CN104268392A - Manufacturing process product reliability decline risk evaluating method based on quality deviation - Google Patents

Abstract

The invention discloses a manufacturing process product reliability decline risk evaluating method based on quality deviation. The manufacturing process product reliability decline risk evaluating method based on the quality deviation includes that step 1, building a manufacturing process product reliability decline risk evaluation index system; step 2, collecting historical data under a stable product manufacturing state according to the index system of the step 1; 3, using a structural equation to inspect the suitability between the index system of the step 1 and the collected historical data of the step 2, if passing the suitability inspection, entering step 4; otherwise, returning to the step 1 to correct the index system; 4, using the index system which passes the inspection for the normal batch production of products, and calculating the manufacturing process product reliability decline risk RS of each batch; 5, calculating the confidence interval of the manufacturing process product reliability decline risk RS of each batch; 6, using the manufacturing process product reliability decline risk RS and confidence interval of each batch to monitor the reliability decline risk undulation situation of each batch under the batch production state of products.

Description

Based on the manufacture process product reliability downslide risk evaluating method of mass deviation

Technical field

The invention provides a kind of manufacture process product reliability downslide risk evaluating method based on mass deviation, belong to risk assessment field.

Background technology

Risk assessment is on the basis identifying venture influence factor, sets up risk evaluation system and draws risk quantum, and with relevant evaluation standard comparing, to determine the process that whether will take appropriate measures.

Reliability weighs the most general and the most effective index whether product operational phase can complete expection requirement, refers to that product completes the ability of predetermined function under defined terms and in official hour.The reliability of product is decided by design, be formed at the fabrication phase, under the prerequisite that design objective is constant, the mass deviations such as manufacture process people, machine, material, method, ring, survey are the risk sources that the product inherent reliability causing the fabrication phase to be formed is degenerated, and namely manufacture process product reliability produces the main cause of downslide risk relative to design objective.For the afterwards reliability estimation method of convectional reliability research based on user's usage data, fail to set up Manufacturing Process product reliability downslide prevention of risk evaluation method effectively, cause goods producer can not understand in time the reliability state of manufactured product, maintenance support activity can not be carried out targetedly, and through often there will be serious quality problems after causing product to come into operation, the present invention is on the basis of clear and definite reliability downslide risk reason, make full use of known quality inspection codominance data a large amount of in manufacture process, give the manufacture process product reliability downslide risk evaluating method based on mass deviation, the downslide prevention of risk carrying out manufacture process product reliability for the manufacture of business is evaluated and monitoring.

Summary of the invention

(1) object of the present invention:

The disappearance of evaluation method of the downslide risk set up for manufacture process product reliability is failed for existing invention and research, the present invention bases oneself upon the fabrication phase forming product reliability, provides the manufacture process product reliability downslide risk evaluating method based on mass deviation.In manufacture process, mass deviation is larger, and the probability of manufacture process product reliability downslide risk also will increase, and the present invention is setting up manufacture process product reliability downslide risk (the i.e. R by mass deviation quantitative expression _s) basis on, add the test stage that can improve result accuracy simultaneously, namely structure based equation the assessment indicator system method of inspection and based on each batch of R under the mass production environment of fiducial interval _sthe method for supervising of fluctuation.Manufacture process product reliability downslide risk evaluating method based on mass deviation provided by the invention, take into full account the overall process qualitative data of the fabrication phase that product reliability is formed and provided the effective test stage improving result accuracy, the downslide risk of evaluation manufacture process product reliability that can be more timely, accurate and sensitive.

(2) technical scheme:

The invention provides a kind of manufacture process product reliability downslide risk evaluating method based on mass deviation, the basic assumption of proposition is as follows:

Suppose that 1 process to measure;

Suppose that 2 process detected values continuously and Normal Distribution;

Suppose that in 3 manufacture processes, reliability design scheme is not changed.

Based on above-mentioned hypothesis, the manufacture process product reliability downslide risk evaluating method based on mass deviation that the present invention proposes, its step as shown in Figure 1, is divided into following 6 steps:

Step 1 sets up manufacture process product reliability downslide Risk Assessment Index System;

Step 2 collects the historical data under product manufacturing stable state according to the index system in step 1;

Step 3 utilizes the suitability of the historical data gathered in the index system and step 2 set up in equation of structure checking procedure 1, if suitability is upchecked, then enters step 4; If suitability inspection is not passed through, then get back to step 1, index system is revised;

Step 4 calculates the manufacture process product reliability downslide risk R of each batch under the index system after upchecking in step 3 is applied to product regular lot production status _s;

Step 5 calculates the manufacture process product reliability downslide risk R of each batch _sfiducial interval;

The manufacture process product reliability downslide risk R that step 6 computation obtains _sand the fluctuation situation of each batch of reliability downslide risk under fiducial interval monitoring product batch production state.

Wherein, described in step 1 " manufacture process product reliability downslide Risk Assessment Index System ", it is formed as shown in Figure 2, refers to by manufacture process procurement risk (V _p), processing risk (V _m), assembling risk (V _a), test risk (V _t) form contain the manufacture process reliability downslide Risk Assessment Index System of 4 grades manufacturing overall process mass deviation; In index system, the IIIth grade of index is all kinds of Critical to qualities of manufacture process, belongs to initial data source; IIth grade of index outsourcing piece yield rate, product composition outsourcing piece ratio etc. are procurement risk (V _p), processing risk (V _m), assembling risk (V _a), test risk (V _t) subordinate's index; The process of establishing of this manufacture process product reliability downslide Risk Assessment Index System is as follows: (1) determines security risk Evaluation Strategy; (2) combing manufacture course of products; (3) the Critical to quality inventory in the stages such as buying, processing, assembling and test is determined.

Wherein, described in step 2 " collecting the historical data under product manufacturing stable state ", refer to when manufacture course of products is in state of statistical control, collect the process of this batch products historical data by the index system set up in step 1, main collection the IIIth grade of indicator measurements.

Wherein, described in step 3 " utilizing the suitability of the historical data gathered in the index system and step 2 set up in equation of structure checking procedure 1 ", refers to covariance matrix implicit in the equation of structure method test rating system utilized in multivariate data analysis and the adaptive degree between historical data covariance matrix S, adaptive degree is higher, represents that the index system set up is more consistent with actual, historical data, namely more close with S; Concrete, weigh by adaptive index with the adaptive degree of S; Described " equation of structure " refers to that the covariance matrix based on variable carrys out a kind of statistical method of relation between situational variables.

Wherein, describedly in step 4 " the manufacture process product reliability downslide risk R of each batch is calculated _s", computation process refers to and utilizes Process Capability with Multivariate index, weighted geometric mean and Hierarchy Analysis Method to calculate the IIth grade, I grade and R in accompanying drawing 1 respectively _sthe risk level of level, and obtain the manufacture process product reliability downslide risk R of this batch _svalue, Risk Calculation formula at different levels is as follows:

$R_S^{\mathrm{II}}={\left[\frac{\underset{i=1}{\overset{m}{\mathrm{\Π}}}({\mathrm{USL}}_i-{\mathrm{LSL}}_i)}{\underset{i=1}{\overset{m}{\mathrm{\Π}}}({\mathrm{UPL}}_i-{\mathrm{LPL}}_i)}\right]}^{\frac{1}{m}}---\left(1\right)$

In formula: USL _iand LSL _ifor the upper and lower limit of qualitative data code requirement, UPL _iand LPL _ifor corresponding to the upper and lower limit of the process area revised in actual production, the volume ratio of both utilizations calculates m represents the number of KQCs, i=1, and 2 ..., m.

$R_S^I={\left[\underset{j=1}{\overset{n}{\mathrm{\Π}}}{\left(R_{S\left(j\right)}^{\mathrm{II}}\right)}^{w_j}\right]}^{\frac{1}{\underset{j=1}{\overset{n}{\mathrm{\Σ}}}{w}_j}}---\left(2\right)$

In formula: w _jrepresent a jth index in assessment indicator system II grade weight (j=1,2 ..., n), w _jfor scope is the integer of 1 ~ 5.

$\begin{array}{c}{R}_S=\underset{k=1}{\overset{4}{\mathrm{\Σ}}}{R}_{S\left(k\right)}^I{w}_k\\ =R_{S\left(1\right)}^I{w}_1+R_{S\left(2\right)}^I{w}_2+R_{S\left(3\right)}^I{w}_3+R_{S\left(4\right)}^I{w}_4\end{array}---\left(3\right)$

In formula: represent the buying of manufacture process, processing, A&T risk respectively, k=1,2,3,4, w _krepresent respective weight.

Wherein, describedly in steps of 5 " the manufacture process product reliability downslide risk R of each batch is calculated _sfiducial interval ", refer to the many groups Normal Type data utilizing the mode of Monte-Carlo analogue simulation to produce to meet this batch of qualitative character, calculate the R often organizing emulated data _svalue also sorts, and obtains final this batch of R according to given degree of confidence _sfiducial interval, computing formula is as follows:

$[k_L,k_U]=[\left(\frac{1-\mathrm{\α}}{2}\right)n+1,n\left(\frac{1+\mathrm{\α}}{2}\right)]---\left(4\right)$

In formula: [k _l, k _u] be R _sthe lower limit of fiducial interval with the upper limit correspond respectively to value corresponding on sorting position, α is the degree of confidence of specifying.

Wherein, described in step 6 " the manufacture process product reliability downslide risk R that computation obtains _sand the fluctuation situation of each batch of reliability downslide risk under fiducial interval monitoring product batch production state ", refer to according to R _sevaluation criterion weigh this batch of R _srisk level while, utilize the method for power gathering-point analysis monitoring batch production process can characterize the parameter ξ of the overlay information of each batch of fiducial interval and the fluctuation situation of positional information, whether offset to monitor manufacture process, parameter ξ computing formula is as follows:

In formula: fiducial interval is the interval calculated in formula (4), accepts the length that burst length is the fiducial interval obtained under stable state.

(3) the manufacture process product reliability downslide risk evaluating method based on mass deviation of the present invention, its using method is as follows:

Step (1) collects product manufacturing stable state historical data;

The suitability of the index system that step (2) utilizes equation of structure inspection institute to set up and history steady state data, suitability is upchecked, then enter step (3); The test fails for suitability, then revise index system, re-starts step (1) and (2);

Step (3) draws the R of each batch of manufacture course of products _svalue and fiducial interval thereof, the fluctuation situation of reliability downslide risk in monitoring manufacture process, if R _svalue or fiducial interval produce unusual fluctuations, then enter step (4); If no exceptions, then continue monitoring;

Step (4) judges that whether real manufacture process is abnormal, if unusual fluctuations are false police, then reenters step (3); If manufacture process occurs abnormal, then analyze reason, adjustment repair process, makes it reenter normal condition.No matter whether unusual fluctuations are that false is alert in step (3), all should minute book time abnormal occur time, disposal route, responsible person.

Wherein, " index system utilizing equation of structure inspection institute to set up and the suitability of history steady state data " described in step (2) refers to and checks the suitability index in the equation of structure whether to meet regulation requirement, has one not meet regulation and requires then to check not pass through.

Wherein, " the R described in step (3) _svalue or fiducial interval produce unusual fluctuations " refer to R _svalue exceeds the fiducial interval generation tendency skew of regulation evaluation criterion or each batch.

(4) advantage and effect:

I. for the reliability assessment of property afterwards of existing invention and research, fail to set up the real-time estimating method for the downslide risk of manufacture process product reliability, the present invention bases oneself upon the fabrication phase forming product reliability, provide the manufacture process product reliability downslide risk evaluating method based on mass deviation, for before product comes into operation, just the Possible waves situation grasping product reliability in advance in the fabrication phase provides effective tool support, achieves the target in manufacture process real-time monitoring product reliability downslide risk.

Ii. after manufacture process value or fiducial interval generation unusual fluctuations are reported to the police, contribute to equally diagnosing manufacture process, reject the abnormal factors affecting product reliability and glide, reach the lasting guarantee of product reliability in the fabrication phase.

Accompanying drawing explanation

Fig. 1 is the FB(flow block) of the method for the invention

Fig. 2 is manufacture process product reliability downslide Risk Assessment Index System figure

Fig. 3 is structural equation model normalizing parameter figure

Fig. 4 is fiducial interval Monte-Carlo analog result

Fig. 5 is density sequence { ρ _npower gathering-point

In figure, symbol description is as follows:

R _srefer to manufacture process product reliability downslide risk

V _prefer to procurement risk

V _mrefer to processing risk

V _arefer to assembling risk

V _trefer to test risk

λ refers to horizontal energy accumulation

δ refers to the variable characterizing fiducial interval overlay information and positional information

{ ρ _nrefer to the density sequence of δ

Embodiment

Below in conjunction with accompanying drawing and example, the present invention is described in further details.

Following instance carries out according to the flow process shown in accompanying drawing 1, specifically described below:

A kind of manufacture process product reliability downslide risk evaluating method based on mass deviation of the present invention, as shown in Figure 1, its step is as follows:

Step 1 sets up manufacture process product reliability downslide Risk Assessment Index System.According to accompanying drawing 2, for certain electronic product batch production process, set up manufacture process reliability downslide risk initial evaluation index system as shown in table 1:

Certain electronic product manufacture process reliability downslide risk initial evaluation index system of table 1

Wherein the index such as V111, V112, V113 corresponds to the Critical to quality index of the assessment indicator system the IIIth grade in accompanying drawing 2.

Step 2 collects the historical data under product manufacturing stable state according to the index system that step 1 is set up.Usually can inquire about the modes such as passing quality record here, the data volume of collection, usually between 20-200, as space is limited, only lists 20 item number certificates of procurement risk 6 indexs below, as shown in table 2:

Table 2 product manufacturing stable state historical data

Step 3 utilizes the suitability of the historical data gathered in the index system and step 2 set up in equation of structure checking procedure 1.First, the inspection of reliability and validity need be carried out to qualitative data under the stable state gathered, reliability and validity is realized by Cronbach's side reaction coefficient and factor analysis exploratory, and through SPSS19.0 software test, data sample reliability and validity all meets the requirement carrying out equation of structure inspection.

Afterwards, the structural equation model of the index system set up in application AMOS17.0 software building step 1, and to wherein implicit covariance matrix and the adaptive degree between the data sample covariance matrix S gathered in step 2 is tested.As shown in Figure 3, procurement risk, processing risk, assembling risk and test risk are to R _spath coefficient value be respectively 0.576,0.702,0.601 and 0.571, and all have passed significance test, show that this four indices is to R _sthere is significant impact, meet and suppose above.Meanwhile, risk is processed with assembling risk to R _shave the greatest impact, be the key risk in this manufacture course of products.

Adopt the most general overall fit goodness index as the adaptive index of structural equation model, comprise following two classes: (1) is Fitting optimization index definitely: chi-square value, GFI, RMSEA; (2) increment goodness of fit index: CFI, IFI, TLI.The AMOS17.0 running software result of structural equation model suitability index is as shown in table 3, and wherein, as key index, RMSEA characterizes the implicit covariance matrix of each degree of freedom drag and the average difference values between data sample covariance matrix S, it has been generally acknowledged that RMSEA<0.05 model will have extraordinary adaptive degree, in table 3, RMSEA equals 0.048 and meets the requirements.In table 3, all the other indices also all reach acceptable level, and the assessment indicator system set up in expression table 1 is reasonable and satisfactory by equation of structure inspection, upchecks.

Table 3 equation of structure Fitness Test result

Step 4 determines manufacture process product reliability downslide risk R _svalue.Process Capability with Multivariate index, weighted geometric mean and Hierarchy Analysis Method is utilized to calculate the IIth grade, I grade and R in accompanying drawing 2 respectively _sthe risk level of level, and obtain the manufacture process product reliability downslide risk R of this batch _svalue.

Process Capability with Multivariate index is utilized to calculate the manufacture risk level of the IIth grade

$R_S^{\mathrm{II}}={\left[\frac{\underset{i=1}{\overset{m}{\mathrm{\&Pi;}}}({\mathrm{USL}}_i-{\mathrm{LSL}}_i)}{\underset{i=1}{\overset{m}{\mathrm{\&Pi;}}}({\mathrm{UPL}}_i-{\mathrm{LPL}}_i)}\right]}^{\frac{1}{m}}---\left(6\right)$

USL in above formula _iand LSL _ifor the upper and lower limit of qualitative data code requirement, UPL _iand LPL _ifor corresponding to the upper and lower limit of the process area revised in actual production, the volume ratio of both utilizations calculates the number of m representation quality characteristic, i=1,2 ..., m.

Utilize the buying of weighted geometric mean Calculation Estimation index system I grade, processing, A&T risk

$R_S^I={\left[\underset{j=1}{\overset{n}{\mathrm{\&Pi;}}}{\left(R_{S\left(j\right)}^{\mathrm{II}}\right)}^{w_j}\right]}^{\frac{1}{\underset{j=1}{\overset{n}{\mathrm{\&Sigma;}}}{w}_j}}---\left(7\right)$

Wherein w _jrepresent a jth index in assessment indicator system II grade weight (j=1,2 ..., n), w _jfor scope is the integer of 1 ~ 5.

Finally, analytical hierarchy process determination manufacture process product reliability downslide risk R is utilized _s:

$\begin{array}{c}{R}_S=\underset{k=1}{\overset{4}{\mathrm{\&Sigma;}}}{R}_{S\left(k\right)}^I{w}_k\\ =R_{S\left(1\right)}^I{w}_1+R_{S\left(2\right)}^I{w}_2+R_{S\left(3\right)}^I{w}_3+R_{S\left(4\right)}^I{w}_4\end{array}---\left(8\right)$

Wherein represent the buying of manufacture process, processing, A&T risk respectively, k=1,2,3,4.

Manufacture process product reliability downslide risk R _svalue larger, represent manufacture process product reliability and will face larger downslide risk probability and cause larger reliability downslide degree.Definition R _d(Degree) represent the grade of manufacture process product reliability downslide risk, be shown below:

$R_D=\left\{\begin{array}{cc}1,& 1.67<R_S\\ 2,& 1.33<R_S\&le;1.67\\ 3,& 1<R_S\&le;1.33\\ 4,& 0.67<R_S\&le;1\\ 5,& R_S\&le;0.67\end{array}\right.---\left(9\right)$

R _ddistribution range is 1 ~ 5, R _dbe worth larger expression manufacture process product reliability downslide risk class higher, the risk probability namely glided and degree of risk higher.Equally, the risk probability glided and degree of risk are also divided into 5 grades, respectively corresponding glide risk probability " low ", " lower ", " height ", " higher ", " high ", and the degree of risk " little " of downslide, " less ", " generally ", " larger ", " seriously ".

In conjunction with (6)-(8) formula, obtain batch R of 1 (this electronic product batch production process 35 totally batches) _svalue:

$\begin{array}{c}{R}_S=R_{S\left(1\right)}^I{w}_1+R_{S\left(2\right)}^I{w}_2+R_{S\left(3\right)}^I{w}_3+R_{S\left(4\right)}^I{w}_4\\ =1.098\&times;0.131+0.908\&times;0.484+1.027\&times;0.271+1.093\&times;0.114\\ =0.986\end{array}$

According to the criterion of formula (9), R _dbe 4 grades, represent that the risk probability that this batch manufacturing process product reliability glides and degree of risk all belong to the level of " less ", in tolerance interval.

Step 5 determines manufacture process product reliability downslide risk R _sfiducial interval.According to the feature of certain batch of actual production data, utilize the mode of Monte-carlo analogue simulation to produce and meet the n group Normal Type simulated data of this batch of qualitative character (when production is stablized, Quality Checkout Data is generally normal distribution), and calculate the R often organizing qualitative data _svalue.To R _ssort from small to large, thus under appointment confidence alpha, R _sthe lower limit of fiducial interval with the upper limit correspond respectively to sorting position [k _l, k _u] upper corresponding value:

$[k_L,k_U]=[\left(\frac{1-\mathrm{\&alpha;}}{2}\right)n+1,n\left(\frac{1+\mathrm{\&alpha;}}{2}\right)]---\left(10\right)$

When given degree of confidence requires to be 80%, Matlab is utilized to carry out 10000 analog computation R _svalue (when number realization is 10000, result be tending towards convergence, as shown in Figure 4), and result is sorted, according to formula (10), sort the 1001st and the 9000th corresponding value [0.9603,1.0016], R in required batch 1 is _sthe lower limit of fiducial interval and the upper limit.

The manufacture process product reliability downslide risk R that step 6 computation obtains _sand the fluctuation situation of each batch of reliability downslide risk under fiducial interval monitoring product batch production state.According to the real time data of the historical production data under product stable state and normal production batch, under stable state can be obtained respectively and normal production batch R _sfiducial interval and the fiducial interval obtained under stable state is defined as reception interval.Meanwhile, be the fluctuation situation utilizing fiducial interval to monitor each batch of reliability downslide risk, first, introduce variable ξ and monitor normal production batch R _sfiducial interval relative to the overlapping cases of reception interval:

ξ value is larger, and show that two interval overlaps are more, consistance is higher.Obviously, ξ can reflect two interval conforming information, but the relation both failing to reflect between position.For this reason, deflection function f (α is then introduced _r) investigate the information of both position relativity shifts:

Wherein α _rrepresent the mid point of normal production batch fiducial interval, α represents the mid point of reception interval, α _rdepart from α larger, f (α _r) by more close-1 or 1.

Finally, in conjunction with overlapping cases and the positional information of each batch of fiducial interval and reception interval, by variable δ=ξ × f (α _r) the power gathering-point analysis of numeric distribution, can reflect and normally produce each batch of R _sfiducial interval relative to reception interval under stable state tendency and fluctuation situation.

For this electronic product, with the fiducial interval [0.9601 of history stable state production data, 1.0014] as reception interval, and contrast the lower fiducial interval of 35 different batches of normal production status and the overlapping of reception interval and positional information, to δ=ξ × f (α according to formula (11), (12) _r) carry out power gathering-point analysis.

Concrete, at the distributed area [-4 × 10 of δ ^-3, 4 × 10 ^-3] in scope, computational length is 0.5 × 10 ^-3the M that counts contained by each sub-range _n, thus try to achieve the density sequence { ρ of δ _n(ρ _n=M _n/ 35, n=1,2 ..., 35), and obtain { ρ accordingly _nλ horizontal energy accumulation, as shown in Figure 5.

In 35 batches, the average of variable ξ is 95.61%, and represent that the overlapping degree of each batch of fiducial interval is higher, consistance is better.But can find out from accompanying drawing 5, the distribution trend of δ is many between [0,0.5], and general trend is to the left.This shows that manufacture process entirety there occurs certain departing from, and needs to be further improved follow-up.

Claims (8)

1., based on a manufacture process product reliability downslide risk evaluating method for mass deviation, suppose as follows:

Suppose that 1 process to measure;

Suppose that 2 process detected values continuously and Normal Distribution;

Suppose that in 3 manufacture processes, reliability design scheme is not changed;

Based on above-mentioned hypothesis, a kind of manufacture process product reliability downslide risk evaluating method based on mass deviation of the present invention, is characterized in that: be divided into following 6 steps:

Step 1 sets up manufacture process product reliability downslide Risk Assessment Index System;

Step 2 collects the historical data under product manufacturing stable state according to the index system in step 1;

Step 3 utilizes the suitability of the historical data gathered in the index system and step 2 set up in equation of structure checking procedure 1, if suitability is upchecked, then enters step 4; If suitability inspection is not passed through, then get back to step 1, index system is revised;

Step 4 calculates the manufacture process product reliability downslide risk R of each batch under the index system after upchecking in step 3 is applied to product regular lot production status _s;

Step 5 calculates the manufacture process product reliability downslide risk R of each batch _sfiducial interval;

The manufacture process product reliability downslide risk R that step 6 computation obtains _sand the fluctuation situation of each batch of reliability downslide risk under fiducial interval monitoring product batch production state.

2. a kind of manufacture process product reliability downslide risk evaluating method based on mass deviation according to claim 1, it is characterized in that: described in step 1 " manufacture process product reliability downslide Risk Assessment Index System ", refers to by manufacture process procurement risk V _p, processing risk V _m, assembling risk V _a, test risk V _tthe manufacture process reliability downslide Risk Assessment Index System of 4 grades containing manufacture overall process mass deviation of composition; In index system, the IIIth grade of index is all kinds of Critical to qualities of manufacture process, belongs to initial data source; IIth grade of index outsourcing piece yield rate, product composition outsourcing piece ratio are procurement risk V _p, processing risk V _m, assembling risk V _a, test risk V _tsubordinate's index.

3. a kind of manufacture process product reliability downslide risk evaluating method based on mass deviation according to claim 1, it is characterized in that: described in step 1 " setting up manufacture process product reliability downslide Risk Assessment Index System ", its process of establishing is as follows: (1) determines security risk Evaluation Strategy; (2) combing manufacture course of products; (3) determine to purchase, processing, assembling and the Critical to quality inventory of test phase.

4. a kind of manufacture process product reliability downslide risk evaluating method based on mass deviation according to claim 1, it is characterized in that: described in step 2 " collecting the historical data under product manufacturing stable state ", refer to when manufacture course of products is in state of statistical control, collect the process of this batch products historical data by the index system set up in step 1, collect the IIIth grade of indicator measurements.

5. a kind of manufacture process product reliability downslide risk evaluating method based on mass deviation according to claim 1, it is characterized in that: described in step 3 " utilizing the suitability of the historical data gathered in the index system and step 2 set up in equation of structure checking procedure 1 ", refers to covariance matrix implicit in the equation of structure method test rating system utilized in multivariate data analysis and the adaptive degree between historical data covariance matrix S, adaptive degree is higher, represents that the index system set up is more consistent with actual, historical data, namely more close with S; Concrete, weighed by adaptive index with the adaptive degree of S; Described " equation of structure " refers to that the covariance matrix based on variable carrys out a kind of statistical method of relation between situational variables.

6. a kind of manufacture process product reliability downslide risk evaluating method based on mass deviation according to claim 1, is characterized in that: describedly in step 4 " calculate the manufacture process product reliability downslide risk R of each batch _s", computation process refers to and utilizes Process Capability with Multivariate index, weighted geometric mean and Hierarchy Analysis Method to calculate the IIth grade, I grade and R respectively _sthe risk level of level, and obtain the manufacture process product reliability downslide risk R of this batch _svalue, Risk Calculation formula at different levels is as follows:

$R_S^{\mathrm{II}}={\left[\frac{\underset{i=1}{\overset{m}{\mathrm{\&Pi;}}}({\mathrm{USL}}_i-{\mathrm{LSL}}_i)}{\underset{i=1}{\overset{m}{\mathrm{\&Pi;}}}({\mathrm{UPL}}_i-{\mathrm{LPL}}_i)}\right]}^{\frac{1}{m}}---\left(1\right)$

In formula: USL _iand LSL _ifor the upper and lower limit of qualitative data code requirement, UPL _iand LPL _ifor corresponding to the upper and lower limit of the process area revised in actual production, the volume ratio of both utilizations calculates m represents the number of KQCs, i=1, and 2 ..., m;

$R_S^I={\left[\underset{j=1}{\overset{n}{\mathrm{\&Pi;}}}{\left(R_{S\left(j\right)}^{\mathrm{II}}\right)}^{w_j}\right]}^{\frac{1}{\underset{j=1}{\overset{n}{\mathrm{\&Sigma;}}}{w}_j}}---\left(2\right)$

In formula: w _jrepresent a jth index in assessment indicator system II grade weight (j=1,2 ..., n), w _jfor scope is the integer of 1 ~ 5;

$\begin{array}{c}{R}_S=\underset{k=1}{\overset{4}{\mathrm{\&Sigma;}}}{R}_{S\left(k\right)}^I{w}_k\\ =R_{S\left(1\right)}^I{w}_1+R_{S\left(2\right)}^I{w}_2+R_{S\left(3\right)}^I{w}_3+R_{S\left(4\right)}^I{w}_4\end{array}---\left(3\right)$

In formula: represent the buying of manufacture process, processing, A&T risk respectively, k=1,2,3,4, w _krepresent respective weight.

7. a kind of manufacture process product reliability downslide risk evaluating method based on mass deviation according to claim 1, is characterized in that: describedly in steps of 5 " calculate the manufacture process product reliability downslide risk R of each batch _sfiducial interval ", refer to the many groups Normal Type data utilizing the mode of Monte-Carlo analogue simulation to produce to meet this batch of qualitative character, calculate the R often organizing emulated data _svalue also sorts, and obtains final this batch of R according to given degree of confidence _sfiducial interval, computing formula is as follows:

$[k_L,k_U]=[\left(\frac{1-\mathrm{\&alpha;}}{2}\right)n+1,n\left(\frac{1+\mathrm{\&alpha;}}{2}\right)]---\left(4\right)$

In formula: [k _l, k _u] be R _sthe lower limit of fiducial interval with the upper limit correspond respectively to value corresponding on sorting position, α is the degree of confidence of specifying.

8. a kind of manufacture process product reliability downslide risk evaluating method based on mass deviation according to claim 1, is characterized in that: " the manufacture process product reliability downslide risk R that computation obtains described in step 6 _sand the fluctuation situation of each batch of reliability downslide risk under fiducial interval monitoring product batch production state ", refer to according to R _sevaluation criterion weigh this batch of R _srisk level while, utilize the method for power gathering-point analysis monitoring batch production process can characterize the parameter ξ of the overlay information of each batch of fiducial interval and the fluctuation situation of positional information, whether offset to monitor manufacture process, parameter ξ computing formula is as follows:

In formula: fiducial interval is the interval calculated in formula (4), accepts the length that burst length is the fiducial interval obtained under stable state.