107194606 A CN107194606 A CN 107194606A
- Authority
- CN
- China
- Prior art keywords
- value
- influence factor
- factor
- quality
- machine tool
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
- G06Q10/0639—Performance analysis of employees; Performance analysis of enterprise or organisation operations
- G06Q10/06395—Quality analysis or management
Landscapes
- Business, Economics & Management (AREA)
- Human Resources & Organizations (AREA)
- Engineering & Computer Science (AREA)
- Strategic Management (AREA)
- Development Economics (AREA)
- Economics (AREA)
- Entrepreneurship & Innovation (AREA)
- Educational Administration (AREA)
- Operations Research (AREA)
- Marketing (AREA)
- Game Theory and Decision Science (AREA)
- Quality & Reliability (AREA)
- Tourism & Hospitality (AREA)
- Physics & Mathematics (AREA)
- General Business, Economics & Management (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Factory Administration (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
Mass analysis method is assembled the invention discloses a kind of Digit Control Machine Tool part, the dimensionality reduction of numerous quality inspection indexs is realized using PCA first, form the principal component being mutually independent to a certain extent, then implementation procedure Digit Control Machine Tool part assembled is regarded as a Uncertain information system, introduce evidence theory to analyze come " uncertainty " to the information system, so as to realize the identification of influencing factors of quality in the assembling of Digit Control Machine Tool part.The present invention takes into full account the redundancy and the feature of Uncertain information system of extensive quality inspection achievement data in Digit Control Machine Tool part assembling process, the identification of Digit Control Machine Tool part assembling process influencing factors of quality is realized, so as to effectively overcome the shortcomings of manufacture or the assembling process mass analysis method of main flow.
Description
Technical field
The present invention relates to product assembly quality analysis field, specially a kind of Digit Control Machine Tool part assembling quality analysis side
Method.
Background technology
Quality analysis, control and improvement in Digit Control Machine Tool assembling process are to ensure and persistently lifted NC Machine Quality
Important step.A series of nodes such as part assembling, general assembly constitute the assembling process of Digit Control Machine Tool, and wherein part assembling is numerical control
The key link that machine mass ensures.Part assembling is made up of a series of process of man-machine participations, there is substantial amounts of quality inspection
Index is tested, final assembling quality is come out by a series of quality inspection index characterizations.Because structure of numerically controlled machine-tool is huge, cause
Assembly process is various, and the forward-backward correlation of process causes some test ratings to there is interactional relation, is reflecting total constitution
Existence information redundancy when amount is evaluated.Further, since combined influence of the assembling quality by many factors in implementation procedure, to dress
When being analyzed with quality, which factor have influence on assembling quality, influence degree how the problems such as be difficult solve.Based on numerical control machine
Bed part assembling process is by the feature of many factors combined influence, and can be considered as one has obvious probabilistic information system
System, and the existing method based on statistical Process Control and process capability analysis does not account for information when carrying out quality analysis
The uncertain feature of system.
The content of the invention
The quality analysis problem assembled for Digit Control Machine Tool part, the present invention realizes crowd using PCA first
The dimensionality reduction of multimass test rating, forms the principal component being mutually independent to a certain extent, then fills Digit Control Machine Tool part
The implementation procedure matched somebody with somebody is regarded as a Uncertain information system, introduces evidence theory and enters come " uncertainty " to the information system
Row analysis, so as to realize the identification of influencing factors of quality in the assembling of Digit Control Machine Tool part.
The technical scheme is that：
A kind of Digit Control Machine Tool part assembling mass analysis method, it is characterised in that：Comprise the following steps：
Step 1：Quality inspection index to numerical control machine tool component assembling process carries out PCA dimension-reduction treatment.
The PCA dimension-reduction treatment is comprised the following steps that：
Step 1.1 for certain type Digit Control Machine Tool assembling process, if the quality inspection index of its certain part assembling is with master
P are had before componential analysis dimension-reduction treatment.According to the quality record of the part assembling process in n production batch, statistics p
The qualification rate or average data of quality inspection index.Using each production batch as a sample, then sample data square is obtained
Battle array X=(xi,t)n×pForWherein xi,tRepresent t (t=1,2 ..., p) individual quality inspection index
I-th, (i=1,2 ..., the n) data in batch, are represented with qualification rate or average value.To X=(xi,t)n×pCarry out at standardization
Reason, i.e.,
Sample data matrix X'=(x' after step 1.2 pair standardizationi,t)n×p, calculate its correlation matrix COR=
(corr,j)p×p, wherein
Step 1.3 obtains correlation matrix COR=(corr,j)p×pCharacteristic value and by size order arrange, i.e. ev1、
ev2、…、evp, corresponding characteristic vector is followed successively by Ev1、Ev2、…、Evp, wherein Evt=(Evt1,Evt2,...,Evtp), t=
1,2,...,p。
Step 1.4 calculate successively t (t=1,2 ..., p) the contribution rate con of individual quality inspection indext, have
Step 1.5 chooses the individual composition conducts of preceding k (k≤p) that contribution rate of accumulative total reaches more than 85% according to contribution rate size
Principal component, with ev1、ev2、…、evkCorresponding characteristic vector Ev1、Ev2、…、EvkConstruct the matrix EV of p × k dimensions；Make X "=
X'EV, so that n × p standardization sample data matrix X' tieed up to be converted into the new matrix X " of n × k dimensions.K is only remained in X "
Individual principal component, it is achieved thereby that the dimension-reduction treatment of quality inspection index.
Step 2：Set up the possibility set of factors Y={ y of influence assembling quality in Digit Control Machine Tool part assembling process1,
y2,…,yN, there is N=11：y1The physiological characteristic situation of=personnel；y2=psychology of seeking quickness；y3=auxiliary equipment fault level；y4=
Temperature humidity factor；y5=measurer trueness error；y6=protocol adherence situation；y7=degree of fatigue；y8=post is skilled in technique journey
Degree；y9The old degree of wear of=measurer；y10=idea of leaving things to chance；y11=it is mixed into impurity situation.Category is divided into four classes：Human factor
={ y1,y2,y7,y8,y10}；Apparatus factor={ y3,y5,y9}；Management factors={ y6}；Environmental factor={ y4,y11}.By property
It is divided into two classes：Quantitative class factor={ y4,y5,y9}；Qualitative class factor={ y1,y2,y3,y6,y7,y8,y10,y11}。
Step 3：The identification of Digit Control Machine Tool part assembling process influencing factors of quality is carried out based on evidence theory.With Digit Control Machine Tool
Part assembling process is object, and the set of all possible sexual factors of quality will be influenceed to be set to framework of identification, i.e. Θ={ y1,
y2,...,yN, wherein yi(i=1,2 ..., N) it is i-th of influence factor.Set power set 2 is possible on ΘΘCarry out table
Show, when the member in Θ have N number of and each element objectionable intermingling, Θ power set 2ΘElement number be 2N.If A is influence
Single factor test or the factor combination of quality are assembled, m (A) is the basic probability assignment function of A on identification framework Θ, represents the letter to A
Ren Du, meetsAndM (A) is represented by m:2Θ→ [0,1], meets m (A)>0 A is referred to as burnt member.
Evidence in the identification of Digit Control Machine Tool part assembling process influencing factors of quality is exactly the quality shape occurred in known assembling process
Quality inspection achievement data after condition, the dimensionality reduction obtained by step 1 is constituted.
What the Digit Control Machine Tool part assembling process influencing factors of quality was recognized comprises the following steps that：
Step 3.1 carries out standardization processing to the corresponding influence factor value of n production batch, obtains matrix Env=
(envi,t)n×N, envi,tExpression production batch i (i=1,2 ..., influence factor y n)t(t=1,2 ..., N) value.
The possibility sexual factor of influence assembling quality is divided into quantitative class influence factor and fixed in Digit Control Machine Tool part assembling process
Property class influence factor.The value of quantitative class influence factor can be obtained according to actual conditions.The value of qualitative class influence factor is by policymaker
Specific evaluation of estimate is provided after integrated survey actual conditions.
Preset 5 opinion ratings：{G1,G2,G3,G4,G5}={ is very poor, poor, typically, good, very well }, wherein G1And G5Respectively
For the liminal value D of certain assembling influencing factors of quality1With Up limit D5Corresponding opinion rating.In the influence factor
On, the property value for being equivalent to opinion rating is followed successively by { D1,D2,D3,D4,D5}.If opinion rating G1,G2,G3,G4,G5Corresponding effect
It is respectively with value：E(G1)=0, E (G2)=0.25, E (G3)=0.5, E (G4)=0.75, E (G5)=1.Each production batch pair
The actual value of one group of influence factor is answered, actual value here has point value, interval value, three kinds of forms of qualitative value.The influence of n batch
Matrix Env=(env can be used after the normalized processing of factor valuei,t)n×NRepresent, envi,tExpression batch i (i=1,2 ..., n)
Influence factor yt(t=1,2 ..., N) value.
Quantitatively the standardization processing of class influence factor is：
Use βjTo represent that influence factor value belongs to opinion rating GjSubjection degree.
When influence factor value is point value a, if Dj≤a≤Dj+1(j=1,2 ..., 4), then envi,t=βj·E(Gj)+
βj+1·E(Gj+1), its
When influence factor value is interval value [a, b]：
If Dj≤a≤b≤Dj+1(j=1,2 ..., 4), then envi,t=βj·E(Gj)+βj+1·E(Gj+1), wherein
If Dj≤a≤Dj+1And Dj+1≤b≤Dj+2(j=1,2 ..., 3), then envi,t=βj·E(Gj)+βj+1·E(Gj+1)+βj+2·
E(Gj+2), wherein
If Dj≤a≤Dj+1And Dq≤b≤Dq+1(j=1,2 ..., 4, q=1,2 ..., 4, j<Q-1), then envi,t=βj·E
(Gj)+...+βq+1·E(Gq+1), wherein
The opinion rating value of utility of qualitative class influence factor is directly tried to achieve according to influence factor value correspondence.
Step 3.2 is by Env=(envi,t)n×NWith X "=(x "i,j)n×kSynthesized, calculate influence factor under each principal component
Effectiveness value matrix P=(pt,j)N×k, wherein
Step 3.3 introduces influence factor weight, adjusts the degree of concern to influence factor with weight, it is assumed that each influence factor
Respective weights are ωt, ωt∈ (0,1), weighted value is bigger, and explanation policymaker is higher to the trusting degree of the influence factor, uncertain
It is lower.Based on this, the basic probability assignment value to all burnt members is weighted normalized, carries out general under different decision attributes
Rate is distributed, and then obtains the weighting basic probability assignment value of all burnt membersI.e.
Wherein l<2N。
Step 3.4 withInputted as evidence, carry out evidence fusion, i.e.,Wherein K is normaliztion constant, is hadSolve each
The synthesis basic probability assignment value m (A of influence factori)。
Step 3.5 is with based on trusting, interval multiple attribute decision making (MADM) is regular to carry out the influence of Digit Control Machine Tool part assembling process quality
The identification of factor.
On Digit Control Machine Tool part assembling process influencing factors of quality framework of identification Θ, all influence factors are calculated respectively
Belief function value Bel (Ai) and likelihood function value Pl (Ai), wherein belief function valueRepresent to AiIt is total
Degree of belief, likelihood function valueRepresent to AiUncertainty, hereConstruction is trusted interval
[Bel(Ai),Pl(Ai)]。
Trust interval with each influence factor is that foundation is recognized, and obtains the maximum influence factor of possibility.Specifically distinguish
Knowing rule is：
If influence factor AiIt is important to influence factor AjDegree be P (Ai>Aj), if AiAnd AjTrust interval be respectively
[Bel(Ai),Pl(Ai)] and [Bel (Aj),Pl(Aj)], then haveWherein P (Ai>Aj)∈[0,1].That
, influence factor partial ordering relation is：If P (Ai>Aj)>0.5, then AiCompare AjIt is important, it is designated asIf P (Ai>Aj)<0.5,
Then AiThere is no AjIt is important, it is designated as Ai＜ Aj；If P (Ai>Aj)=0.5, then AiWith AjThere is no difference, be designated as Ai~Aj；For any
Three influence factor Ai、AjAnd AkIf, P (Ai>Aj)>0.5 and P (Aj>Ak)>0.5, then AiCompare AkIt is important, it is designated as
The beneficial effects of the invention are as follows：
Method is easy rationally, easily realizes.Take into full account that extensive quality inspection refers in Digit Control Machine Tool part assembling process
The redundancy of data and the feature of Uncertain information system are marked, Digit Control Machine Tool part assembling process influencing factors of quality is realized
Identification, overcomes the shortcomings of manufacture or the assembling process mass analysis method of main flow.
Brief description of the drawings
Fig. 1 is the schematic flow sheet that Digit Control Machine Tool part assembles mass analysis method.
Embodiment
The present invention is described with reference to specific embodiment, so that advantages and features of the invention can be easier to by this area skill
Art personnel understand, apparent are clearly defined so as to be made to protection scope of the present invention.
Digit Control Machine Tool part proposed by the present invention assembles mass analysis method, realizes crowd using PCA first
The dimensionality reduction of multimass test rating, forms the principal component being mutually independent to a certain extent, then fills Digit Control Machine Tool part
The implementation procedure matched somebody with somebody is regarded as a Uncertain information system, introduces evidence theory and enters come " uncertainty " to the information system
Row analysis, so as to realize the identification of influencing factors of quality in the assembling of Digit Control Machine Tool part, as shown in Figure 1.Comprise the following steps that：
Step 1：Quality inspection index to numerical control machine tool component assembling process carries out PCA dimension-reduction treatment.
The PCA dimension-reduction treatment is comprised the following steps that：
Step 1.1 for certain type Digit Control Machine Tool assembling process, if the quality inspection index of its certain part assembling is with master
P are had before componential analysis dimension-reduction treatment.According to the quality record of the part assembling process in n batch, p quality is counted
The qualification rate or average data of test rating.Using each production batch as a sample, then sample data matrix X=is obtained
(xi,t)n×pForWherein xi,tRepresent t (t=1,2 ..., p) individual quality inspection index is i-th
(i=1,2 ..., the n) data in batch, are represented with qualification rate or average value.To X=(xi,t)n×pStandardization processing is carried out, i.e.,
Sample data matrix X'=(x' after step 1.2 pair standardizationi,t)n×p, calculate its correlation matrix COR=
(corr,j)p×p, wherein
Step 1.3 obtains correlation matrix COR=(corr,j)p×pCharacteristic value and by size order arrange, i.e. ev1、
ev2、…、evp, corresponding characteristic vector is followed successively by Ev1、Ev2、…、Evp, wherein Evt=(Evt1,Evt2,...,Evtp), t=
1,2,...,p。
Step 1.4 calculate successively t (t=1,2 ..., p) the contribution rate con of individual quality inspection indext, have
Step 1.5 chooses the individual composition conducts of preceding k (k≤p) that contribution rate of accumulative total reaches more than 85% according to contribution rate size
Principal component, with ev1、ev2、…、evkCorresponding characteristic vector Ev1、Ev2、…、EvkConstruct the matrix EV of p × k dimensions；Make X "=
X'EV, so that n × p standardization sample data matrix X' tieed up to be converted into the new matrix X " of n × k dimensions.K is only remained in X "
Individual principal component, it is achieved thereby that the dimension-reduction treatment of quality inspection index.
Step 2：Set up the possibility set of factors Y={ y of influence assembling quality in Digit Control Machine Tool part assembling process1,
y2,…,yN, there is N=11：y1The physiological characteristic situation of=personnel；y2=psychology of seeking quickness；y3=auxiliary equipment fault level；y4=
Temperature humidity factor；y5=measurer trueness error；y6=protocol adherence situation；y7=degree of fatigue；y8=post is skilled in technique journey
Degree；y9The old degree of wear of=measurer；y10=idea of leaving things to chance；y11=it is mixed into impurity situation.Category is divided into four classes：Human factor
={ y1,y2,y7,y8,y10}；Apparatus factor={ y3,y5,y9}；Management factors={ y6}；Environmental factor={ y4,y11}.By property
It is divided into two classes：Quantitative class factor={ y4,y5,y9}；Qualitative class factor={ y1,y2,y3,y6,y7,y8,y10,y11}。
Step 3：The identification of Digit Control Machine Tool part assembling process influencing factors of quality is carried out based on evidence theory.With Digit Control Machine Tool
Part assembling process is object, and the set of all possible sexual factors of quality will be influenceed to be set to framework of identification, i.e. Θ={ y1,
y2,...,yN, wherein yi(i=1,2 ..., N) it is i-th of influence factor.Set power set 2 is possible on ΘΘCarry out table
Show, when the member in Θ have N number of and each element objectionable intermingling, Θ power set 2ΘElement number be 2N.Digit Control Machine Tool
Evidence in the identification of part assembling process influencing factors of quality is exactly the quality condition occurred in known assembling process, by step
Quality inspection achievement data after 1 dimensionality reduction obtained is constituted.
What the Digit Control Machine Tool part assembling process influencing factors of quality was recognized comprises the following steps that：
Step 3.1 carries out standardization processing to the corresponding influence factor value of n production batch, obtains matrix Env=
(envi,t)n×N, envi,tExpression production batch i (i=1,2 ..., influence factor y n)t(t=1,2 ..., N) value.
The possibility sexual factor of influence assembling quality is divided into quantitative class influence factor and fixed in Digit Control Machine Tool part assembling process
Property class influence factor.The value of quantitative class influence factor can be obtained according to actual conditions, and the value of qualitative class influence factor is by policymaker
Specific evaluation of estimate is provided after integrated survey actual conditions.
Preset 5 opinion ratings：{G1,G2,G3,G4,G5}={ is very poor, poor, typically, good, very well }, wherein G1And G5Respectively
For the liminal value D of certain assembling influencing factors of quality1With Up limit D5Corresponding opinion rating.In the influence factor
On, the property value for being equivalent to opinion rating is followed successively by { D1,D2,D3,D4,D5}.If opinion rating G1,G2,G3,G4,G5Corresponding effect
It is respectively with value：E(G1)=0, E (G2)=0.25, E (G3)=0.5, E (G4)=0.75, E (G5)=1.Each production batch pair
The actual value of one group of influence factor is answered, actual value here has point value, interval value, three kinds of forms of qualitative value.The influence of n batch
Matrix Env=(env can be used after the normalized processing of factor valuei,t)n×NRepresent, envi,tExpression batch i (i=1,2 ..., n)
Influence factor yt(t=1,2 ..., N) value.
Quantitatively the standardization processing of class influence factor is：
Use βjTo represent that influence factor value belongs to opinion rating GjSubjection degree.
When influence factor value is point value a, if Dj≤a≤Dj+1(j=1,2 ..., 4), then envi,t=βj·E(Gj)+
βj+1·E(Gj+1), wherein
When influence factor value is interval value [a, b]：
If Dj≤a≤b≤Dj+1(j=1,2 ..., 4), then envi,t=βj·E(Gj)+βj+1·E(Gj+1), wherein
If Dj≤a≤Dj+1And Dj+1≤b≤Dj+2(j=1,2 ..., 3), then envi,t=βj·E(Gj)+βj+1·E(Gj+1)+βj+2·
E(Gj+2), wherein
If Dj≤a≤Dj+1And Dq≤b≤Dq+1(j=1,2 ..., 4, q=1,2 ..., 4, j<Q-1), then envi,t=βj·E
(Gj)+...+βq+1·E(Gq+1), wherein
The opinion rating value of utility of qualitative class influence factor is directly tried to achieve according to influence factor value correspondence.
Step 3.2 is by Env=(envi,t)n×NWith X "=(x "i,j)n×kSynthesized, calculate influence factor under each principal component
Effectiveness value matrix P=(pt,j)N×k, wherein
Step 3.3 introduces influence factor weight, adjusts the degree of concern to influence factor with weight, it is assumed that each influence factor
Respective weights are ωt, ωt∈ (0,1), weighted value is bigger, and explanation policymaker is higher to the trusting degree of the influence factor, uncertain
It is lower.Based on this, the basic probability assignment value to all burnt members is weighted normalized, carries out general under different decision attributes
Rate is distributed, and then obtains the weighting basic probability assignment value of all burnt membersI.e.
Wherein l<2N。
Step 3.4 withInputted as evidence, carry out evidence fusion, i.e.,Wherein K is normaliztion constant, is hadSolve each
The synthesis basic probability assignment value m (A of influence factori)。
Step 3.5 is with based on trusting, interval multiple attribute decision making (MADM) is regular to carry out the influence of Digit Control Machine Tool part assembling process quality
The identification of factor.
On Digit Control Machine Tool part assembling process influencing factors of quality framework of identification Θ, all influence factors are calculated respectively
Belief function value Bel (Ai) and likelihood function value Pl (Ai), wherein belief function valueRepresent to AiIt is total
Degree of belief, likelihood function valueRepresent to AiUncertainty, hereConstruction is trusted interval
[Bel(Ai),Pl(Ai)]。
Trust interval with each influence factor is that foundation is recognized, and obtains the maximum influence factor of possibility.Specifically distinguish
Knowing rule is：
If influence factor AiIt is important to influence factor AjDegree be P (Ai>Aj), if AiAnd AjTrust interval be respectively
[Bel(Ai),Pl(Ai)] and [Bel (Aj),Pl(Aj)], then haveWherein P (Ai>Aj)∈[0,1].That
, influence factor partial ordering relation is：If P (Ai>Aj)>0.5, then AiCompare AjIt is important, it is designated asIf P (Ai>Aj)<0.5,
Then AiThere is no AjIt is important, it is designated as Ai＜ Aj；If P (Ai>Aj)=0.5, then AiWith AjThere is no difference, be designated as Ai~Aj；For any
Three influence factor Ai、AjAnd AkIf, P (Ai>Aj)>0.5 and P (Aj>Ak)>0.5, then AiCompare AkIt is important, it is designated as
The present invention is described with reference to specific embodiment：
Embodiment：
The lubricating oil part assembling process of certain model Digit Control Machine Tool includes 17 procedures：
(1) part repaiies hair and falls blunt and clean；
(2) outer separator in checking；
(3) one groups of bearing holder (housing, cover)s enter main shaft and bearing block, and nut is compressed；
(4) adjust end cap and load gland；
(5) precision is done at the beginning of spindle assemblies；
(6) main shaft is loaded after rotor heating；
(7) another group of bearing loads main shaft, and nut compresses gland and bearing；
(8) spindle assemblies dynamic balancing；
(9) motor stator loads body shell and positioned with bearing；
(10) body shell leak test is carried out；
(11) main shaft dismounting one group of bearing of right-hand member and end cap；
(12) spindle assemblies are inserted in grinding carriage body shell, and end cap compresses bearing；
(13) magnetic grid is filled；
(14) precision, static rigidity and magnetic bar signal inspection；
(15) auxiliary support assemblies are assembled；
(16) adpting flange axle and coupling spindle；
(17) right-hand member protective cover is installed.
This 17 procedure has 16 quality inspection indexs.
The initial data of the lubricating oil part assembling process quality inspection index of 8 production batch is obtained, as shown in table 1.
The initial data of the lubricating oil part assembling process quality inspection index of 18 production batch of table
Step 1：According to the initial data of the lubricating oil part assembling process quality inspection index of 8 production batch, to lubricating oil
16 quality inspection indexs of part assembling process carry out PCA dimension-reduction treatment.
Using each production batch as a sample, then sample data matrix X=(x are obtainedi,t)8×16, the sample after standardization
Notebook data matrix is X'=(x'i,t)8×16, through PCA dimension-reduction treatment, by the standardization sample data square of 8 × 16 dimensions
X' is converted into new the matrix X ", X " of 8 × 5 dimensions and only remains 5 principal components, as shown in table 2, realizes lubricating oil part and assembled
The dimensionality reduction of journey quality inspection index.
Number of principal components evidence after the dimensionality reduction of table 2
| Production batch | Principal component 1 | Principal component 2 | Principal component 3 | Principal component 4 | Principal component 5 |
| 1 | -1.3206 | 0.4505 | 1.9005 | 2.7904 | -0.0157 |
| 2 | -2.7963 | -1.2688 | -1.2606 | -0.4737 | -0.2135 |
| 3 | 1.2503 | 2.3476 | 0.4210 | -0.7951 | -2.2240 |
| 4 | 2.4006 | -4.2119 | 1.6564 | -0.7803 | 0.0162 |
| 5 | -1.0364 | 2.5937 | 1.6359 | -1.6709 | 1.4925 |
| 6 | 2.2235 | 0.6642 | -1.8009 | 0.4264 | 0.9354 |
| 7 | -3.1493 | -1.2246 | -1.1438 | -0.2377 | -0.2813 |
| 8 | 2.4283 | 0.6495 | -1.4086 | 0.7410 | 0.2903 |
Step 2：Set up the possibility set of factors Y={ y of influence assembling quality in lubricating oil part assembling process1,y2,…,
yN, N=11, wherein y1The physiological characteristic situation of=personnel；y2=psychology of seeking quickness；y3=auxiliary equipment fault level；y4=temperature
Humidity factor；y5=measurer trueness error；y6=protocol adherence situation；y7=degree of fatigue；y8=post skills involved in the labour；y9
The old degree of wear of=measurer；y10=idea of leaving things to chance；y11=it is mixed into impurity situation.Category is divided into four classes：Human factor=
{y1,y2,y7,y8,y10}；Apparatus factor={ y3,y5,y9}；Management factors={ y6}；Environmental factor={ y4,y11}.By property point
For two classes：Quantitative class factor={ y4,y5,y9}；Qualitative class factor={ y1,y2,y3,y6,y7,y8,y10,y11}。
Step 3：The identification of Digit Control Machine Tool part assembling process influencing factors of quality is carried out based on evidence theory.
The influence factor data of the lubricating oil part assembling process of 8 production batch of step 3.1 are as shown in table 3.
The influence factor data of the lubricating oil part assembling process of 38 production batch of table
After normalized processing, the standardization data matrix Env=of 8 batches, 11 influence factors is obtained
(envi,t)8×11。
Step 3.2 is according to 8 production batch tried to achieve, data matrix X "=(x " of 5 principal componentsi,j)8×5, by Env
=(envi,t)8×11With X "=(x "i,j)8×5Synthesized, calculate to obtain the effectiveness value matrix P=of each influence factor under each principal component
(pt,j)11×5, as shown in table 4.
The value of utility of each influence factor under each principal component of table 4
| Influence factor | Principal component 1 | Principal component 2 | Principal component 3 | Principal component 4 | Principal component 5 |
| y1 | 0.4984 | 0.5472 | 0.2511 | 0.3804 | 0.3371 |
| y2 | 0.9597 | 0.1386 | 0.6160 | 0.5678 | 0.1622 |
| y3 | 0.3404 | 0.1493 | 0.4733 | 0.0759 | 0.7943 |
| y4 | 0.5853 | 0.2575 | 0.3517 | 0.0540 | 0.3112 |
| y5 | 0.2238 | 0.8407 | 0.8308 | 0.5308 | 0.5285 |
| y6 | 0.7513 | 0.2543 | 0.5853 | 0.7792 | 0.1656 |
| y7 | 0.2551 | 0.8143 | 0.5497 | 0.9340 | 0.6020 |
| y8 | 0.5060 | 0.2435 | 0.9172 | 0.1299 | 0.2630 |
| y9 | 0.6991 | 0.9293 | 0.2858 | 0.5688 | 0.6541 |
| y10 | 0.8909 | 0.3500 | 0.7572 | 0.4694 | 0.6892 |
| y11 | 0.9593 | 0.1966 | 0.7537 | 0.0119 | 0.7482 |
Step 3.3 sets the respective weights of each influence factor to be ω1,ω2,…,ω11=0.9,0.8,0.6,0.8,0.9,0.5,
0.6,0.8,0.8,0.7,0.6.Calculate to obtain all burnt first weighting basic probability assignment values.By taking principal component 1 as an example, under the decision attribute,
The weighting basic probability assignment value of single factor test is followed successively by：
The weighting basic probability assignment value of four class factors is followed successively by：Human factorApparatus factorManagement factorsEnvironmental factor
And meetIt can equally obtain
The weighting basic probability assignment value of each burnt member under to remaining 4 principal component.
Step 3.4 weighting basic probability assignment value of each burnt member using under 5 principal components is inputted as evidence, is carried out evidence and is melted
Close, the synthesis basic probability assignment value for solving each burnt member is：
m(y1)=0.0499；m(y2)=0.0622；m(y3)=0.0378；m(y4)=0.0489；m(y5)=0.0212；m
(y6)=0.0185；m(y7)=0.0384；m(y8)=0.0739；m(y9)=0.0330；m(y10)=0.0559；m(y11)=
0.0672；m({y1,y2,y7,y8,y10)=0.2809；m({y3,y5,y9)=0.0920；m({y4,y11)=0.1161；m
(Θ)=0.0042.
Step 3.5 is with based on trusting, interval multiple attribute decision making (MADM) is regular to carry out lubricating oil part assembling process influencing factors of quality
Identification.The belief function value Bel (A of each burnt member are calculated respectivelyj) and likelihood function value Pl (Aj), and construct trust interval [Bel
(Aj),Pl(Aj)], as shown in table 5.
Belief function value, likelihood function value and the trust interval of each burnt member of table 5
| Influence factor | Bel(Aj) | Pl(Aj) | [Bel(Aj),Pl(Aj)] |
| y1 | 0.0499 | 0.3350 | [0.0499,0.3350] |
| y2 | 0.0622 | 0.3473 | [0.0622,0.3473] |
| y3 | 0.0378 | 0.1340 | [0.0378,0.1340] |
| y4 | 0.0489 | 0.1692 | [0.0489,0.1692] |
| y5 | 0.0212 | 0.1174 | [0.0212,0.1174] |
| y6 | 0.0185 | 0.0227 | [0.0185,0.0227] |
| y7 | 0.0384 | 0.3235 | [0.0384,0.3235] |
| y8 | 0.0739 | 0.3590 | [0.0739,0.3590] |
| y9 | 0.0330 | 0.1292 | [0.0330,0.1292] |
| y10 | 0.0559 | 0.3410 | [0.0559,0.3410] |
| y11 | 0.0672 | 0.1875 | [0.0672,0.1875] |
| {y1,y2,y7,y8,y10} | 0.5612 | 0.5654 | [0.5612,0.5654] |
| {y3,y5,y9} | 0.1840 | 0.1882 | [0.1840,0.1882] |
| {y4,y11} | 0.2322 | 0.2364 | [0.2322,0.2364] |
The trust interval of single factor test is compared, the good and bad relation for obtaining 11 single factor tests is as shown in table 6.
The good and bad relation of 6 11 single factor tests of table
| y1 | y2 | y3 | y4 | y5 | y6 | y7 | y8 | y9 | y10 | y11 | |
| y1 | 0.5 | 0.4784 | 0.7794 | 0.7057 | 0.8230 | 1.0000 | 0.5202 | 0.4579 | 0.7920 | 0.4895 | 0.6606 |
| y2 | 0.5 | 0.8117 | 0.7361 | 0.8552 | 1.0000 | 0.5417 | 0.4795 | 0.8243 | 0.5110 | 0.6909 | |
| y3 | 0.5 | 0.3931 | 0.5863 | 1.0000 | 0.2507 | 0.1576 | 0.5249 | 0.2048 | 0.3085 | ||
| y4 | 0.5 | 0.6836 | 1.0000 | 0.3226 | 0.2351 | 0.6291 | 0.2795 | 0.4239 | |||
| y5 | 0.5 | 0.9851 | 0.2072 | 0.1141 | 0.4387 | 0.1613 | 0.2319 | ||||
| y6 | 0.5 | 0 | 0 | 0 | 0 | 0 | |||||
| y7 | 0.5 | 0.4377 | 0.7619 | 0.4693 | 0.6322 | ||||||
| y8 | 0.5 | 0.8550 | 0.5316 | 0.7198 | |||||||
| y9 | 0.5 | 0.1922 | 0.2864 | ||||||||
| y10 | 0.5 | 0.6754 | |||||||||
| y11 | 0.5 |
Equally, 4 kinds of classification factors：Human factor={ y1,y2,y7,y8,y10}；Apparatus factor={ y3,y5,y9}；Management because
Element={ y6}；Environmental factor={ y4,y11Good and bad relation it is as shown in table 7.
The good and bad relation of 74 kinds of classification factors of table
| Human factor | Apparatus factor | Management factors | Environmental factor | |
| Human factor | 0.5 | 1.0000 | 1.0000 | 1.0000 |
| Apparatus factor | 0.5 | 1.0000 | 0 | |
| Management factors | 0.5 | 0 | ||
| Environmental factor | 0.5 |
According to identification rule, by table 6,7, in the factor of influence lubricating oil part assembling quality, single factor test is ordered as：The influence factor for coming front three is：y8=post is skilled in technique journey
Degree, y2=psychology of seeking quickness, y10=idea of leaving things to chance；Classification factor is ordered as： Influence maximum is human factor, next to that environmental factor and apparatus factor, management factors influence are minimum.
Claims (1)
1. a kind of Digit Control Machine Tool part assembles mass analysis method, it is characterised in that comprise the following steps：
Step 1：Quality inspection index to numerical control machine tool component assembling process carries out PCA dimension-reduction treatment.
The PCA dimension-reduction treatment is comprised the following steps that：
Step 1.1 for certain type Digit Control Machine Tool assembling process, if the quality inspection index of its certain part assembling is with principal component
P are had before analytic approach dimension-reduction treatment.According to the quality record of the part assembling process in n production batch, p quality is counted
The qualification rate or average data of test rating.Using each production batch as a sample, then sample data matrix X=is obtained
(xi,t)n×pForWherein xi,tRepresent t (t=1,2 ..., p) individual quality inspection index is i-th
(i=1,2 ..., the n) data in batch, are represented with qualification rate or average value.To X=(xi,t)n×pStandardization processing is carried out, i.e.,
Sample data matrix X'=(x' after step 1.2 pair standardizationi,t)n×p, calculate its correlation matrix COR=
(corr,j)p×p, wherein
Step 1.3 obtains correlation matrix COR=(corr,j)p×pCharacteristic value and by size order arrange, i.e. ev1、
ev2、…、evp, corresponding characteristic vector is followed successively by Ev1、Ev2、…、Evp, wherein Evt=(Evt1,Evt2,...,Evtp), t=
1,2,...,p。
Step 1.4 calculate successively t (t=1,2 ..., p) the contribution rate con of individual quality inspection indext, have
Step 1.5 according to contribution rate size, choose contribution rate of accumulative total reach more than 85% preceding k (k≤p) individual composition as it is main into
Point, with ev1、ev2、…、evkCorresponding characteristic vector Ev1、Ev2、…、EvkConstruct the matrix EV of p × k dimensions；X "=X'EV is made,
So as to which n × p standardization sample data matrix X' tieed up to be converted into the new matrix X " of n × k dimensions.Only remained in X " k it is main into
Point, it is achieved thereby that the dimension-reduction treatment of quality inspection index.
Step 2：Set up the possibility set of factors Y={ y of influence assembling quality in Digit Control Machine Tool part assembling process1,y2,…,
yN, there is N=11：y1The physiological characteristic situation of=personnel；y2=psychology of seeking quickness；y3=auxiliary equipment fault level；y4=temperature is wet
Degree factor；y5=measurer trueness error；y6=protocol adherence situation；y7=degree of fatigue；y8=post skills involved in the labour；y9=
The old degree of wear of measurer；y10=idea of leaving things to chance；y11=it is mixed into impurity situation.Category is divided into four classes：Human factor={ y1,
y2,y7,y8,y10}；Apparatus factor={ y3,y5,y9}；Management factors={ y6}；Environmental factor={ y4,y11}.It is divided into two by property
Class：Quantitative class factor={ y4,y5,y9}；Qualitative class factor={ y1,y2,y3,y6,y7,y8,y10,y11}。
Step 3：The identification of Digit Control Machine Tool part assembling process influencing factors of quality is carried out based on evidence theory.With Digit Control Machine Tool part
Assembling process is object, and the set of all possible sexual factors of quality will be influenceed to be set to framework of identification, i.e. Θ={ y1,y2,...,
yN, wherein yi(i=1,2 ..., N) it is i-th of influence factor.Set power set 2 is possible on ΘΘTo represent, work as Θ
In member when have N number of and each element objectionable intermingling, Θ power set 2ΘElement number be 2N.Digit Control Machine Tool part is assembled
Evidence in the identification of procedure quality influence factor is exactly the quality condition occurred in known assembling process, is obtained by step 1
Quality inspection achievement data after dimensionality reduction is constituted.
What the Digit Control Machine Tool part assembling process influencing factors of quality was recognized comprises the following steps that：
Step 3.1 carries out standardization processing to the corresponding influence factor value of n production batch, obtains matrix Env=
(envi,t)n×N, envi,tExpression production batch i (i=1,2 ..., influence factor y n)t(t=1,2 ..., N) value.
The possibility sexual factor of influence assembling quality is divided into quantitative class influence factor and qualitative class in Digit Control Machine Tool part assembling process
Influence factor.The value of quantitative class influence factor can be obtained according to actual conditions.The value of qualitative class influence factor is integrated by policymaker
Specific evaluation of estimate is provided after investigating actual conditions.
Preset 5 opinion ratings：{G1,G2,G3,G4,G5}={ is very poor, poor, typically, good, very well }, wherein G1And G5Respectively certain
Assemble the liminal value D of influencing factors of quality1With Up limit D5Corresponding opinion rating.In the influence factor, etc.
Valency is followed successively by { D in the property value of opinion rating1,D2,D3,D4,D5}.If opinion rating G1,G2,G3,G4,G5Corresponding value of utility
Respectively：E(G1)=0, E (G2)=0.25, E (G3)=0.5, E (G4)=0.75, E (G5)=1.Each production batch correspondence one
The actual value of group influence factor, actual value here has point value, interval value, three kinds of forms of qualitative value.The influence factor of n batch
Matrix Env=(env can be used by being worth after normalized processingi,t)n×NRepresent, envi,tExpression batch i (i=1,2 ..., influence n)
Factor yt(t=1,2 ..., N) value.
Quantitatively the standardization processing of class influence factor is：
Use βjTo represent that influence factor value belongs to opinion rating GjSubjection degree.
When influence factor value is point value a, if Dj≤a≤Dj+1(j=1,2 ..., 4), then envi,t=βj·E(Gj)+βj+1·E
(Gj+1), whereinβj+1=1- βj。
When influence factor value is interval value [a, b]：
If Dj≤a≤b≤Dj+1(j=1,2 ..., 4), then envi,t=βj·E(Gj)+βj+1·E(Gj+1), whereinβj+1=1- βj；
If Dj≤a≤Dj+1And Dj+1≤b≤Dj+2(j=1,2 ..., 3), then envi,t=βj·E(Gj)+βj+1·E(Gj+1)+βj+2·E
(Gj+2), wherein
If Dj≤a≤Dj+1And Dq≤b≤Dq+1(j=1,2 ..., 4, q=1,2 ..., 4, j<Q-1), then envi,t=βj·E(Gj)
+...+βq+1·E(Gq+1), wherein
The opinion rating value of utility of qualitative class influence factor is directly tried to achieve according to influence factor value correspondence.
Step 3.2 is by Env=(envi,t)n×NWith X "=(x "i,j)n×kSynthesized, calculate the effect of influence factor under each principal component
With value matrix P=(pt,j)N×k, wherein
Step 3.3 introduces influence factor weight, adjusts the degree of concern to influence factor with weight, it is assumed that pair of each influence factor
It is ω to answer weightt, ωt∈ (0,1), weighted value is bigger, and explanation policymaker is higher to the trusting degree of the influence factor, uncertain
It is lower.Based on this, the basic probability assignment value to all burnt members is weighted normalized, carries out general under different decision attributes
Rate is distributed, and then obtains the weighting basic probability assignment value of all burnt membersI.e.
Wherein l<2N。
Step 3.4 withInputted as evidence, carry out evidence fusion, i.e.,Wherein K is normaliztion constant, is hadSolve each
The synthesis basic probability assignment value m (A of influence factori)。
Step 3.5 is with based on trusting, interval multiple attribute decision making (MADM) is regular to carry out Digit Control Machine Tool part assembling process influencing factors of quality
Identification.
On Digit Control Machine Tool part assembling process influencing factors of quality framework of identification Θ, the trust of all influence factors is calculated respectively
Functional value Bel (Ai) and likelihood function value Pl (Ai), wherein belief function valueRepresent to AiTotal trust
Degree, likelihood function valueRepresent to AiUncertainty, hereConstruction trusts interval [Bel
(Ai),Pl(Ai)]。
Trust interval with each influence factor is that foundation is recognized, and obtains the maximum influence factor of possibility.Specific identification rule
It is then：
If influence factor AiIt is important to influence factor AjDegree be P (Ai>Aj), if AiAnd AjTrust interval be respectively [Bel (Ai),Pl
(Ai)] and [Bel (Aj),Pl(Aj)], then have
Wherein P (Ai>Aj)∈[0,1].So, influence factor partial ordering relation is：If P (Ai>Aj)>0.5, then AiCompare AjIt is important, it is designated asIf P (Ai>Aj)<0.5, then AiThere is no AjIt is important, it is designated asIf P (Ai>Aj)=0.5, then AiWith AjIt is not poor
Not, it is designated as Ai~Aj；For any three influence factor Ai、AjAnd AkIf, P (Ai>Aj)>0.5 and P (Aj>Ak)>0.5, then AiCompare Ak
It is important, it is designated as
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710434211.XA CN107194606A (en) | 2017-06-09 | 2017-06-09 | A kind of Digit Control Machine Tool part assembles mass analysis method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710434211.XA CN107194606A (en) | 2017-06-09 | 2017-06-09 | A kind of Digit Control Machine Tool part assembles mass analysis method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN107194606A true CN107194606A (en) | 2017-09-22 |
Family
ID=59877845
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710434211.XA Pending CN107194606A (en) | 2017-06-09 | 2017-06-09 | A kind of Digit Control Machine Tool part assembles mass analysis method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107194606A (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108073074A (en) * | 2017-12-15 | 2018-05-25 | 西安交通大学 | A kind of assembling quality evaluating method based on Machine Tool Feeding System kinetic characteristic |
| CN108956122A (en) * | 2018-08-22 | 2018-12-07 | 清华大学 | A kind of assembling quality detection method based on structural dynamic characteristics |
| CN109711663A (en) * | 2018-11-15 | 2019-05-03 | 国网山东省电力公司淄博供电公司 | Substation's oil-immersed transformer status assessment and modification method and system based on big data analysis |
| CN110363374A (en) * | 2019-05-16 | 2019-10-22 | 南京理工大学 | A kind of quantitative analysis method of substandard product influence factor |
-
2017
- 2017-06-09 CN CN201710434211.XA patent/CN107194606A/en active Pending
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108073074A (en) * | 2017-12-15 | 2018-05-25 | 西安交通大学 | A kind of assembling quality evaluating method based on Machine Tool Feeding System kinetic characteristic |
| CN108073074B (en) * | 2017-12-15 | 2019-12-24 | 西安交通大学 | Assembly quality control method based on motion characteristics of machine tool feeding system |
| CN108956122A (en) * | 2018-08-22 | 2018-12-07 | 清华大学 | A kind of assembling quality detection method based on structural dynamic characteristics |
| CN109711663A (en) * | 2018-11-15 | 2019-05-03 | 国网山东省电力公司淄博供电公司 | Substation's oil-immersed transformer status assessment and modification method and system based on big data analysis |
| CN109711663B (en) * | 2018-11-15 | 2021-03-02 | 国网山东省电力公司淄博供电公司 | Transformer substation oil-immersed transformer state evaluation and correction method and system based on big data analysis |
| CN110363374A (en) * | 2019-05-16 | 2019-10-22 | 南京理工大学 | A kind of quantitative analysis method of substandard product influence factor |
| CN110363374B (en) * | 2019-05-16 | 2022-08-12 | 南京理工大学 | Quantitative analysis method for unqualified product influence factors |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107194606A (en) | A kind of Digit Control Machine Tool part assembles mass analysis method | |
| Chang et al. | Applying the concept of exponential approach to enhance the assessment capability of FMEA | |
| CN108664010A (en) | Generating set fault data prediction technique, device and computer equipment | |
| CN103247008A (en) | Quality evaluation method of electricity statistical index data | |
| Choi et al. | A simulation study of an automotive foundry plant manufacturing engine blocks | |
| CN110705887A (en) | Low-voltage transformer area operation state comprehensive evaluation method based on neural network model | |
| CN110942084A (en) | Loss reduction measure making method based on synchronous line loss abnormity identification | |
| CN110321631A (en) | One kind is towards marine diesel fuselage qualitative character process reliability appraisal procedure | |
| CN116485267B (en) | Intelligent building health state evaluation system based on multidimensional data | |
| Zhang et al. | An early fault elimination method of computerized numerical control machine tools | |
| CN111445037A (en) | Steam turbine fault diagnosis method, system and equipment based on evidence theory | |
| Nurhadi et al. | The Effect of Experience and Professionalism on Auditor's Performance in Jakarta and Surabaya Accounting Firms during Pandemic the COVID–19 | |
| CN112561231A (en) | Switch cabinet quality judgment method and system | |
| CN106295712A (en) | A kind of fault detection method and system | |
| CN104933258A (en) | Design method for improving reliability of horizontal machining centre | |
| CN113094816B (en) | Method for constructing comprehensive working condition vibration spectrum and long-life test spectrum of armored vehicle | |
| CN102280883B (en) | Wide area pattern analysis method for dynamic simulation validation of power system | |
| Ding et al. | Key performance indicators for regional construction supervision systems in China | |
| Schuh et al. | Consideration of risk management in global production footprint design | |
| Ngadiman et al. | Relationship between human errors in maintenance and overall equipment effectiveness in food industries | |
| CN114331175A (en) | Centralized statistical evaluation method and system for urban safety performance data | |
| Dudewicz | Basic statistical methods | |
| Kolodenkova et al. | Method for selecting diagnostic parameters using information technologies under a set of various types of information | |
| Edossa et al. | Reducing the defect rate of final products through spc tools: A case study on ammunition cartridge production factory | |
| Lu et al. | A Quality Evaluation Method for Software Testing about Safety-Critical Software |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| WD01 | Invention patent application deemed withdrawn after publication | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20170922 |