CN108646684A - A kind of multi-product production line production cycle prediction technique based on mobility measurement - Google Patents
A kind of multi-product production line production cycle prediction technique based on mobility measurement Download PDFInfo
- Publication number
- CN108646684A CN108646684A CN201810538128.1A CN201810538128A CN108646684A CN 108646684 A CN108646684 A CN 108646684A CN 201810538128 A CN201810538128 A CN 201810538128A CN 108646684 A CN108646684 A CN 108646684A
- Authority
- CN
- China
- Prior art keywords
- station
- logic
- product
- link
- indicate
- 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.)
- Granted
Links
Classifications
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
- G05B19/00—Programme-control systems
- G05B19/02—Programme-control systems electric
- G05B19/418—Total factory control, i.e. centrally controlling a plurality of machines, e.g. direct or distributed numerical control [DNC], flexible manufacturing systems [FMS], integrated manufacturing systems [IMS], computer integrated manufacturing [CIM]
- G05B19/41885—Total factory control, i.e. centrally controlling a plurality of machines, e.g. direct or distributed numerical control [DNC], flexible manufacturing systems [FMS], integrated manufacturing systems [IMS], computer integrated manufacturing [CIM] characterised by modeling, simulation of the manufacturing system
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
- G05B2219/00—Program-control systems
- G05B2219/30—Nc systems
- G05B2219/32—Operator till task planning
- G05B2219/32339—Object oriented modeling, design, analysis, implementation, simulation language
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
- Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
- Y02P90/02—Total factory control, e.g. smart factories, flexible manufacturing systems [FMS] or integrated manufacturing systems [IMS]
Abstract
The invention belongs to line management technical field, the multi-product production line production cycle prediction technique of the invention based on mobility measurement provide in the case of multi-product mixture manufacturing, have the characteristics that the mobility of " go here and there and go here and there " production line is measured and the production cycle prediction technique towards product accordingly.This method includes：Initial data, history data and the future period planning data of extraction and each station of process for producing line；The case where being actually passed through link according to product builds the logic station of product, and concatenating logic station obtains the logic production line of product；According to basic data and logic product line group at the mobility of each logic station and the mobility transmitted between logic station on calculating Product Logic production line；Based on the queuing time of each logic station of VUT equation calculation future periods, the accumulative queuing time for summing it up each logic station and effective process time obtain the whole production cycle of Product Logic production line.
Description
Technical field
The invention belongs to line management technical fields, and in particular to the production cycle under the conditions of multi-product mixture manufacturing is pre-
It surveys.
Background technology
Production cycle (Cycle Time, CT) is that the key performance of product (such as semiconductor chip packaging test) production line refers to
Mark one of, it refer to a collection of product be used as in product (WIP) in production line time-consuming mean value.Production cycle is to formulate
The key parameter of the production schedule, if the CT of energy Accurate Prediction production line, the production schedule formulated also will be more rationally raw
Production efficiency is just mutually due for promotion.It includes many processes that semiconductor chip packaging, which tests production line, and each process corresponds to corresponding work
Stand (Workstation), and product is used as and is handled followed by each station according to process route in product.Station is being processed
There are many Dynamic Uncertain factors including failure, mold changing etc. in the process, when these factors can make the processing of station
Between exist fluctuation, the process time including this Dynamic Uncertain factor for considering station is called effective process time
(Effective Process Time, EPT) calls the wave phenomenon of this effective process time the mobility of station, becomes
The presence of dynamic property causes the average processing time even if each station on production line equal and feed rate be equal to it is average plus
Between working hour, product still may undergo the stand-by period before station, and the addition of stand-by period is so that the prediction of production cycle becomes
It obtains more complicated.《Factory's physics》The VUT equations of proposition describe stand-by period and mobility (V), utilization rate (U) well
Quantitative relationship between effective process time (T), if can precisive go out the change transmitted between the mobility and station of station
Dynamic property can relatively accurately calculate the stand-by period before each product station, finally by the queuing for adding up each station
Time and process time can be obtained by the production cycle of whole production line.
Referring to Fig. 1, usually composed in parallel by a plurality of link inside the station of semiconductor chip packaging test production line, these
Link is composed in series by the equipment of several changeable parameters again, can be processed by change parameter same link different types of
Product.In practical operation, the same station in the same production cycle may mix again passes through multiple product.This multi-product is mixed
Symphysis is produced and the equipment compositing characteristic of " go here and there and-simultaneously-go here and there " makes effective process time of station forms to become sufficiently complex, and then is led
The measurement complexity for having caused station mobility, there is presently no corresponding scientific and reasonable measures, this becomes based on VUT
The biggest obstacle of this kind of production line production cycle of prediction equation.
Although application No. is 200810201503.X, apply for the patent of entitled " production cycle target measurement and system "
In application, disclose it is a kind of by being directly fitted the relational implementation production cycle of utilization rate and queuing time prediction (or survey
Amount) method, although this method can also reach the target of production cycle prediction, it only considered utilization rate parameter and is built
Mould, to which once change (product and link variations) this method has occurred by failure in mobility between the production cycle；In addition, the party
Case does not distinguish the period of each product in the case of multi-product mixed flow, to calculate the life of multiple product entirety
Produce the period.Station is properly termed as by black box due to it, cannot distinguish between each equipment in parallel in station (equipment corresponds to link)
And the specific influence by the different product of same station on the production cycle, it also can not just optimize for further performance and provide more
Fine-grained direction guidance.
Invention content
The goal of the invention of the present invention is：For the difficult point and deficiency of existing production cycle prediction technique, pass through mechanism
The station mobility in the case of multi-product mixing and more equipment parallel connections is measured in analysis realization, and then based on the realization pair of VUT equations
Have the characteristics that the production cycle prediction of this production line, while analysis foundation is provided for the optimization of such manufacture system.
The technical scheme is that：A kind of multi-product production line production cycle prediction technique based on mobility measurement,
It includes the following steps：
Step 1：The process data of each physics station in extraction and process for producing line, the process data includes original performance
Data, history shut down data, the scheduling of production data of history scheduling of production data and future period；
Step 2：Be actually passed through in each physics station by product link constitute product present physical station logic work
It stands, the logic station for identical product of connecting obtains the logic production line of product；
Step 3：Based on the process data of the corresponding physics station of each logic station, the logic production of each product is measured
Effective process time of each logic station, mobility on line：Every link for measuring each logic station first is corresponding effectively
Process time, mobility；Regression model is built again merges effective process time of all link of each logic station, variation
Property, obtain effective process time, the mobility of each logic station；
Step 4：Queuing time based on each logic station of VUT equation calculation future periods about different product, iteration
Queuing time and process time on accumulative logic production line obtain the whole production cycle of each product.
The multi-product production line production cycle prediction technique measured based on mobility of the present invention proposed is based on ratio
The method that the mode of weighting proposes multiple product processing mobility on convergence single link, this method can be in mixture manufacturing
Product calculates the link under any products combination based on the respective basic mobility of each product after changing and processes mobility,
To solve the mobility metric question in the case of multi-product mixed flow；By introducing regression model in parallel a plurality of in station
The rule that the processing mobility of link is converged to station overall processing mobility is portrayed, and is realized to link structures in parallel
Mobility is measured so that the measurement of mobility can adapt to the variation of link structures, to solve the mobility of link in parallel
Metric question；And then can be divided in conjunction with logic station and carry out parameter extraction, multi-product mixed flow situation may be implemented by VUT equations
Under prediction to the specific products production cycle.
Further, in step 3, when measuring every link corresponding effective processing of each product about each logic station
Between, mobility be specially：
Indicate that logic station identifier, j indicate link identifiers with i, t indicates product category identifier, based on link's
Shut down link couples of t kind products p of j-th strip of data type calculating logic station itEffective time teijtAnd when effectively processing
Between standard deviation sigmaeijt：
If the shutdown data of current link, which only include to take the lead, shuts down data, according to formulaCalculate effectively processing
Time teijt, further according toCalculate corresponding effective process time standard
Poor σeijt；
If the shutdown data of current link only include non-take the lead and shut down data, according to formulaCalculating has
Imitate process time teijt, further according toIt calculates corresponding
Effective process time standard deviation sigmaeijt；
If the shutdown data of current link include taking the lead to shut down data and non-shutdown data of taking the lead, according to formulaCalculate effective process time teijt, further according to
Calculate corresponding effective process time standard deviation sigmaeijt；
Wherein, t0ijtIndicate link couples of t kind products p of j-th strip of logic station itTheoretical process time, wherein ptBelong to
In product set linkijCorresponding category set, linkijIndicate the product set of the j-th strip link by logic station i；
C0ijtIndicate link couples of t kind products p of j-th strip of logic station itTheoretical process time the coefficient of variation；σ0ijt=C0ijt×
t0ijt；Parameter Aij=mfij/(mfij+mrij), wherein mfijWhen indicating to shut down caused by the failure of the j-th strip link of logic station i
Between be spaced mean value, mrijIndicate the time interval mean value of the fault restoration of the j-th strip link of logic station i；σrijIndicate logic work
Stand i j-th strip link fault restoration time interval standard deviation；Wherein tsijIndicate changing for the j-th strip link of logic station i
Mould time interval mean value, σsijIndicate the variance at the mould replacing time interval of the j-th strip link of logic station i, NsijIndicate logic work
Stand i j-th strip link twice change the mold between institute's converted products par；
Fusion multiple product obtain each product the j-th strip link of logic station i effective process timeAnd become
Dynamic propertyWhereinK is product identifiers, and M is indicated through excessive
The product category number of preceding link, r0tIndicate t kind products ptIngredient proportion, nitIndicate t kind products ptThe logic work of process
It stands the link quantity in parallel of i, miktIndicate t kind products ptWith k-th of product pkThe link's passed through simultaneously on logic station i
Number, wherein pk∈ P, P indicate the product set of mixture manufacturing.
Further, the regression model is specially：yγ=fγ(x)=w'x+e, wherein x=(x1x2x3x4x5x6x7) ' table
Show feature vector, x1Indicate effective process time of all link of current logic stationMean value, x2Indicate current logic
Effective process time of all link of stationVariance, x3Indicate the mobility of all link of current logic station
Root mean square, x4Indicate effective process time of all link of current logic stationMaximum value, x5Indicate current logic
Effective process time of all link of stationMinimum value, x6Indicate the mobility of all link of current logic stationMaximum value, x7Indicate the mobility of all link of current logic stationMinimum value；Symbol (×) ' indicate to turn
It sets, w indicates that coefficient vector, e indicate error vector, yγSubscript γ=1,2, wherein y1After indicating logic station overall fusion
Effective process time, i.e. y1For the effect process time of the arbitrary logical work station of producty2After indicating logic station overall fusion
Mobility, i.e. y2For the arbitrary logical work station of product
Further, step 4 is specially：
Step 401：Calculate the expected utilization rate of the following production cycleWherein uiIndicate logic station i's
Expected utilization rate, rtFor the rate to feed intake, rbFor effective processing speed of logic production line bottleneck station, reiFor logic station i
Effective processing speed, miFor the link numbers in parallel of logic station i；
Step 402：Queuing time based on each product of VUT equation calculations before each logic station, in addition effectively adding
The production cycle that can be obtained single logic station between working hour recycles each logic station and executes the step and accumulative production week
Phase can be obtained the prediction production cycle of logic production line entirety.
In conclusion by adopting the above-described technical solution, the beneficial effects of the invention are as follows：
(1) it is explicitly modeled for factors of variability such as multi-product, parallel connection link structures because of the present invention, so as to
In production product and after link structures change, to measure out the mobility after corresponding variation, to realize it is corresponding in the case of
Production cycle calculate.
(2) in the case of multi-product mixed flow may be implemented in method proposed by the invention, to production cycle of specific products into
Row calculates, and is more in line with the actual demand of semiconductor chip packaging test production line.
Description of the drawings
Fig. 1 is the internal structure chart for the single physical station that semiconductor chip packaging tests production line；
Fig. 2 is the module frame chart of the method for the present invention；
Fig. 3 is the specific implementation flow chart of the method for the present invention.
Specific implementation mode
To make the object, technical solutions and advantages of the present invention clearer, with reference to embodiment and attached drawing, to this hair
It is bright to be described in further detail.
The present invention is using the coefficient of variation (Coefficient of Variation, CV) of effective process time as variation
Property Measure Indexes, based on the index predict multi-product production line production cycle, realize the present invention production cycle prediction side
When method, it can lead to and build different processing modules to realize.Referring to Fig. 2, in order to realize the production cycle prediction technique of the present invention, structure
4 processing modules are built, including：Input data processing module, Product Logic production line structure module, mobility metric module and life
Period forecasting module is produced, referring to Fig. 3, the specific implementation procedure of each module is：
(1) input data processing module：The process data of extraction and each physics station of process for producing line, and will extraction and place
Reason result passes to Product Logic production line structure module.
Wherein, process data includes original performance data (abbreviation initial data), history shutdown data, history scheduling of production
The scheduling of production data of data and future period.
Initial data is the theoretical parameter that each link of physics station processes certain product, including：Theoretical process time, theory
The coefficient of variation of process time；History shuts down the outage record data that data are each link, is divided into take the lead shutdown and non-take the lead and stops
Machine data.It takes the lead and shuts down data such as：Disorderly closedown and repair for event sequence；Non- take the lead shuts down data such as：Change the mold shutdown intervals sequence
The product quantity sequence data processed between row, mold changing；Scheduling of production data refer to the feelings that each product passes through each physics station link
Condition, i.e., the link item numbers that each product occupies, is divided into history scheduling of production data and the scheduling of production data with future period.
(2) Product Logic production line builds module：Link situations are actually passed through according to the product being collected into, build product
Logic station, concatenating logic station obtain the logic production line of product and pass to mobility metric module and production cycle prediction
Module.
For specific product pk, corresponding logic station is expressed as the product by being somebody's turn to do in i-th of physics station
The link of physics station gathers, and is specifically represented by formula (1),
wik={ lis|s∈Sik} (1)
Wherein, lisRepresent the s articles link, S of i-th of physics stationikRepresent pkBy the number of the link of physics station i
Set.
Based on this, product pkLogic production line availability vector be expressed as formula (2),
plk=(w1k,w2k,...,wnk) n∈W (2)
For i-th logic station (for specific product, only there are one logic station in physics station, therefore work
Station identifier can be general) jth link, the product set being processed by it can be expressed as formula (3),
linkij={ pk} pk∈P,i∈W,j∈Li (3)
Wherein, pkIndicate product, under be designated as product identifiers, P represents the product set of mixture manufacturing, and W represents physics life
The physics station number set of producing line, LiRepresent the link number set of i-th of physics station.
(3) mobility metric module：Measure the mobility of each logic station and effectively processing on Product Logic production line
Time and the mobility transmitted between logic station, and measurement results are passed into production cycle prediction module.
For the specific link and specific products of some logic station, theoretical mobility is the variation of theoretical process time
Coefficient, measure formulas are (4)：
In formula (4), t0For theoretical process time, σ0For theoretical process time standard deviation, and C0、t0Object can be directly based upon
Each link processes the theoretical parameter acquisition of certain product in the initial data at science and engineering station.
Step 301：Coefficient of variation C based on theoretical process time0, theoretical process time t0With theoretical process time standard
Poor σ0Relationship measure correspondence effective process time, the mobility of every link in conjunction with the shutdown data type of each link.
Because of the shutdown data of link, it is divided into and takes the lead shutdown and non-take the lead shuts down data, the present invention is according to included by it accordingly
Shutdown data type the specific link of some logic station is obtained about particular kind of production using corresponding metric form
Effective process time of product and mobility.
(1) for mobility of taking the lead, using shutdown mobility is example caused by typical failure, the stream of measurements of mobility
Cheng Wei：Step1：Basic parameter shown in table 1 is obtained according to disorderly closedown and repair for event sequence statistic；
1 disorderly closedown mobility of table measures basic parameter
Step2：Availability parameter A is calculated according to formula (5)；
Step3：The effective process time for considering disorderly closedown factor is calculated according to formula (6)；
Step4：According to formula (7) corresponding mobility of taking the lead is calculated in conjunction with the theoretical mobility measured out
Cp；
(2) for non-mobility of taking the lead, typically to change the mold caused shutdown as example, mobility measurement flow is：
Step1：Basic parameter shown in table 2 is obtained according to mold changing sequence of events statistics；
The mold changing of table 2 shuts down mobility and measures basic parameter
Step2：The effective process time for considering mold changing shutdown factor is calculated according to formula (8)；
Step3：It is calculated corresponding non-in conjunction with the theoretical mobility measured out and process time according to formula (9)
Mobility of taking the lead Cnp：
(3) for taking the lead and the case where non-mobility of taking the lead all exists, then when needing first to calculate the effective processing taken the lead
Between teWith mobility Cp(being calculated respectively according to formula 6,7), then by teAnd CpIt is considered as theoretical process time and theoretical mobility
It brings formula (8) (9) into and can be obtained by have merged and take the lead with effective process time of non-shutdown of taking the lead and mobility to get to melting
Close the metric taken the lead with the mobility of non-shutdown of taking the leadWhereinAnd byObtain effective process time that fusion is taken the lead with non-shutdown of taking the lead.
Step 302：The flow of previous step is acted on to all products on all link of any logic station i, then may be used
To obtain effective process time ts of the every link about each producteijtWith mobility CeijtIt is (corresponding above-mentioned to take the lead, non-account for
First, take the lead the non-mobility C to take the lead under fusion Three modelsp、Cnp、Cfuse).This step will be according to each product on link
Shared ratio converges mixing by the mobility of the multi-product of single link and when effectively processing in a manner of average weighted
Between, obtain effective process times and mobility of the specific link about specific products logic station.
For with certain link described in formula (3), multi-product mobility and the convergence formula of effective process time are such as
Shown in formula (10)：
Wherein,Link couples of t kind products p of j-th strip of logic station i is indicated respectivelytTheoretical processing when
Between, mobility, wherein subscript i is physics station (logic station) identifier, k is product identifiers, and subscript j is link marks
Symbol, linkijRepresent all product set by i-th of logic station j-th strip link, ptFor the t of link mixture manufacturings
Kind product, M indicate product kind number, teijtIndicate effective processing of the j-th strip link of i-th of logic station about t kind products
Time, σeijtIndicate effective process time standard deviations of the j-th strip link about t kind products of i-th of logic station, and σeijt
=teijt×Ceijt, CeijtIndicate mobilities of the j-th strip link of i-th of logic station about t kind products, r0tIt is ptProduct
Original ingredient proportion, nitBe t products pass through logic station i link quantity in parallel, referred to as product about logic station and
Row degree has formula (11) based on formula (2).
nit=Count { wit} (11)
Wherein, Count operators represent the quantity of set of computations kind element.Meanwhile miktRepresent ptProduct passes through pkProduct pair
The quantity of link in the logic station answered, similarly there is formula (12)：
mikt=Count { wik∩wit} (12)
Step 303：It is fitted different in logic station for specific logic station by establishing linear regression model (LRM)
The rule that the mobility of link is converged to logic station entirety mobility；And it is whole based on the model metrics logic station trained
Mobility.Table 3 is the input and output of the regression model and corresponding symbol and calculation.
3 regression model of table outputs and inputs parameter
Shown in corresponding regression model such as formula (13).
yγ=fγ(x)=w'x+e, γ=1,2 (13)
Wherein, yiSubscript i=1,2, x=(x1x2x3x4x5x6x7) ' indicate feature vector, symbol (×) ' expression transposition, w
It is coefficient vector, e indicates error vector.Above-mentioned model can train (training objective using least square method：Minimize error
It is vectorial e), shown in the algorithm such as formula (14) of least square.
W*=arg min (y-Xw) ' (y-Xw) (14)
Wherein, X indicates the feature vector of sample, y ∈ { y1,y2}。
Step 304：Mobility can also flow other than it can be had an impact in station between station, a upper logic station
The mobility of effective process time will be transmitted to downstream logic station with time departure separation fluctuation, be directly becoming downstream and patrol
The arrival time separation fluctuation of station is collected, therefore the present invention measures the rule of mobility transmission using formula (15)：
Wherein, Ca() indicates the arrival mobility of station in bracket, i.e. Ce(i) indicate that the arrival of logic station i changes
Property, i.e., according to the mobility after the obtained logic station overall fusion of formula (13)；Ca(i+1) it indicates under logic station i
Swim the arrival mobility of station, uiIndicate the expected utilization rate of station i, miIndicate the equipment in parallel (or parallel connection link) of station i
Quantity.
(4) production cycle prediction module：Based on the queuing time of each logic station of VUT equation calculation future periods, repeatedly
Queuing time and effective process time on generation totalling logic production line obtain the whole production cycle of product.
Step 401：The expected utilization rate of the following production cycle estimates that the utilization rate of production line can use formula (16) to count
It calculates.
Wherein, uiIndicate the expected utilization rate of logic station i, rtFor the rate to feed intake, rbFor logic production line bottleneck station
Effective processing speed, reiFor effective processing speed of logic station i, miFor the quantity of logic station i parallel connections link.
Step 402：Queuing time based on each product of VUT equation calculations before each logic station, in addition effectively adding
The production cycle that can be obtained single logic station between working hour recycles each logic station and executes the step and accumulative production week
Phase can be obtained the prediction production cycle of logic production line entirety, and specific execution flow is as follows：
Step1：The identifier i=1 to logic station is initialized, is initialised to up to time interval mobility Ca, i.e., initially
Change Ca(1) value, generally rule of thumb initializes Ca, C can be usually seta(1)=0；
Step2：It is calculated up to time interval mobility C based on abovementioned stepsa(then directly it is pre- for first station
If value, non-first station then calculated according to formula (15)), mobility Ce, effective process time teAnd expected utilization rate u tri-
A basic parameter.
Step3：Using the queuing time before VUT equation calculation logic station, shown in VUT equations such as formula (17).
Then for any logic station i, any product pkK, product pkTo the queuing time of logic station iFor：WhereinIndicate logic station i about product pkEffective process time,Indicate logic station i about product pkMobility, uiIndicate the expected utilization rate of logic station i, miIt indicates to pass through logic
The quantity of the link in parallel of station i.
Step4：The production cycle of single logic station is calculated according to formula (18)：
Step5：To logic station i+1, cycle executes Step2~4 in step 402, until the last one logic station,
To obtain production cycle of each logic station about current production.
Step6：Sum it up each production cycle of logic station about productWhole logic production line is obtained to close
In the production cycle of product.
The above description is merely a specific embodiment, any feature disclosed in this specification, except non-specifically
Narration, can be replaced by other alternative features that are equivalent or have similar purpose；Disclosed all features or all sides
Method or in the process the step of, other than mutually exclusive feature and/or step, can be combined in any way.
Claims (5)
1. a kind of multi-product production line production cycle prediction technique based on mobility measurement, which is characterized in that including following step
Suddenly：
Step 1：The process data of each physics station in extraction and process for producing line, the process data includes original performance number
Data, the scheduling of production data of history scheduling of production real data and future period are shut down according to, history；
Step 2：Link is actually passed through into product in the logic station of present physical station, series connection in each physics station by product
The logic station of identical product obtains the logic production line of product；
Step 3：Based on the process data of the corresponding physics station of each logic station, measure on the logic production line of each product
Effective process time of each logic station, mobility：The corresponding effective processing of every link of each logic station is measured first
Time, mobility；Regression model is built again and merges effective process time of all link of each logic station, mobility, is obtained
Effective process time, mobility to each logic station；
Step 4：Queuing time based on each logic station of VUT equation calculation future periods about different product, iteration are accumulative
Queuing time and process time on logic production line obtain the whole production cycle of each product.
2. the method as described in claim 1, which is characterized in that in step 3, measure each product about the every of each logic station
Link corresponding effective process times, mobility are specially：
Indicate that logic station identifier, j indicate that link identifiers, t indicate product category identifier with i；
Link couples of t kind products p of j-th strip of shutdown data type calculating logic station i based on linktEffective time teijt、
And effective process time standard deviation sigmaeijt：
If the shutdown data of current link, which only include to take the lead, shuts down data, according to formulaCalculate effective process time
teijt, further according toCalculate corresponding effective process time standard deviation
σeijt；
If the shutdown data of current link only include non-take the lead and shut down data, according to formulaIt calculates and effectively adds
T between working houreijt, further according toIt calculates corresponding effective
Process time standard deviation sigmaeijt；
If the shutdown data of current link include taking the lead to shut down data and non-shutdown data of taking the lead, according to formula
Calculate effective process time teijt, further according toMeter
Calculate corresponding effective process time standard deviation sigmaeijt；
Wherein, t0ijtIndicate link couples of t kind products p of j-th strip of logic station itTheoretical process time, wherein ptBelong to production
Product set linkijCorresponding category set, linkijIndicate the product set of the j-th strip link by logic station i；C0ijtTable
Show link couples of t kind products p of j-th strip of logic station itTheoretical process time the coefficient of variation；σ0ijt=C0ijt×t0ijt；
Parameter Aij=mfij/(mfij+mrij), wherein mfijIndicate downtime interval caused by the failure of the j-th strip link of logic station i
Mean value, mrijIndicate the time interval mean value of the fault restoration of the j-th strip link of logic station i；σrijIndicate the of logic station i
The time interval standard deviation of the fault restoration of j link；Wherein tsijBetween the mould replacing time for indicating the j-th strip link of logic station i
Every mean value, σsijIndicate the variance at the mould replacing time interval of the j-th strip link of logic station i, NsijIndicate the jth of logic station i
The par of institute's converted products between the changing the mold twice of link；
Fusion multiple product obtain each product the j-th strip link of logic station i effective process timeAnd mobilityWhereinK is product identifiers, and M is indicated by current
The product category number of linkl, r0tIndicate t kind products ptIngredient proportion, nitIndicate t kind products ptThe logic station of process
The link quantity in parallel of i, miktIndicate t kind products ptWith k-th of product pkThe link's passed through simultaneously on logic station i
Number, wherein pk∈ P, P indicate the product set of mixture manufacturing.
3. method as claimed in claim 2, which is characterized in that regression model is specially：yγ=fγ(x)=w'x+e；
Wherein x=(x1x2x3x4x5x6x7) ' indicate feature vector, x1Indicate effective processing of all link of current logic station
TimeMean value, x2 indicate current logic station all link effective process timeVariance, x3Expression is currently patrolled
Collect the mobility of all link of stationRoot mean square, x4When indicating effective processing of all link of current logic station
BetweenMaximum value, x5Indicate effective process time of all link of current logic stationMinimum value, x6Indicate current
The mobility of all link of logic stationMaximum value, x7Indicate the mobility of all link of current logic station
Minimum value；Symbol () ' indicate that transposition, w indicate that coefficient vector, e indicate error vector, yγSubscript γ=1,2, wherein y1
Indicate effective process time after logic station overall fusion, i.e. y1For the effect process time of the arbitrary logical work station of product
y2Indicate the mobility after logic station overall fusion, i.e. y2For the arbitrary logical work station of product
4. method as claimed in claim 3, which is characterized in that step 4 is specially：
Step 401：Calculate the expected utilization rate of the following production cycleWherein uiIndicate the expection of logic station i
Utilization rate, rtFor the rate to feed intake, rbFor effective processing speed of logic production line bottleneck station, reiFor having for logic station i
Imitate processing speed, miFor the link numbers in parallel of logic station i；
Step 402：Queuing time based on each product of VUT equation calculations before each logic station, in addition when effectively processing
Between can be obtained production cycle of single logic station, each logic station is recycled and executes the step and the accumulative production cycle is
The prediction production cycle of logic production line entirety can be obtained.
5. method as claimed in claim 4, which is characterized in that step 402 is specially：
402-1：The identifier i=1 to logic station is initialized, is initialised to up to time interval mobility, it will be between arrival time
It is denoted as C every mobility initial valuea(1)；
402-2：Utilize VUT equation calculation current productions pkTo the queuing time of logic station iWhereinWhereinIndicate logic station i about product pkEffective process time,Indicate logic station i about product pkMobility, uiIndicate the expected utilization rate of logic station i, miIt indicates to pass through logic
The quantity of the link in parallel of station i；
402-3：Calculating logic station i is about product pkProduction cycleWherein
402-4：More new identifier i=i+1, and according to formulaIt calculates
The time departure separation fluctuation C of current logic stationa(i+1)；
402-5：Cycle executes step 402-2~402-4, until the last one logic station, to obtain each logic station
Production cycle；
402-6：Sum it up the production cycle of each logic stationWhole logic production line is obtained about product pkProduction
Period.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810538128.1A CN108646684B (en) | 2018-05-30 | 2018-05-30 | Multi-product production line production period prediction method based on variability measurement |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810538128.1A CN108646684B (en) | 2018-05-30 | 2018-05-30 | Multi-product production line production period prediction method based on variability measurement |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN108646684A true CN108646684A (en) | 2018-10-12 |
| CN108646684B CN108646684B (en) | 2020-10-23 |
Family
ID=63758663
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810538128.1A Active CN108646684B (en) | 2018-05-30 | 2018-05-30 | Multi-product production line production period prediction method based on variability measurement |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108646684B (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109740813A (en) * | 2018-12-29 | 2019-05-10 | 上海华力微电子有限公司 | Online product batches run the analyzing and predicting method of goods state in wafer manufacture |
| CN111857081A (en) * | 2020-08-10 | 2020-10-30 | 电子科技大学 | Chip packaging test production line performance control method based on Q-learning reinforcement learning |
| CN113780907A (en) * | 2020-06-10 | 2021-12-10 | 宝山钢铁股份有限公司 | Consumable demand and inventory control method, device, system and storage medium |
| CN113807513A (en) * | 2020-06-12 | 2021-12-17 | 中芯国际集成电路制造(上海)有限公司 | Method, system and computing device for determining impact factor weight of semiconductor production process |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20070239300A1 (en) * | 2006-03-30 | 2007-10-11 | Burda Richard G | Method and system for setting rates and targets in a range management system |
| CN101727098A (en) * | 2008-10-21 | 2010-06-09 | 中芯国际集成电路制造(上海)有限公司 | Cycle time target measuring method and system |
| CN105045236A (en) * | 2015-07-21 | 2015-11-11 | 江苏云道信息技术有限公司 | Assembly line production scheduling method and system |
| CN107944695A (en) * | 2017-11-21 | 2018-04-20 | 电子科技大学 | One kind is based on the probabilistic chaining key setting buffers method of production |
-
2018
- 2018-05-30 CN CN201810538128.1A patent/CN108646684B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20070239300A1 (en) * | 2006-03-30 | 2007-10-11 | Burda Richard G | Method and system for setting rates and targets in a range management system |
| CN101727098A (en) * | 2008-10-21 | 2010-06-09 | 中芯国际集成电路制造(上海)有限公司 | Cycle time target measuring method and system |
| CN105045236A (en) * | 2015-07-21 | 2015-11-11 | 江苏云道信息技术有限公司 | Assembly line production scheduling method and system |
| CN107944695A (en) * | 2017-11-21 | 2018-04-20 | 电子科技大学 | One kind is based on the probabilistic chaining key setting buffers method of production |
Non-Patent Citations (1)
| Title |
|---|
| 胡强: "芯片封装测试生产线生产周期预测与优化研究", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109740813A (en) * | 2018-12-29 | 2019-05-10 | 上海华力微电子有限公司 | Online product batches run the analyzing and predicting method of goods state in wafer manufacture |
| CN109740813B (en) * | 2018-12-29 | 2020-11-24 | 上海华力微电子有限公司 | Analysis and prediction method for on-line product batch running state in wafer manufacturing |
| CN113780907A (en) * | 2020-06-10 | 2021-12-10 | 宝山钢铁股份有限公司 | Consumable demand and inventory control method, device, system and storage medium |
| CN113780907B (en) * | 2020-06-10 | 2023-09-12 | 宝山钢铁股份有限公司 | Consumable demand and inventory control method, device, system and storage medium |
| CN113807513A (en) * | 2020-06-12 | 2021-12-17 | 中芯国际集成电路制造(上海)有限公司 | Method, system and computing device for determining impact factor weight of semiconductor production process |
| CN111857081A (en) * | 2020-08-10 | 2020-10-30 | 电子科技大学 | Chip packaging test production line performance control method based on Q-learning reinforcement learning |
Also Published As
| Publication number | Publication date |
|---|---|
| CN108646684B (en) | 2020-10-23 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108646684A (en) | A kind of multi-product production line production cycle prediction technique based on mobility measurement | |
| CN104600699B (en) | A kind of distribution net work structure method of estimation based on MINLP model model | |
| CN106503919A (en) | A kind of power distribution network evaluation methodology based on power supply zone characteristic | |
| WO2022021726A1 (en) | Pmu-based power system state estimation performance evaluation method | |
| CN103676881A (en) | Dynamic bottleneck analytical method of semiconductor production line | |
| CN107133721A (en) | Power distribution network based on gray theory measures bad data relation factor analysis method | |
| CN111488896A (en) | Distribution line time-varying fault probability calculation method based on multi-source data mining | |
| CN103324858A (en) | Three-phase load flow state estimation method of power distribution network | |
| CN110378549A (en) | A kind of transmission tower bird pest grade appraisal procedure based on FAHP- entropy assessment | |
| CN111177128B (en) | Metering big data batch processing method and system based on improved outlier detection algorithm | |
| CN109494726A (en) | Stability of power system online evaluation method based on DLRNN neural network | |
| CN106021724A (en) | Energy efficiency evaluation method of machine tool product manufacturing system based on AHM and entropy method | |
| CN109375050A (en) | A kind of electric system key sections discrimination method based on primary fault route cut set | |
| CN106372440B (en) | A kind of adaptive robust state estimation method of the power distribution network of parallel computation and device | |
| CN106503861A (en) | Wind power forecasting method based on many meteorological sources wind speed fusions of probability statistics and particle group optimizing | |
| CN107944716A (en) | Based on the modified substation's electrical energy measurement cycle balance abnormality diagnostic method of learning outcome | |
| CN112508254B (en) | Method for determining investment prediction data of transformer substation engineering project | |
| Ni et al. | A review of line loss analysis of the low-voltage distribution system | |
| WO2024087129A1 (en) | Generative adversarial multi-head attention neural network self-learning method for aero-engine data reconstruction | |
| CN106159941A (en) | A kind of power system state estimation method considering actual measurement error propagation characteristic | |
| CN116467658A (en) | Equipment fault tracing method based on Markov chain | |
| CN105699043B (en) | A kind of wind tunnel sensors that improve measure stability and the method for precision | |
| CN115564310A (en) | Reliability evaluation method for new energy power system based on convolutional neural network | |
| CN115879605A (en) | Aircraft final assembly system transient production capacity calculation method based on pulsating production mode | |
| CN114861993A (en) | Regional photovoltaic power generation prediction method based on federal learning and deep neural network |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |