CN107358348A - A kind of method that plan fault-tolerance is improved based on OEE - Google Patents

A kind of method that plan fault-tolerance is improved based on OEE Download PDF

Info

Publication number
CN107358348A
CN107358348A CN201710544025.1A CN201710544025A CN107358348A CN 107358348 A CN107358348 A CN 107358348A CN 201710544025 A CN201710544025 A CN 201710544025A CN 107358348 A CN107358348 A CN 107358348A
Authority
CN
China
Prior art keywords
mrow
case
msub
time
oee
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
Application number
CN201710544025.1A
Other languages
Chinese (zh)
Inventor
李迅波
王振林
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Chengdu Dianke Zhaopin Technology Co Ltd
Original Assignee
Chengdu Dianke Zhaopin Technology Co Ltd
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Chengdu Dianke Zhaopin Technology Co Ltd filed Critical Chengdu Dianke Zhaopin Technology Co Ltd
Priority to CN201710544025.1A priority Critical patent/CN107358348A/en
Publication of CN107358348A publication Critical patent/CN107358348A/en
Pending legal-status Critical Current

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
    • G06Q10/0637—Strategic management or analysis, e.g. setting a goal or target of an organisation; Planning actions based on goals; Analysis or evaluation of effectiveness of goals
    • G06Q10/06375—Prediction of business process outcome or impact based on a proposed change

Abstract

The invention belongs to MES system field, discloses a kind of method based on OEE raising plan fault-tolerances, including：Start process is monitored by underlying device, obtains the time consumed at present in set-up procedure of starting shooting, using nearest neighbor method, start is provided and prepares remaining time；Breakdown maintenance process is monitored by underlying device, the time consumed during obtaining at present, after matching, the residue for predicting whole maintenance process takes Trepair；Debugging process is monitored by underlying device, obtains the value of parameter in current debugging process, after matching, the residue for predicting whole debugging process takes TdebuggEtc..The present invention is solved in production process, can not effectively it be connected because dragging the phase to wait between process, planned dispatching can not be tackled in time, can not effectively implement so as to the production schedule, in the production process of reality by artificial experience remove plan for adjustment and the problems such as without standard reference.

Description

A kind of method that plan fault-tolerance is improved based on OEE
Technical field
The invention belongs to MES system field, more particularly to a kind of method that plan fault-tolerance is improved based on OEE.
Background technology
MES system is formed a connecting link in whole enterprise integration information system, is production activity and management activity information communication Bridge.Do not implement MES, control integration is an empty verbiage.For Facing to Manufacturing processing industry and the ERP of process industrial, take off It will be unable to go tissue, management and optimization production according to the market demand from MES.MES system is the production management skill towards shop layer Art and real time information system, it is to implement Enterprise Agile manufacturing strategy, realizes the basic fundamental means of Workshop Production Agility. Because MES system is emphasized to control and coordinates, modern manufacturing industry information system is set not only have good planning system, and can make Plan the execution system put into effect.Therefore rapid promote is come in the enterprises of MES abroad between a few years, and to enterprise Industry brings huge economic benefit.Chinese Enterprises, which implement business management software based on MRPIIERP, to be had nearly 20 years History, in the industry cycle, there is for the huge ERP manufacturers of number and exploitation legion, but this part manufacturer only has fewer parts can be to MES direction is developed, and its reason is：MES system is combined closely with Industry Control, and it is researched and developed and implements to need very strong industry Automation foundation and industry spot engineering experience, this is technical threshold higher together, by the IT vendor of many ERP types Keep outside of the door.But with the development of industry and economic scale, have substantial portion of IT vendor and automation manufacturer, manufacture type Combination of enterprise works closely, and in recent years, MES manufacturers emerge in large numbers like the mushrooms after rain, its technology and implement means also increasingly into It is ripe.Cores of the production planning system MPS as MES, plays conclusive effect in whole MES system. In practical application, for various reasons, the actual conditions especially in production process can not be fed back in time, and MPS's is accurate Property and agility can not often meet actual demand, and the plan of causing can not be implemented well, can be brought on the contrary to enterprise negative The difficult situation of load.
For abundant excavating equipment and the potential value of operating personnel, enterprise's volume competitiveness is improved, international manufacturing industry proposes Overall equipment efficiency OEE concept.In recent years, some manufacturing industry were constantly carrying out this method.It is actual to prove that OEE is One fabulous fiducial tool, analyzed by each subitems of OEE models, it accurately can clearly tell your device efficiency such as What, which link of production is how many loss, and you can carry out those improvement work.Long-term uses OEE instruments, enterprise Industry can easily find the bottleneck for influenceing production efficiency, and be improved and track and find out weak link, make all departments Responsibility more refine and clearly.Although this methods of OEE are all being carried out by numerous manufacturers, worldwide research shows, The average OEE of currently manufactured industry is only 60%.The OEE of global industry averagely serious hope values should be that 85% (world-class OEE refers to Number) or it is higher.It is clear that current OEE indexes also have many leeway that can be improved, particularly in our countries.
In actual production scheduling, it is necessary in face of uncertain factor, thus process time be mostly can not be completely true Fixed, it can only be estimated according to conventional experience, so the precision of plan just determines whether whole MES systems can be effective Ground performs.Conventional Forecasting Methodology having time serial method, multiple linear regression method, gray system theory and ANN Network, in addition, the algorithm of the case-based reasioning also in intelligent algorithm.
OEE methods are widely used in major workshop, and the statistics that its data mainly comes from production line employee is converged Always, then administrative staff are transferred to go comprehensive analysis, mainly including Homes Using TV (OR), speed load factor (PR) and the percentage of A-class goods (QR).
The technology of plan fault-tolerance is improved now, on the one hand, mainly by artificial experience, according to the feelings actually occurred Condition, which is done, artificially to be judged, its problem is that the emergency case tackled in production process is not prompt enough, may also manually be judged by accident Or the situation of carelessness.On the other hand, also have to using reasoning by cases prediction process time, its problem is that whole workshop is given birth to The actual conditions being likely to occur during production can not be tackled comprehensively.
In summary, problems of the prior art are：The production schedule can not tackle production in time in existing MES During actual conditions and adjust, cause fault-tolerance insufficient；
And in the existing MES production schedules, difficulty is effectively predict that future is possible to according to present case The situation of generation, so as to accomplish to do some preparations in advance；Cause the reaction sensitivity of system low, the accuracy of plan is poor.
The content of the invention
The problem of existing for prior art, the invention provides it is a kind of based on OEE improve plan fault-tolerance method,
The present invention is achieved in that a kind of method based on OEE raising plan fault-tolerances, including case description, case Retrieval, modification reuse and preserved renewal；Specially：Using the time prediction method of case-based reasioning, treated in the past is asked Topic is described as the case being made up of problem characteristic collection and solution, and organizes in a certain way, and case library is arrived in storage In；When there is new problem solution, case library is retrieved, case matching is carried out, finds out and the same or analogous several cases of new problem Example；By directly quoting identical case or modification similar cases, the solution of new problem is obtained, while preserved newly in case library Experience, to realize dynamic learning.
Further, the case description specifically includes：
The method that case description in database uses facing relation data model, for case i, in relation data In storehouse, input variable combination is expressed as problem characteristic vector xi=(xi1,xi2,...,xim), wherein i is represented the in database I bars record, that is, i-th of case among case library, and 1,2 ... m, represent the time point sampled in case；
The Case Retrieval specifically includes：Using nearest neighbor method, certain kind of inspection target case and case in case library Property matching degree, and by calculate each attributes match Chengdu weighted sum determine best match case；
The case modification reuses and case is modified using weighted mean method, if retrieved from case library and Present case case similarity the most similar is d (xi,x0)<5%, then each component in these cases is averaging Value, finally obtains an average vector, xi=(xi1,xi2,...,xim), add coefficient d 1, d2 obtains xmax=d1xi, xmin= d2xiIf the component in present case exceedes the component value in minimum vector sum maximum vector, using minimum component and maximum Component replaces；
The case is preserved and specifically included：To prevent that case library is excessive, the case for meeting equation below is only preserved：
min d(xi,x0) ＜ δ.
Further, described matching process uses nearest neighbor method, checks certain kind of target case and case in case library Property matching degree, and by calculate each attributes match Chengdu weighted sum determine best match case；The formula of nearest neighbor method For：
Wherein, x0And xiThe problem of being i-th case in present case and case library respectively characteristic vector, s (xij,x0j) It is x0And xiFeature j similarity, wjIt is feature j attribute weight；s(xi,x0) it is that present case is similar to case id Degree, its value is smaller, and matching degree is better；
Similarity Measure in nearest method, using Euclidean distance mode, Euclidean distance formula is：
d(xi, x.) present case and case i weighted euclidean distance are represented, its value is smaller, illustrates that two cases get over phase Seemingly；Wherein xijRepresent i-th case, the time-consuming situation of jth step, x0jIn the case monitored in real time in expression production process J steps take, wjRepresent the weight of correlation step；Case Retrieval uses d (xi, x.) minimum case carries out time prediction, The residue for predicting whole start set-up procedure takes TbootPrepare。
Further, the specific steps of the method based on OEE raising plan fault-tolerances include：
Start process is monitored by underlying device, the time consumed at present in set-up procedure of starting shooting is obtained, using nearest Adjacent method, provide start and prepare remaining time；
Breakdown maintenance process is monitored by underlying device, the time consumed during obtaining at present, after matching, predicted The residue of whole maintenance process takes Trepair；
Debugging process is monitored by underlying device, the value of parameter in current debugging process is obtained, after matching, predicts whole The residue of individual debugging process takes Tdebugg；
By underlying device monitor production process, the process of production is divided startup stage, plateau, slow down and stop rank Section；
By underlying device monitor production process, according to caused percent defective, using stochastic pattern time series forecasting, If the waste product number collected at regular intervals is y1, y2, y3 ..., yn, first-order difference, Ran Houji are taken to sample first CalculateSampled correlation function, parameter is estimated using autoregressive model；Predict follow-up number of rejects Nwaste.
Further, before monitoring start process by underlying device, need to carry out：
Establish the case library of start time：According to process, board, personnel, preparation process, a case library is obtained；
Establish breakdown maintenance case library：According to board, fault type, maintenance personal, maintenance step, a case is obtained Storehouse；
Establish debug time case library：According to process, board, personnel, specifications and models, the different debugging ginsengs sampled every time Number, obtains a case library；
Establish velocity standard curve：According to board, operating personnel, velocity standard curve is formulated；
Percent defective is established with reference to table.
Further, the startup stage, including：According to the waveform of real-time data of collection, startup stage secondary song will be used The equation of line goes to be fitted, Remaining Stages normative reference rate curve, obtains the remaining predicted time T of production processproduce；
Y=ax2+bx+c；
The plateau includes：According to the waveform of real-time data of collection, plateau is used into the method for moving average, remained Remaining stage reference standard speed curve, obtain the remaining predicted time T of debugging processproduce；
The deceleration stop phase includes：According to the waveform of real-time data of collection, deceleration stop phase is used into secondary song The equation model of line, obtain the remaining predicted time T of production processproduce；
Y=ax2+bx+c；
Wherein, the coefficient that a, b, c determine according to three points collected in production process.
Further, it is described by underlying device monitor production process, according to caused percent defective, using the stochastic pattern time Serial anticipation method, if the waste product number collected at regular intervals is y1, y2, y3 ..., yn, in, specifically include：
1) first-order difference is taken to sample：
Wherein, parameter ykIn k be k-th of sampled point；
2) calculateSampled correlation function, the sequence of calculation { Yt, t ∈ T } sample auto-covarianceWith sample auto-correlation Function：
3) autoregressive model is used, p rank autoregression model AR (p) form is：
The parameter Estimation of model uses Matrix Estimation method：
Wherein, the parameter Estimation matrix of AR (1) model is：
The parameter Estimation matrix of AR (2) model is：
4) follow-up number of rejects Nwaste is predicted more than.
Further, the Forecasting Methodology of the remaining time, including：
A) preparatory stage before starting shooting, the total time that prediction completion task also needs to：
Ttotal=TbootPrepare+Tdebugg+Tproduce,
Wherein, TdebuggWith reference to the average time in debugging case library, TproduceWith reference to according to mark rate curve production institute Time；
B) debug the stage, the total time that prediction completion task also needs to：
Ttotal=Tdebugg+Tproduce,
Wherein, TproduceWith reference to according to the mark rate curve production time used；
C) breakdown maintenance stage in production process, the total time that prediction completion task also needs to：
Ttotal=Trepair+Tproduce
Wherein, TproduceWith reference to according to the mark rate curve production time used；
D) in production process, the total time that completion task also needs to is predicted：
Ttotal=Tproduce.
Further, the method for waste product monitoring includes：
By RwasteWith percent defective with reference to the maximum percent defective R in tablemaxIt is compared, if Rwaste-Rmax>=0, pass through machine Device equipment produces alarm signal, waits artificial determination, is then shut down in time if signal is accurate.
Another object of the present invention is to provide a kind of system based on OEE raising plan fault-tolerances.
Advantages of the present invention and good effect are：The method that plan fault-tolerance is improved based on OEE of the present invention, is solved In production process, can not effectively it be connected because dragging the phase to wait between process, planned dispatching can not be tackled in time, so as to raw Production plan can not effectively be implemented, and remove plan for adjustment by artificial experience in the production process of reality and asked without standard reference etc. Topic.
The present invention compared with prior art, see the table below：Predicted time error information can be obtained：
It can be found that the present invention is in application case reasoning from table, the prediction error to the time significantly reduces.
Brief description of the drawings
Fig. 1 is the flow chart of the method provided in an embodiment of the present invention based on OEE raising plan fault-tolerances.
Embodiment
In order to make the purpose , technical scheme and advantage of the present invention be clearer, with reference to embodiments, to this hair It is bright to be further elaborated.It should be appreciated that the specific embodiments described herein are merely illustrative of the present invention, not For limiting the present invention.
Problems of the prior art are：The production schedule can not be tackled actual in production process in time in existing MES Situation and adjust, fault-tolerance deficiency present situation.
The application principle of the present invention is described in detail below in conjunction with the accompanying drawings.
It is provided in an embodiment of the present invention based on OEE improve plan fault-tolerance method, including case description, Case Retrieval, Modification reuses and preserved renewal；Specially：Using the time prediction method of case-based reasioning, former the problem for the treatment of, is retouched The case being made up of problem characteristic collection and solution is stated into, and is organized in a certain way, is stored in case library； When there is new problem solution, case library is retrieved, case matching is carried out, finds out and the same or analogous several cases of new problem；It is logical Cross and directly quote identical case or modification similar cases, obtain the solution of new problem, while new experience is preserved in case library, To realize dynamic learning.
The case description specifically includes：
The method that case description in database uses facing relation data model, for case i, in relation data In storehouse, input variable combination is expressed as problem characteristic vector xi=(xi1,xi2,...,xim), wherein i is represented the in database I bars record, that is, i-th of case among case library, and 1,2 ... m, represent the time point sampled in case；
The Case Retrieval specifically includes：Using nearest neighbor method, certain kind of inspection target case and case in case library Property matching degree, and by calculate each attributes match Chengdu weighted sum determine best match case；
The case modification reuses and case is modified using weighted mean method, if retrieved from case library and Present case case similarity the most similar is d (xi,x0)<5%, then each component in these cases is averaging Value, finally obtains an average vector, xi=(xi1,xi2,...,xim), add coefficient d 1, d2 obtains xmax=d1xi, xmin= d2xiIf the component in present case exceedes the component value in minimum vector sum maximum vector, using minimum component and maximum Component replaces；
The case is preserved and specifically included：To prevent that case library is excessive, the case for meeting equation below is only preserved：
min d(xi,x0) ＜ δ.
Fig. 1 is the method provided in an embodiment of the present invention based on OEE raising plan fault-tolerances, is specifically included：
Start process is monitored by underlying device, the time consumed at present in set-up procedure of starting shooting is obtained, using nearest Adjacent method, provide start and prepare remaining time；
Breakdown maintenance process is monitored by underlying device, the time consumed during obtaining at present, after matching, predicted The residue of whole maintenance process takes Trepair；
Debugging process is monitored by underlying device, the value of parameter in current debugging process is obtained, after matching, predicts whole The residue of individual debugging process takes Tdebugg；
By underlying device monitor production process, the process of production is divided startup stage, plateau, slow down and stop rank Section；
By underlying device monitor production process, according to caused percent defective, using stochastic pattern time series forecasting, If the waste product number collected at regular intervals is y1, y2, y3 ..., yn, first-order difference, Ran Houji are taken to sample first CalculateSampled correlation function, parameter is estimated using autoregressive model；Predict follow-up number of rejects Nwaste.
With reference to specific embodiment, the invention will be further described.
The method provided in an embodiment of the present invention that plan fault-tolerance is improved based on OEE, including：
Establish the case library of start time
According to process, board, personnel, preparation process (more thin better), so as to obtain a case library, following example：
。
Establish breakdown maintenance case library
According to board, fault type, maintenance personal, maintenance step (more thin better), so as to obtain a case library, such as Lower example：
。
Establish debug time case library
According to process, board, personnel, specifications and models, the different tuning parameters sampled every time, so as to obtain a case Storehouse, following example：
。
Establish velocity standard curve：
According to board, operating personnel, velocity standard curve is formulated.
Percent defective is established with reference to table：
。
Start process is monitored by underlying device, the time consumed at present in set-up procedure of starting shooting is obtained, using nearest Adjacent method, provide start and prepare remaining time.Described matching process uses nearest neighbor method, checks in target case and case library Certain attributes match degree of case, and the weighted sum by calculating each attributes match Chengdu determines best match case；Recently The formula of adjacent method is：
Wherein, x0And xiThe problem of being i-th case in present case and case library respectively characteristic vector, s (xij,x0j) It is x0And xiFeature j similarity, wjIt is feature j attribute weight.s(xi,x0) it is that present case is similar to case id Degree, its value is smaller, and matching degree is better；
Similarity Measure in nearest method, using Euclidean distance mode, Euclidean distance formula is：
d(xi, x.) present case and case i weighted euclidean distance are represented, its value is smaller, illustrates that two cases get over phase Seemingly；Wherein xijRepresent i-th case, the time-consuming situation of jth step, x0jIn the case monitored in real time in expression production process J steps take, wjRepresent the weight of correlation step；Case Retrieval uses d (xi, x.) minimum case carries out time prediction, The residue for predicting whole start set-up procedure takes TbootPrepare。
Breakdown maintenance process is monitored by underlying device, the time consumed during obtaining at present, method is with being used above It is consistent in nearest neighbor method.After finding the case of matching, the residue for predicting whole maintenance process takes Trepair；
Debugging process is monitored by underlying device, obtains the value of parameter in current debugging process, method is with using arest neighbors Consistent in method, difference is wjThe weight of relevant parameter is represented, after finding the case of matching, predicts the surplus of whole debugging process Remaining time-consuming Tdebugg；
By underlying device monitor production process, according to the characteristic of Workshop Production, the process of production is divided into 3 sections, that is, opened Dynamic stage, plateau, deceleration stop phase；
Startup stage：According to the waveform of real-time data of collection, startup stage is gone to be fitted using the equation of conic section, Remaining Stages normative reference rate curve, obtain the remaining predicted time T of production processproduce；
Y=ax2+bx+c
Plateau：According to the waveform of real-time data of collection, plateau is used into the simple method of moving average, remaining rank Section normative reference rate curve, obtains the remaining predicted time T of debugging processproduce；
Deceleration stop phase：According to the waveform of real-time data of collection, the equation by deceleration stop phase using conic section Go to be fitted, obtain the remaining predicted time T of production processproduce；
Y=ax2+ bx+c, wherein, a, b, the coefficient that c determines according to three points collected in production process.
By underlying device monitor production process, according to caused percent defective, using stochastic pattern time series forecasting, If the waste product number collected at regular intervals is y1, y2, y3 ..., yn,
First-order difference is taken to sample first：
Wherein, parameter ykIn k be k-th of sampled point；
Then calculateSampled correlation function, the sequence of calculation { Yt, t ∈ T } sample auto-covarianceWith sample from phase Close function：
Using autoregressive model, p rank autoregression model AR (p) form is：
The parameter Estimation of model uses Matrix Estimation method：
In addition, the parameter Estimation matrix of AR (1) model is：
In addition, the parameter Estimation matrix of AR (2) model is：
Follow-up number of rejects Nwaste is predicted more than.
With reference to explanation, the invention will be further described.
Remaining time predicts explanation：
1st, preparatory stage before starting shooting, the total time that prediction completion task also needs to：
Ttotal=TbootPrepare+Tdebugg+Tproduce
Wherein, TdebuggWith reference to the average time in debugging case library, TproduceWith reference to according to mark rate curve production institute Time；
2nd, debug the stage, the total time that prediction completion task also needs to：
Ttotal=Tdebugg+Tproduce
Wherein, TproduceWith reference to according to the mark rate curve production time used；
Breakdown maintenance stage in production process, the total time that prediction completion task also needs to：
Ttotal=Trepair+Tproduce
Wherein, TproduceWith reference to according to the mark rate curve production time used；
In production process, the total time that completion task also needs to is predicted：
Ttotal=Tproduce
Step S112：Waste product monitors explanation：
By RwasteWith percent defective with reference to the maximum percent defective R in tablemaxIt is compared, once Rwaste-Rmax>=0, immediately Alarm signal is produced by machinery equipment, waits artificial determination, if signal accurately can then be shut down in time, lower loss.
The foregoing is merely illustrative of the preferred embodiments of the present invention, is not intended to limit the invention, all essences in the present invention All any modification, equivalent and improvement made within refreshing and principle etc., should be included in the scope of the protection.

Claims (10)

  1. A kind of 1. method that plan fault-tolerance is improved based on OEE, it is characterised in that described that plan fault-tolerance is improved based on OEE Method includes case description, Case Retrieval, modification reuses and preservation renewal；Specially：It is pre- using the time of case-based reasioning Survey method, the case that former the problem for the treatment of is described as being made up of problem characteristic collection and solution, and in a certain way Organize, store in case library；When there is new problem solution, case library is retrieved, case matching is carried out, finds out and new problem Same or analogous several cases；By directly quoting identical case or modification similar cases, the solution of new problem is obtained, while New experience is preserved in case library, to realize dynamic learning.
  2. 2. the method for plan fault-tolerance is improved based on OEE as claimed in claim 1, it is characterised in that the case description tool Body includes：
    Case description in database using facing relation data model method, for case i, in relational database, Input variable combination is expressed as problem characteristic vector xi=(xi1,xi2,...,xim), wherein i represents i-th note in database Record, that is, i-th of case among case library, 1,2 ... m, represent the time point sampled in case；
    The Case Retrieval specifically includes：Using nearest neighbor method, certain attribute of inspection target case and case in case library With degree, and the weighted sum by calculating each attributes match Chengdu determines best match case；
    The case modification is reused and case is modified using weighted mean method, if being retrieved from case library and current case Example case similarity the most similar is d (xi,x0)<5%, then each component in these cases is averaged, finally To an average vector, xi=(xi1,xi2,...,xim), add coefficient d 1, d2 obtains xmax=d1xi, xmin=d2xiIf current case Component in example exceedes the component value in minimum vector sum maximum vector, then is replaced using minimum component and largest component；
    The case is preserved and specifically included：To prevent that case library is excessive, the case for meeting equation below is only preserved：
    mind(xi,x0) ＜ δ.
  3. 3. the method that plan fault-tolerance is improved based on OEE as described in right 1, it is characterised in that described matching process uses Nearest neighbor method, checks certain attributes match degree of case in target case and case library, and by calculate each attributes match into Weighted sum all determines best match case；The formula of nearest neighbor method is：
    <mrow> <mi>s</mi> <mrow> <mo>(</mo> <msub> <mi>x</mi> <mi>i</mi> </msub> <mo>,</mo> <msub> <mi>x</mi> <mn>0</mn> </msub> <mo>)</mo> </mrow> <mo>=</mo> <munderover> <mo>&amp;Sigma;</mo> <mrow> <mi>j</mi> <mo>=</mo> <mn>1</mn> </mrow> <mi>m</mi> </munderover> <mi>s</mi> <mrow> <mo>(</mo> <msub> <mi>x</mi> <mrow> <mi>i</mi> <mi>j</mi> </mrow> </msub> <mo>,</mo> <msub> <mi>x</mi> <mrow> <mn>0</mn> <mi>j</mi> </mrow> </msub> <mo>)</mo> </mrow> <mo>;</mo> </mrow>
    Wherein, x0And xiThe problem of being i-th case in present case and case library respectively characteristic vector, s (xij,x0j) it is x0With xiFeature j similarity, wjIt is feature j attribute weight；s(xi,x0) it is present case and case id similarities, its value Smaller, matching degree is better；
    Similarity Measure in nearest method, using Euclidean distance mode, Euclidean distance formula is：
    <mrow> <mi>d</mi> <mrow> <mo>(</mo> <msub> <mi>x</mi> <mi>i</mi> </msub> <mo>,</mo> <msub> <mi>x</mi> <mn>0</mn> </msub> <mo>)</mo> </mrow> <mo>=</mo> <msqrt> <mrow> <msup> <msub> <mi>&amp;Sigma;w</mi> <mi>j</mi> </msub> <mn>2</mn> </msup> <msup> <mrow> <mo>(</mo> <msub> <mi>x</mi> <mrow> <mi>i</mi> <mi>j</mi> </mrow> </msub> <mo>-</mo> <msub> <mi>x</mi> <mrow> <mn>0</mn> <mi>j</mi> </mrow> </msub> <mo>)</mo> </mrow> <mn>2</mn> </msup> </mrow> </msqrt> <mo>;</mo> </mrow>
    d(xi, x.) present case and case i weighted euclidean distance are represented, its value is smaller, illustrates that two cases are more similar；Wherein xijRepresent i-th case, the time-consuming situation of jth step, x0jJ steps in the case monitored in real time in expression production process It is time-consuming, wjRepresent the weight of correlation step；Case Retrieval uses d (xi, x.) minimum case carries out time prediction, predict whole The residue of individual start set-up procedure takes TbootPrepare。
  4. 4. the method for plan fault-tolerance is improved based on OEE as claimed in claim 1, it is characterised in that described to be improved based on OEE Planning the specific steps of the method for fault-tolerance includes：
    Start process is monitored by underlying device, obtains the time consumed at present in set-up procedure of starting shooting, using nearest neighbor method, Provide start and prepare remaining time；
    Breakdown maintenance process is monitored by underlying device, the time consumed during obtaining at present, after matching, predicts whole dimension The residue for repairing process takes Trepair；
    Debugging process is monitored by underlying device, the value of parameter in current debugging process is obtained, after matching, predicts whole debugging The residue of process takes Tdebugg；
    By underlying device monitor production process, the process of production is divided startup stage, plateau, deceleration stop phase；
    By underlying device monitor production process, according to caused percent defective, using stochastic pattern time series forecasting, if every The waste product number that a period of time collects is y1, y2, y3 ..., yn, takes first-order difference to sample first, then calculates's Sampled correlation function, parameter is estimated using autoregressive model；Predict follow-up number of rejects Nwaste.
  5. 5. the method that plan fault-tolerance is improved based on OEE as described in right 4, it is characterised in that opened by underlying device monitoring Machine crosses Cheng Qian, needs to carry out：
    Establish the case library of start time：According to process, board, personnel, preparation process, a case library is obtained；
    Establish breakdown maintenance case library：According to board, fault type, maintenance personal, maintenance step, a case library is obtained；
    Establish debug time case library：According to process, board, personnel, specifications and models, the different tuning parameters sampled every time, obtain To a case library；
    Establish velocity standard curve：According to board, operating personnel, velocity standard curve is formulated；
    Percent defective is established with reference to table.
  6. 6. the method that plan fault-tolerance is improved based on OEE as described in right 4, it is characterised in that
    The startup stage, including：According to the waveform of real-time data of collection, startup stage is gone to intend using the equation of conic section Close, Remaining Stages normative reference rate curve, obtain the remaining predicted time T of production processproduce；
    Y=ax2+bx+c；
    The plateau includes：According to the waveform of real-time data of collection, plateau is used into the method for moving average, Remaining Stages Normative reference rate curve, obtain the remaining predicted time T of debugging processproduce；
    The deceleration stop phase includes：According to the waveform of real-time data of collection, by deceleration stop phase using conic section Equation model, obtain the remaining predicted time T of production processproduce；
    Y=ax2+bx+c；
    Wherein, the coefficient that a, b, c determine according to three points collected in production process.
  7. 7. the method that plan fault-tolerance is improved based on OEE as described in right 4, it is characterised in that
    It is described by underlying device monitor production process, according to caused percent defective, using stochastic pattern time series forecasting, if The waste product number collected at regular intervals is y1, y2, y3 ..., yn, in, specifically include：
    1) first-order difference is taken to sample：
    <mrow> <msub> <mo>&amp;dtri;</mo> <mrow> <mi>y</mi> <mi>k</mi> </mrow> </msub> <mo>=</mo> <msub> <mi>y</mi> <mi>k</mi> </msub> <mo>-</mo> <msub> <mi>y</mi> <mrow> <mi>k</mi> <mo>-</mo> <mn>1</mn> </mrow> </msub> <mo>,</mo> <mrow> <mo>(</mo> <mi>k</mi> <mo>&amp;GreaterEqual;</mo> <mn>2</mn> <mo>)</mo> </mrow> <mo>,</mo> </mrow>
    Wherein, parameter ykIn k be k-th of sampled point；
    2) calculateSampled correlation function, the sequence of calculation { Yt, t ∈ T } sample auto-covarianceAnd sample autocorrelation function：
    <mrow> <mover> <msub> <mi>r</mi> <mi>k</mi> </msub> <mo>^</mo> </mover> <mo>=</mo> <msub> <mover> <mi>r</mi> <mo>^</mo> </mover> <mrow> <mo>-</mo> <mi>k</mi> </mrow> </msub> <mo>=</mo> <mfrac> <mn>1</mn> <mi>N</mi> </mfrac> <munderover> <mo>&amp;Sigma;</mo> <mrow> <mi>t</mi> <mo>=</mo> <mn>1</mn> </mrow> <mrow> <mi>N</mi> <mo>-</mo> <mi>k</mi> </mrow> </munderover> <msub> <mi>y</mi> <mi>t</mi> </msub> <msub> <mi>y</mi> <mrow> <mi>t</mi> <mo>+</mo> <mi>k</mi> </mrow> </msub> <mo>,</mo> <mi>k</mi> <mo>=</mo> <mn>0</mn> <mo>,</mo> <mn>1</mn> <mo>,</mo> <mn>2</mn> <mo>,</mo> <mo>...</mo> <mo>,</mo> <mi>N</mi> <mo>-</mo> <mn>1</mn> <mo>,</mo> </mrow>
    <mrow> <msub> <mover> <mi>&amp;rho;</mi> <mo>^</mo> </mover> <mi>k</mi> </msub> <mo>=</mo> <msub> <mover> <mi>&amp;rho;</mi> <mo>^</mo> </mover> <mrow> <mo>-</mo> <mi>k</mi> </mrow> </msub> <mo>=</mo> <mfrac> <mover> <msub> <mi>&amp;gamma;</mi> <mi>k</mi> </msub> <mo>^</mo> </mover> <mover> <msub> <mi>&amp;gamma;</mi> <mn>0</mn> </msub> <mo>^</mo> </mover> </mfrac> <mo>,</mo> <mi>k</mi> <mo>=</mo> <mn>0</mn> <mo>,</mo> <mn>1</mn> <mo>,</mo> <mn>2</mn> <mo>...</mo> <mo>,</mo> <mi>N</mi> <mo>-</mo> <mn>1</mn> <mo>;</mo> </mrow>
    3) autoregressive model is used, p rank autoregression model AR (p) form is：
    The parameter Estimation of model uses Matrix Estimation method：
    Wherein, the parameter Estimation matrix of AR (1) model is：
    The parameter Estimation matrix of AR (2) model is：
    4) follow-up number of rejects Nwaste is predicted more than.
  8. 8. the method that plan fault-tolerance is improved based on OEE as described in right 4, it is characterised in that the prediction of the remaining time Method, including：
    A) preparatory stage before starting shooting, the total time that prediction completion task also needs to：
    Ttotal=TbootPrepare+Tdebugg+Tproduce,
    Wherein, TdebuggWith reference to the average time in debugging case library, TproduceWith reference to according to mark rate curve production it is used when Between；
    B) debug the stage, the total time that prediction completion task also needs to：
    Ttotal=Tdebugg+Tproduce,
    Wherein, TproduceWith reference to according to the mark rate curve production time used；
    C) breakdown maintenance stage in production process, the total time that prediction completion task also needs to：
    Ttotal=Trepair+Tproduce
    Wherein, TproduceWith reference to according to the mark rate curve production time used；
    D) in production process, the total time that completion task also needs to is predicted：
    Ttotal=Tproduce.
  9. 9. the method that plan fault-tolerance is improved based on OEE as described in right 4, it is characterised in that the method for waste product monitoring includes：
    <mrow> <msub> <mi>R</mi> <mrow> <mi>w</mi> <mi>a</mi> <mi>s</mi> <mi>t</mi> <mi>e</mi> </mrow> </msub> <mo>=</mo> <mfrac> <mrow> <msub> <mi>Num</mi> <mrow> <mi>w</mi> <mi>a</mi> <mi>s</mi> <mi>t</mi> <mi>e</mi> </mrow> </msub> </mrow> <mrow> <msub> <mi>Num</mi> <mrow> <mi>t</mi> <mi>o</mi> <mi>t</mi> <mi>a</mi> <mi>l</mi> </mrow> </msub> </mrow> </mfrac> <mo>,</mo> </mrow>
    By RwasteWith percent defective with reference to the maximum percent defective R in tablemaxIt is compared, if Rwaste-Rmax>=0, pass through machinery equipment Alarm signal is produced, waits artificial determination, is then shut down in time if signal is accurate.
  10. 10. a kind of method based on OEE raising plan fault-tolerances as described in right 1 is based on OEE raising plan fault-tolerances System.
CN201710544025.1A 2017-07-05 2017-07-05 A kind of method that plan fault-tolerance is improved based on OEE Pending CN107358348A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201710544025.1A CN107358348A (en) 2017-07-05 2017-07-05 A kind of method that plan fault-tolerance is improved based on OEE

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201710544025.1A CN107358348A (en) 2017-07-05 2017-07-05 A kind of method that plan fault-tolerance is improved based on OEE

Publications (1)

Publication Number Publication Date
CN107358348A true CN107358348A (en) 2017-11-17

Family

ID=60291704

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201710544025.1A Pending CN107358348A (en) 2017-07-05 2017-07-05 A kind of method that plan fault-tolerance is improved based on OEE

Country Status (1)

Country Link
CN (1) CN107358348A (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110688757A (en) * 2019-09-30 2020-01-14 吉林大学 Method for realizing OEE dynamic based on big data driving

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101383027A (en) * 2008-09-12 2009-03-11 同济大学 Environmental emergency scheme generating method and system
CN104778369A (en) * 2015-04-20 2015-07-15 河海大学 Method and system for decision making and early warning based on ground subsidence monitoring
CN105787610A (en) * 2014-12-18 2016-07-20 中国科学院沈阳自动化研究所 Case-based reasoning method capable of supporting time sequence matching

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101383027A (en) * 2008-09-12 2009-03-11 同济大学 Environmental emergency scheme generating method and system
CN105787610A (en) * 2014-12-18 2016-07-20 中国科学院沈阳自动化研究所 Case-based reasoning method capable of supporting time sequence matching
CN104778369A (en) * 2015-04-20 2015-07-15 河海大学 Method and system for decision making and early warning based on ground subsidence monitoring

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110688757A (en) * 2019-09-30 2020-01-14 吉林大学 Method for realizing OEE dynamic based on big data driving
CN110688757B (en) * 2019-09-30 2022-10-18 吉林大学 Method for realizing OEE dynamic based on big data driving

Similar Documents

Publication Publication Date Title
KR101825881B1 (en) Method of managing a manufacturing process and system using the same
CN114626615B (en) Production process monitoring and management method and system
CN106250288A (en) Root alarm analysis and identification method based on data mining
Goyal et al. Optimization of condition-based maintenance using soft computing
CN105574593B (en) Track state static detection and control system and method based on cloud computing and big data
CN104407589A (en) Workshop manufacturing process-oriented active sensing and anomaly analysis method of real-time generating performance
Ungermann et al. Data analytics for manufacturing systems–a data-driven approach for process optimization
CN117172414A (en) Building curtain construction management system based on BIM technology
CN111126217A (en) Intelligent operation and maintenance management system for power transmission line based on intelligent identification
Wang et al. Innovative smart scheduling and predictive maintenance techniques
CN117172509B (en) Construction project distribution system based on decoration construction progress analysis
D'Emilia et al. Data validation techniques for measurements systems operating in a Industry 4.0 scenario a condition monitoring application
Nicoletti Lean and automate manufacturing and logistics
CN106054832B (en) Multivariable-based dynamic online monitoring method and device for intermittent chemical production process
CN110781206A (en) Method for predicting whether electric energy meter in operation fails or not by learning meter-dismantling and returning failure characteristic rule
CN117709617A (en) MES-based intelligent scheduling system for production workshop
CN107358348A (en) A kind of method that plan fault-tolerance is improved based on OEE
Chen et al. Exploring equipment electrocardiogram mechanism for performance degradation monitoring in smart manufacturing
CN115640980A (en) Power grid engineering cost dynamic management system based on target control
Breitenbach et al. A systematic literature review of machine learning tools for supporting supply chain management in the manufacturing environment
CN115471135A (en) Environment-friendly service analysis system
Ming Decision approach of maintenance for urban rail transit based on equipment supervision data mining
Huang et al. Towards automated model calibration and validation in rail transit simulation
Albayrak et al. Digitalization of a Steel Pipe Production Factory: STEEL4. 0-A Family of Products Developed on Routes from Industry 3.0 to Industry 4.0
Lee et al. Killer applications of industrial AI

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
RJ01 Rejection of invention patent application after publication
RJ01 Rejection of invention patent application after publication

Application publication date: 20171117