CN108510180A - The computational methods of performance interval residing for a kind of production equipment - Google Patents
The computational methods of performance interval residing for a kind of production equipment Download PDFInfo
- Publication number
- CN108510180A CN108510180A CN201810264936.3A CN201810264936A CN108510180A CN 108510180 A CN108510180 A CN 108510180A CN 201810264936 A CN201810264936 A CN 201810264936A CN 108510180 A CN108510180 A CN 108510180A
- Authority
- CN
- China
- Prior art keywords
- index
- cluster
- data
- equipment
- performance
- 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
- 238000004519 manufacturing process Methods 0.000 title claims abstract description 34
- 238000000205 computational method Methods 0.000 title claims abstract description 9
- 238000000034 method Methods 0.000 claims description 42
- 238000012423 maintenance Methods 0.000 claims description 12
- 239000000463 material Substances 0.000 claims description 8
- 230000000694 effects Effects 0.000 claims description 7
- 238000003064 k means clustering Methods 0.000 claims description 6
- 238000005265 energy consumption Methods 0.000 claims description 5
- 238000012797 qualification Methods 0.000 claims description 5
- 238000004364 calculation method Methods 0.000 claims description 4
- 238000005457 optimization Methods 0.000 claims description 3
- 230000004888 barrier function Effects 0.000 claims 1
- 230000032258 transport Effects 0.000 claims 1
- 238000011156 evaluation Methods 0.000 abstract description 22
- 238000004422 calculation algorithm Methods 0.000 abstract description 7
- 238000012544 monitoring process Methods 0.000 abstract description 5
- 230000002452 interceptive effect Effects 0.000 abstract description 2
- 239000011159 matrix material Substances 0.000 description 6
- 241001269238 Data Species 0.000 description 4
- 239000002994 raw material Substances 0.000 description 4
- 238000005516 engineering process Methods 0.000 description 3
- 238000004458 analytical method Methods 0.000 description 2
- 238000013210 evaluation model Methods 0.000 description 2
- 238000005259 measurement Methods 0.000 description 2
- 238000012545 processing Methods 0.000 description 2
- 238000011160 research Methods 0.000 description 2
- 238000012549 training Methods 0.000 description 2
- 230000002159 abnormal effect Effects 0.000 description 1
- 238000013459 approach Methods 0.000 description 1
- 238000007635 classification algorithm Methods 0.000 description 1
- 230000002860 competitive effect Effects 0.000 description 1
- 230000007812 deficiency Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000012854 evaluation process Methods 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 238000003754 machining Methods 0.000 description 1
- 238000013178 mathematical model Methods 0.000 description 1
- 238000010606 normalization Methods 0.000 description 1
- 238000007500 overflow downdraw method Methods 0.000 description 1
- 238000013139 quantization Methods 0.000 description 1
- 238000012706 support-vector machine Methods 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
- 239000013598 vector Substances 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
- G06Q10/0639—Performance analysis of employees; Performance analysis of enterprise or organisation operations
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/23—Clustering techniques
- G06F18/232—Non-hierarchical techniques
- G06F18/2321—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions
- G06F18/23213—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions with fixed number of clusters, e.g. K-means clustering
-
- 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
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/04—Manufacturing
-
- 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
- Y02P90/30—Computing systems specially adapted for manufacturing
Landscapes
- Engineering & Computer Science (AREA)
- Business, Economics & Management (AREA)
- Human Resources & Organizations (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Strategic Management (AREA)
- Economics (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Development Economics (AREA)
- Entrepreneurship & Innovation (AREA)
- Marketing (AREA)
- Tourism & Hospitality (AREA)
- Educational Administration (AREA)
- General Business, Economics & Management (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Quality & Reliability (AREA)
- Evolutionary Computation (AREA)
- Evolutionary Biology (AREA)
- Bioinformatics & Computational Biology (AREA)
- Game Theory and Decision Science (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Artificial Intelligence (AREA)
- Operations Research (AREA)
- General Engineering & Computer Science (AREA)
- Life Sciences & Earth Sciences (AREA)
- Probability & Statistics with Applications (AREA)
- Manufacturing & Machinery (AREA)
- Health & Medical Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Primary Health Care (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a kind of computational methods of performance interval residing for production equipment, it is related to calculating the study on evaluation way for judging performance interval situation residing for equipment according to section and operation Real-time Monitoring Data.Compared to traditional clustering algorithm, have the advantages that cluster result is more accurate, precision higher, each index in actual production is influenced size to equipment performance to take into account, highlights the weight of larger impact index, final performance cluster centre is made to be more in line with equipment operation actual conditions;By defining maximum cluster radius in mathematics level, it is different from other kinds of fuzzy division, show that an accurate performance interval divides;The acquisition window size in specific mathematics level is defined, when the data in acquisition window are all divided into same section, equipment performance status evaluation can be just obtained, specify the precision of evaluation, also avoids interfering data to influence caused by evaluation result individually.
Description
Technical field
The present invention relates to a kind of historical datas based on equipment operation to be modeled to equipment performance interval division, and
The study on evaluation way for judging performance interval situation residing for equipment is calculated according to section and operation Real-time Monitoring Data.
Background technology
With the development of industrial technology and the expansion of manufacturing enterprise's scale, manufacturing technology is by previous semi-artificial semi-mechanization
Gradually develop to the intelligentized direction of comprehensive mechanization, the more and more extensive utilization of highly complex accurate and expensive equipment
In on enterprise's production line.The manufacturing equipment operation of " Production line type " enterprise has the characteristics that security requirement is high, production is continuous, one
Denier equipment performance declines or breaks down, and can not only increase enterprise's maintenance cost, the influence to enterprise's prouctiveness also will be non-
Chang Ju great takes specific aim maintenance measure that producing line is made to keep good prouctiveness so understanding each equipment performance situation in time, right
It retains its competitive edge in fierce market competition in enterprise and plays very important effect.
Conceptually make simple general statement however, being currently presented as the division in equipment performance section more, for
Selected and performance interval the division of equipment performance index lacks standard, quantization model, and is drawn for performance interval
Point and carry out in the main stream approach research of performance evaluation, using simple analytic hierarchy process (AHP) or simple data fusion method,
Precision is relatively low, larger to index weights subjective impact and lack representativeness therefore more in terms of interval division and evaluation accuracy
It selects and tax power is carried out to multi objective to calculate the objective weighted model of weight coefficient based on decision matrix information founding mathematical models.
Wherein comentropy is to uncertain measurement, and the lower uncertainty of entropy is smaller, and indication information amount is bigger, and weight is bigger, otherwise also
So, but ignore correlation between index.CRITIC be according between each index to specific strength and conflicting come combination weights, mainly according to
The related coefficient between index is relied to be calculated to be used to assign power, two methods, which combine, can be very good to utilize actual information data and index
Between the degree of correlation obtain each index more reasonably objective weight.K-means algorithms think each index of analyzed sample poly-
Effect is identical in class, and support vector machines is for the big sample training difficulty of data volume larger and main supports two classification
Algorithm, the equipment performance this practical application more than two classification of classifying need Multiple Training that could complete to classify.For above-mentioned
The deficiency of each method, proposition one kind are combined to assign based on CRITIC combining information entropy methods and be weighed to improve K-means clustering algorithms,
Clustering is carried out to device history data, obtains relatively accurate equipment performance subregion.
Invention content
The present invention is for studying less situation and absent standardized to equipment performance interval division at present, with high accuracy set
Standby method of evaluating performance provides a kind of historical data monitored when run based on equipment to the progress of equipment performance interval division
Model and judge according to section and Real-time Monitoring Data the evaluation method of performance interval residing for equipment.This method, which is directed to, uses K-
Means algorithms act on identical disadvantage in each Criterion Attribute of analytical equipment in the cluster for dividing performance interval, lead to area
Between the not high problem and information Entropy Method of cluster accuracy the problems such as only ignoring correlation between index according to data information, in conjunction with
CRITIC obtains the comprehensive weight of each index with comentropy method, then goes out to assemble focus by Multiple Attribute Clustering, obtains to equipment
Performance interval divides;
The technical scheme is that a kind of computational methods of performance interval residing for production equipment, and further according in real time
Monitoring data calculate the method for judging performance interval residing for equipment, and this method includes:
Step 1:Choose Key Performance Indicator in target device production activity;
Further, the choosing method of Key Performance Indicator is in the step 1:
The index of the horizontal aspect of quality capability includes:Product standard qualification rate, product quality score, product quality lose shadow
It rings;
Index in terms of Operating ettectiveness situation includes:Equipment fortune production efficiency, mean time between failures, equipment are stopped greatly
Machine failure rate;
Production consumption situation in terms of index include:Material loss rate, the auxiliary material proportion of goods damageds, energy consumption levels;
Index in terms of plant maintenance feedback includes:Average maintenance times averagely safeguard that duration, average safeguard are lost;
From one or more indexs are selected in said one or many aspects respectively as in target device production activity
Key Performance Indicator;
Step 2:The weight of each attribute is determined according to CRITIC and comentropy method respectively, then true using VC Method
Determine combination coefficient optimization Criterion Attribute comprehensive weight;
Step 2.1:The associated weights of each index are calculated using CRITIC methods;
1) data for obtaining index, are normalized each achievement data, and the achievement data that same time period obtains is one
Group;
2) degree of correlation that j-th of index and other indexs are calculated in each group is:Wherein rijIt is i-th
Related coefficient between index and j-th of index;
Calculate the information content C that j-th of index is includedj:
Wherein:σjIndicate j-th index to specific strength;CjBigger, the information content that j-th of evaluation index is included is bigger,
The relative importance of the index is also bigger;
Calculate the associated weights θ of j-th of indexj:
Step 2.2:Using the entropy weight weight of each index of the method for comentropy;
1) index of acquisition is normalized, all data of an index are divided into one group;
2) the comentropy H of j-th of assessment factor is calculatedj:
Wherein:fijIndicate that the probability function of i-th of data in j-th of index, calculation formula are:Indicate that i-th of data of j-th of index after normalizing, m indicate data total number in each index;
3) the entropy weight weight W of each index is calculatedj:
Wherein:N indicates the total number of index;
Step 2.3:The combining weights w of j-th of index is calculated using VC Methodj;
wj=a θj+βWj
Wherein:Againp1,p2..., pnWhat expression was calculated
The entropy weight of each index weighs the correspondence numerical value after rearranging from small to large;
Step 3:K-means clustering methods are used after each index is enclosed corresponding weight, are k by each achievement data cluster
A class, different classes indicate the different performance section residing for target device;
Further, the k-means clustering methods in step 3 are:
Step 3.1:K number is arbitrarily chosen from all data according to as initial cluster center c1,c2,…,ck;
Step 3.2:With c1,c2,…,ckAs initial cluster center, each data are calculated at a distance from each cluster centre, it will
Each data are divided into the clustering cluster belonging to the cluster centre minimum with its distance, and k cluster is obtained after having traversed all data
Cluster;
Step 3.3:To step 3.2 to k clustering cluster in after each data enclose the combining weights that step 2 obtains, weight
Newly calculate the cluster centre c of each cluster1',c'2,…,c'k;
Step 3.4:The cluster centre of a new generation, Zhi Daoman are calculated using step 3.2 and the identical method of step 3.4
Foot cluster termination condition, obtains k cluster centre and k clustering cluster;
Step 3.5:The data for being more than R with distance in the cluster of the clustering cluster are rejected in each clustering cluster, according to wherein R
The threshold value of actual conditions setting, obtains final cluster result.
Further, R/R in the step 3.5max=0.95, wherein RmaxIt indicates in the clustering cluster in data point to cluster
The maximum distance of the heart.
Step 4:Target device achievement data is obtained in real time, and the achievement data for calculating acquisition belongs to which step 3 obtains
A clustering cluster, then assertive goal equipment belong to the corresponding equipment performance section of the clustering cluster.
Further, described to assert that current target is set after the M data continuously obtained belong to the same clustering cluster
It is standby to belong to the corresponding equipment performance section of the clustering cluster.
The present invention has the advantages that cluster result is more accurate, precision higher compared to traditional clustering algorithm, will be practical
Each index takes into account equipment performance influence size in production, highlights the weight of larger impact index, makes final property
Energy cluster centre is more in line with equipment operation actual conditions;By defining maximum cluster radius in mathematics level, it is different from other
The fuzzy division of type show that an accurate performance interval divides;The acquisition window size in specific mathematics level is defined,
When the data in acquisition window are all divided into same section, it can just obtain equipment performance status evaluation, specify evaluation
Precision also avoids interfering data to influence caused by evaluation result individually.
Description of the drawings
Fig. 1 is the specific implementation mode flow chart of the method for the present invention;
Fig. 2 is equipment key performance performance analysis illustraton of model;
Fig. 3 is equipment performance evaluation index layer model figure;
Fig. 4 equipment performances interval division and performance evaluation step.
Specific implementation mode
(such as Fig. 1) is described in detail to the embodiment of the present invention journey below, the present embodiment journey is being with technical solution of the present invention
Under the premise of implemented, give detailed embodiment and specific operating process, but protection scope of the present invention is not limited to
Following implementation routines.
Implement routine and can be mainly divided into following steps:
Step 1:Equipment production active critical performance indicator and structure are formulated with S-I-P-O-C-F models (such as Fig. 2)
Performance evaluation key index analysis model.
There are four aspects for the key dimension of equipment complex performance:
One, quality level.The core of equipment performance evaluation is the product quality of equipment production.For key quality parameters,
Quantitatively evaluating product is inside and outside in quality.
Two, Operating ettectiveness.Can the basic assessment dimension that performance be weighed be that normal operation shape is kept under use environment
State.The operation characteristics such as key evaluation process capability, machining accuracy, reliability.
Three, supplies consumption.In terms of input and output angle, the consumption level to raw material, auxiliary material is also that performance evaluation is important
With reference to.
Four, feedback is safeguarded.Maintenance process mean cost increase production it is slack-off, maintenance times and maintenance time are all performances
Evaluate important feedback index.
Global index is designed a model with quality, efficiency, loss, is safeguarded as longitudinal dimension, customer, production manager, equipment
The key related side such as guardian can be related to different longitudinal dimensions, and overall evaluation index need to be identified fully, embody each crucial correlation
The crucial requirement of different longitudinal direction dimensions, it is ensured that the science of final appraisal results with it is reliable.The pass of equipment performance comprehensive evaluation model
Key performance indicator, as shown in Figure 3.
Step 2:The weight of each attribute is determined according to CRITIC and comentropy method respectively, then true using VC Method
Determine combination coefficient optimization Criterion Attribute comprehensive weight;
Step 2.1:CRITIC methods determine attribute weight
1) by index matrix A=(aij)m×n, the index matrix B=(b of standardization are calculated according to formula (1) formula (2)ij)m×n;It is right
In profit evaluation model index, can generally enable:
For cost type index, can enable:
In formulaI.e.Respectively
J index ajMaximum value and minimum value.Some experimental data after normalization is as shown in table 1:
Table 1 normalizes each Performance Evaluating Indexes data instance
2) each index weights are calculated according to CRITIC French (3) and formula (4).J-th of index conflicts with other indexs
Property quantizating index is:Wherein rijRelated coefficient between evaluation index i and j.
The objective weight of each index is exactly to integrate measurement to specific strength and conflicting.If CjIt indicates to comment for j-th
The information content that valence index is included, then CjIt can be expressed as:
CjBigger, the information content that j-th of evaluation index is included is bigger, and the relative importance of the index is also bigger, institute
With the objective weight of j-th of index:
Each index CRITIC weighted values being calculated through upper formula and experimental data are as shown in table 2:
2 CRITIC methods of table calculate each index weights
Index | Weighted value |
Product standard qualification rate | 0.27 |
Product quality score | 0.15 |
Product quality loss influences | 0.08 |
Equipment runs production efficiency | 0.12 |
Mean time between failures | 0.16 |
The big machine halt trouble rate of equipment | 0.04 |
The raw material proportion of goods damageds | 0.08 |
Energy consumption levels | 0.02 |
Auxiliary material is lost | 0.02 |
Average maintenance times | 0.02 |
Averagely safeguard duration | 0.01 |
It is average to safeguard loss | 0.03 |
Step 2.2:Comentropy obtains attribute objective weight
If it is T={ t to normalize obtained matrixij}m×n, for the assessment factor of more bigger more excellent type, have
For the assessment factor of smaller more excellent type, have
In formula, for the factor of more bigger more excellent type, cmaxIt indicates to best suit claimer, c in all information datas of the factormin
Then indicate the person that least do not meet the requirements;Opposite, for the factor of smaller more excellent type, cmaxIt indicates in all information datas of the factor
Least do not meet the requirements person, cminIt then indicates to best suit claimer.
Using normalized matrix T as object, according to the calculation formula of comentropy, the comentropy of j-th of assessment factor is:
fijThe probability function of each information datas of expression factor j, calculation formula are:
Work as tijWhen=0, the f in formula (6)ij=0, this up-to-date style (7) is invalid.Therefore it needs to be modified (8):
Finally calculate the information entropy weight of each factor:
Wherein:Meet the normalizing condition of weight.
Each indication information entropy weight weight values being calculated through formula above and experimental data are as shown in table 3:
3 information Entropy Method of table calculates each index weights
Index | Weighted value |
Product standard qualification rate | 0.12 |
Product quality score | 0.10 |
Product quality loss influences | 0.06 |
Equipment runs production efficiency | 0.14 |
Mean time between failures | 0.08 |
The big machine halt trouble rate of equipment | 0.10 |
The raw material proportion of goods damageds | 0.06 |
Energy consumption levels | 0.04 |
Auxiliary material is lost | 0.08 |
Average maintenance times | 0.07 |
Averagely safeguard duration | 0.09 |
It is average to safeguard loss | 0.06 |
Step 2.3:Combining weights are determined using VC Method
Wherein T is the coefficient of variation of each components of θ,p1,
p2,…,pnIt is each component rearranging from small to large in CRITIC weight vectors w ', n is the number of attribute.Then β=1- α, generation
Enter wj=a θj+βWjObtain the comprehensive weight w of each indexj。
Each indicator combination weighted value being calculated through each formula of VC Method is as shown in table 4:
4 each indicator combination weighted value of table
Index | Weighted value |
Product standard qualification rate | 0.20 |
Product quality score | 0.13 |
Product quality loss influences | 0.07 |
Equipment runs production efficiency | 0.13 |
Mean time between failures | 0.12 |
The big machine halt trouble rate of equipment | 0.07 |
The raw material proportion of goods damageds | 0.07 |
Energy consumption levels | 0.03 |
Auxiliary material is lost | 0.05 |
Average maintenance times | 0.04 |
Averagely safeguard duration | 0.03 |
It is average to safeguard loss | 0.06 |
Step 3:Each weight and index attribute value are clustered, obtains the division in equipment performance section, and is set in maximum cluster
Heart radius Rmax.Input:Data set X to be clustered, cluster number K;Output:K cluster.
K-means algorithms set k initial cluster centres first, will according to k cluster centre in iteration each time
Data sample is divided into k clustering cluster, and after the completion of division, calculating the center of each cluster, (all object distances is flat i.e. in cluster
Value), as the cluster centre reference point of next round, by successive ignition, the cluster centre of acquisition will become closer to really
Clustering cluster center, object function is smaller and smaller, and the quality of cluster can also become better and better.
It is as follows based on k-means clustering algorithm concrete processing procedures:
(1) k number is arbitrarily chosen according to as initial cluster center c from n normal data objects1,c2,…,ck。
(2) with c1,c2,…,ckAs initial cluster center, the division of clustering cluster is carried out according to following to set of data samples
Principle:If dij(xi,cj) < dim(xi,cm), wherein m=1,2 ..., k;j≠m;I=1,2 ..., n, by data sample xiIt divides
To clustering cluster cjThe inside.
(3) according to formulaRecalculate cluster centre c1',c'2,…,c'k。
(4) if for arbitrary j ∈ (1,2 ..., k), c'j=cjThen terminate, and c1',c'2,…,c'kIt represents last
Otherwise cluster centre enables cj=c'j, (2) are re-executed, until meeting maximum iteration.
(5) cluster result and final normal data cluster centre are exported.
(6) data point X to be measured is taken, r is enabledi=| x-ci| it is point X to i-th cluster class center c of testing dataiDistance, if
max(ri) > R when, R is the set reasonable threshold value of experiment, then it is assumed that the data point is not belonging to any one normal clustering cluster.
The maximum cluster radius R of definition:
Using non-standard numerical method, R is definedmaxSize it is related with the cluster result of historical data, according to different poly-
Class center chooses the historical data 95% dot matrix coverage area nearest apart from cluster centre and makees approximate circle, it is practical big to find out it
It is small.
U/n=0.95, u ∈ (0, n) and Rmax=| xu-ci|
Wherein xuIt is u small from ith cluster center ciThe sampled data points of distance.
According to the research of front it is found that the section in relation to production equipment performance evaluation has been completed division, effect master
To be used for instructing in production application scene, same category of device in the process of running, is commented according to the creation data monitored
Valence equipment performance interval residing at present, because the spherical feature in cluster section, always has some abnormal points and fall on all areas
Between it is external, so needing set one size as the data acquisition window mouth of M, because variation data belong to minority after all, equipment is transported
Performance is chronically at metastable state between the departure date, only when the data of continuous acquisition to window size belong to the poly- of certain one kind
When the maximum radius range of class center, you can judge equipment actual performance situation, so not considering a small number of variation data, only consider normal
Data are advised, processing is omitted for minority variation data when calculating.
Equipment performance interval division and performance evaluation step are as shown in Figure 4;
Go out cluster points (X in equipment different performance section according to historical data clustering1,X2,X3,...,Xk) respectively
Indicate performance interval (V1,V2,V3,...,Vk) number counted in cluster radius, wherein k is classification number, first according to confidence
Degree takes the 5% of each interval clustering points of historical data to be used as each section acquisition window Nx, x=1,2 ..., k, for new equipment
The setting of online monitoring data acquisition window size N, can define N=max { Nx, x=1,2 ..., k, when more than collecting window N
Data belong to a certain section simultaneously and be, can definition apparatus be in certain performance interval.
Claims (5)
1. the computational methods of performance interval, this method residing for a kind of production equipment include:
Step 1:Choose the index that key performance is regarded as in target device production activity;
Step 2:The weight of each attribute is determined according to CRITIC and comentropy method respectively, then group is determined using VC Method
Collaboration number optimization Criterion Attribute comprehensive weight;
Step 2.1:The associated weights of each index are calculated using CRITIC methods;
1) data for obtaining index, are normalized each achievement data, and the achievement data that same time period obtains is one group;
2) degree of correlation that j-th of index and other indexs are calculated in each group is:Wherein rijFor i-th of index
With the related coefficient between j-th of index;
Calculate the information content C that j-th of index is includedj:
Wherein:σjIndicate j-th index to specific strength;
Calculate the associated weights θ of j-th of indexj:
Step 2.2:Using the entropy weight weight of each index of the method for comentropy;
1) index of acquisition is normalized, all data of an index are divided into one group;
2) the comentropy H of j-th of assessment factor is calculatedj:
Wherein:fijIndicate that the probability function of i-th of data in j-th of index, calculation formula are:
Indicate that i-th of data of j-th of index after normalizing, m indicate data total number in each index;
3) the entropy weight weight W of each index is calculatedj:
Wherein:N indicates the total number of index;
Step 2.3:The combining weights w of j-th of index is calculated using VC Methodj;
wj=a θj+βWj
Wherein:Againp1,p2,…,pnIndicate each index being calculated
Entropy weight weight rearrange from small to large after correspondence numerical value;
Step 3:K-means clustering methods are used after each index is enclosed corresponding weight, are k by each achievement data cluster
Class, different classes indicate the different performance section residing for target device;
Step 4:In real time obtain target device achievement data, calculate acquisition achievement data belong to step 3 obtain which gather
Class cluster, then assertive goal equipment belong to the corresponding equipment performance section of the clustering cluster.
2. the computational methods of performance interval residing for a kind of production equipment as described in claim 1, it is characterised in that the step 1
The choosing method of middle Key Performance Indicator is:
The index of the horizontal aspect of quality capability includes:Product standard qualification rate, product quality score, product quality loss influence;
Index in terms of Operating ettectiveness situation includes:Equipment transports the big shutdown event of production efficiency, mean time between failures, equipment
Barrier rate;
Production consumption situation in terms of index include:Material loss rate, the auxiliary material proportion of goods damageds, energy consumption levels;
Index in terms of plant maintenance feedback includes:Average maintenance times averagely safeguard that duration, average safeguard are lost;
From one or more indexs are selected in said one or many aspects respectively as crucial in target device production activity
Performance indicator.
3. the computational methods of performance interval residing for a kind of production equipment as described in claim 1, it is characterised in that the step 3
In k-means clustering methods be:
Step 3.1:K number is arbitrarily chosen from all data according to as initial cluster center c1,c2,…,ck;
Step 3.2:With c1,c2,…,ckAs initial cluster center, each data are calculated at a distance from each cluster centre, by each number
According to being divided into the clustering cluster belonging to the cluster centre minimum with its distance, k clustering cluster is obtained after having traversed all data;
Step 3.3:To step 3.2 to k clustering cluster in after each data enclose the combining weights that step 2 obtains, count again
Calculate the cluster centre c ' of each cluster1,c′2,…,c′k;
Step 3.4:The cluster centre of a new generation is calculated using step 3.2 and the identical method of step 3.4, it is poly- until meeting
Class termination condition obtains k cluster centre and k clustering cluster;
Step 3.5:The data that distance in the cluster with the clustering cluster is more than R are rejected in each clustering cluster, wherein R is according to reality
The threshold value of situation setting, obtains final cluster result.
4. the computational methods of performance interval residing for a kind of production equipment as claimed in claim 3, it is characterised in that the step
R/R in 3.5max=0.95, wherein RmaxIndicate in the clustering cluster data point to the maximum distance of cluster centre.
5. the computational methods of performance interval residing for a kind of production equipment as claimed in claim 4, it is characterised in that it is further,
It is described to assert that current target equipment belongs to the clustering cluster correspondence after M data continuously obtained belong to the same clustering cluster
Equipment performance section.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810264936.3A CN108510180B (en) | 2018-03-28 | 2018-03-28 | Method for calculating performance interval of production equipment |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810264936.3A CN108510180B (en) | 2018-03-28 | 2018-03-28 | Method for calculating performance interval of production equipment |
Publications (2)
Publication Number | Publication Date |
---|---|
CN108510180A true CN108510180A (en) | 2018-09-07 |
CN108510180B CN108510180B (en) | 2021-08-06 |
Family
ID=63379085
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201810264936.3A Active CN108510180B (en) | 2018-03-28 | 2018-03-28 | Method for calculating performance interval of production equipment |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN108510180B (en) |
Cited By (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109144839A (en) * | 2018-10-09 | 2019-01-04 | 郑州云海信息技术有限公司 | The acquisition methods and device of equipment state |
CN110889468A (en) * | 2019-12-23 | 2020-03-17 | 长沙理工大学 | Multi-model uncertain reproduction result analysis method capable of eliminating error information |
CN112446509A (en) * | 2020-11-10 | 2021-03-05 | 中国电子科技集团公司第三十八研究所 | Complex electronic equipment prediction maintenance method |
CN115757076A (en) * | 2022-12-09 | 2023-03-07 | 国网湖北省电力有限公司信息通信公司 | Cluster management fitness calculation method based on equipment operation characteristics and performance indexes |
CN116227754A (en) * | 2023-05-09 | 2023-06-06 | 张家港大裕橡胶制品有限公司 | Production self-adaptive optimization control method and system for rubber gloves |
Citations (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN1684082A (en) * | 2005-02-18 | 2005-10-19 | 北京大学深圳医院 | Evaluation management system and its method for modern hospital performance |
US20110060603A1 (en) * | 2009-09-09 | 2011-03-10 | Capelli Christopher C | Population Adjusted Indexes |
CN105512448A (en) * | 2014-09-22 | 2016-04-20 | 国家电网公司 | Power distribution network health index assessment method |
CN105654229A (en) * | 2015-11-26 | 2016-06-08 | 国家电网公司 | Power grid automation system and equipment running state risk assessment algorithm |
CN106372952A (en) * | 2016-11-14 | 2017-02-01 | 北京创业公社征信服务有限公司 | Objective and subjective weight determining multi-model compositional verification-based enterprise credit assessment method and system |
CN106709654A (en) * | 2016-12-28 | 2017-05-24 | 中南大学 | Global operating condition evaluating and quality tracing method for hydrocracking process |
CN107038460A (en) * | 2017-04-10 | 2017-08-11 | 南京航空航天大学 | A kind of ship monitor shortage of data value complementing method based on improvement KNN |
CN107330458A (en) * | 2017-06-27 | 2017-11-07 | 常州信息职业技术学院 | A kind of fuzzy C-means clustering method of minimum variance clustering of optimizing initial centers |
CN107679677A (en) * | 2017-10-30 | 2018-02-09 | 华南农业大学 | A kind of drought changes the preferential area's system of selection of water |
-
2018
- 2018-03-28 CN CN201810264936.3A patent/CN108510180B/en active Active
Patent Citations (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN1684082A (en) * | 2005-02-18 | 2005-10-19 | 北京大学深圳医院 | Evaluation management system and its method for modern hospital performance |
US20110060603A1 (en) * | 2009-09-09 | 2011-03-10 | Capelli Christopher C | Population Adjusted Indexes |
CN105512448A (en) * | 2014-09-22 | 2016-04-20 | 国家电网公司 | Power distribution network health index assessment method |
CN105654229A (en) * | 2015-11-26 | 2016-06-08 | 国家电网公司 | Power grid automation system and equipment running state risk assessment algorithm |
CN106372952A (en) * | 2016-11-14 | 2017-02-01 | 北京创业公社征信服务有限公司 | Objective and subjective weight determining multi-model compositional verification-based enterprise credit assessment method and system |
CN106709654A (en) * | 2016-12-28 | 2017-05-24 | 中南大学 | Global operating condition evaluating and quality tracing method for hydrocracking process |
CN107038460A (en) * | 2017-04-10 | 2017-08-11 | 南京航空航天大学 | A kind of ship monitor shortage of data value complementing method based on improvement KNN |
CN107330458A (en) * | 2017-06-27 | 2017-11-07 | 常州信息职业技术学院 | A kind of fuzzy C-means clustering method of minimum variance clustering of optimizing initial centers |
CN107679677A (en) * | 2017-10-30 | 2018-02-09 | 华南农业大学 | A kind of drought changes the preferential area's system of selection of water |
Non-Patent Citations (3)
Title |
---|
丁晓琴 等: "基于 AHP 和 CRITIC 综合赋权的 K-means 算法", 《计算机系统应用》 * |
刘恒宇: "基于性能退化的车身焊装生产线动态预防性维护方法研究", 《中国优秀硕士学位论文全文数据库 经济与管理科学辑》 * |
梁寒冰: "基于改进GRA法构建P2P网贷平台风险评价体系", 《财会月刊》 * |
Cited By (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109144839A (en) * | 2018-10-09 | 2019-01-04 | 郑州云海信息技术有限公司 | The acquisition methods and device of equipment state |
CN110889468A (en) * | 2019-12-23 | 2020-03-17 | 长沙理工大学 | Multi-model uncertain reproduction result analysis method capable of eliminating error information |
CN112446509A (en) * | 2020-11-10 | 2021-03-05 | 中国电子科技集团公司第三十八研究所 | Complex electronic equipment prediction maintenance method |
CN112446509B (en) * | 2020-11-10 | 2023-05-26 | 中国电子科技集团公司第三十八研究所 | Prediction maintenance method for complex electronic equipment |
CN115757076A (en) * | 2022-12-09 | 2023-03-07 | 国网湖北省电力有限公司信息通信公司 | Cluster management fitness calculation method based on equipment operation characteristics and performance indexes |
CN115757076B (en) * | 2022-12-09 | 2023-10-27 | 国网湖北省电力有限公司信息通信公司 | Cluster management fitness computing method based on equipment operation characteristics and performance indexes |
CN116227754A (en) * | 2023-05-09 | 2023-06-06 | 张家港大裕橡胶制品有限公司 | Production self-adaptive optimization control method and system for rubber gloves |
Also Published As
Publication number | Publication date |
---|---|
CN108510180B (en) | 2021-08-06 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN108510180A (en) | The computational methods of performance interval residing for a kind of production equipment | |
CN106355030B (en) | A kind of fault detection method based on analytic hierarchy process (AHP) and Nearest Neighbor with Weighted Voting Decision fusion | |
CN106371427B (en) | Industrial process Fault Classification based on analytic hierarchy process (AHP) and fuzzy Fusion | |
CN105117602B (en) | A kind of metering device running status method for early warning | |
CN107247737B (en) | The analysis of platform area default electricity use and method for digging based on electricity consumption | |
CN108564286A (en) | A kind of artificial intelligence finance air control credit assessment method and system based on big data reference | |
CN111324642A (en) | Model algorithm type selection and evaluation method for power grid big data analysis | |
CN104239712B (en) | Real-time evaluation method for anti-interference performance of radar | |
CN110658487A (en) | Meter box and system capable of achieving intelligent electric meter error online estimation | |
CN110309886A (en) | The real-time method for detecting abnormality of wireless sensor high dimensional data based on deep learning | |
CN111429034A (en) | Method for predicting power distribution network fault | |
CN110674993A (en) | User load short-term prediction method and device | |
CN104123678A (en) | Electricity relay protection status overhaul method based on status grade evaluation model | |
CN113887908A (en) | Bridge risk assessment method considering subjective and objective cross fusion weight | |
CN110135708A (en) | Ball valve quality multi-step Fuzzy Comprehensive Evaluation based on AHP and comentropy | |
CN109633518A (en) | A kind of intelligent electric energy meter comprehensive performance evaluation method, apparatus | |
CN113627735A (en) | Early warning method and system for safety risk of engineering construction project | |
CN111949939A (en) | Intelligent electric meter running state evaluation method based on improved TOPSIS and cluster analysis | |
CN109934420A (en) | A kind of method and system for predicting labor turnover | |
CN110705859A (en) | PCA-self-organizing neural network-based method for evaluating running state of medium and low voltage distribution network | |
CN116187835A (en) | Data-driven-based method and system for estimating theoretical line loss interval of transformer area | |
CN115660170A (en) | Multidimensional index weight collaborative optimization data asset management effect differentiation evaluation method and system | |
CN114548494A (en) | Visual cost data prediction intelligent analysis system | |
CN109685133A (en) | The data classification method of prediction model low cost, high discrimination based on building | |
CN108830407A (en) | Sensor distribution optimization method under the conditions of multi-state in monitoring structural health conditions |
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 |