CN108647808A - A kind of manufacturing parameter Optimization Prediction method, apparatus, equipment and storage medium - Google Patents
A kind of manufacturing parameter Optimization Prediction method, apparatus, equipment and storage medium Download PDFInfo
- Publication number
- CN108647808A CN108647808A CN201810322649.3A CN201810322649A CN108647808A CN 108647808 A CN108647808 A CN 108647808A CN 201810322649 A CN201810322649 A CN 201810322649A CN 108647808 A CN108647808 A CN 108647808A
- Authority
- CN
- China
- Prior art keywords
- association
- chain
- monitoring data
- data
- inter
- 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
- 238000000034 method Methods 0.000 title claims abstract description 200
- 238000004519 manufacturing process Methods 0.000 title claims abstract description 57
- 238000005457 optimization Methods 0.000 title claims abstract description 18
- 238000003860 storage Methods 0.000 title claims abstract description 9
- 230000008569 process Effects 0.000 claims abstract description 139
- 238000012544 monitoring process Methods 0.000 claims abstract description 58
- 230000001537 neural effect Effects 0.000 claims abstract description 14
- 238000012935 Averaging Methods 0.000 claims description 7
- 239000011159 matrix material Substances 0.000 claims description 6
- 230000008859 change Effects 0.000 claims description 3
- 238000013481 data capture Methods 0.000 claims description 3
- 238000012795 verification Methods 0.000 claims description 3
- 238000004590 computer program Methods 0.000 claims 1
- 230000000694 effects Effects 0.000 abstract description 7
- 230000006870 function Effects 0.000 description 8
- 238000012098 association analyses Methods 0.000 description 6
- 238000012545 processing Methods 0.000 description 6
- 238000004458 analytical method Methods 0.000 description 5
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 5
- 238000013461 design Methods 0.000 description 4
- 238000006243 chemical reaction Methods 0.000 description 3
- 239000003245 coal Substances 0.000 description 3
- 230000007423 decrease Effects 0.000 description 3
- 238000005065 mining Methods 0.000 description 3
- 230000005619 thermoelectricity Effects 0.000 description 3
- 241000208340 Araliaceae Species 0.000 description 2
- 235000005035 Panax pseudoginseng ssp. pseudoginseng Nutrition 0.000 description 2
- 235000003140 Panax quinquefolius Nutrition 0.000 description 2
- 230000008901 benefit Effects 0.000 description 2
- 230000033228 biological regulation Effects 0.000 description 2
- 238000010219 correlation analysis Methods 0.000 description 2
- 238000013480 data collection Methods 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 235000008434 ginseng Nutrition 0.000 description 2
- 230000010354 integration Effects 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000011160 research Methods 0.000 description 2
- 230000009471 action Effects 0.000 description 1
- QVGXLLKOCUKJST-UHFFFAOYSA-N atomic oxygen Chemical compound [O] QVGXLLKOCUKJST-UHFFFAOYSA-N 0.000 description 1
- 238000009412 basement excavation Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000010224 classification analysis Methods 0.000 description 1
- 238000002485 combustion reaction Methods 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 238000009833 condensation Methods 0.000 description 1
- 230000005494 condensation Effects 0.000 description 1
- 238000013500 data storage Methods 0.000 description 1
- 230000001934 delay Effects 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 235000013399 edible fruits Nutrition 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 239000007789 gas Substances 0.000 description 1
- 238000010438 heat treatment Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 230000002073 mitogenetic effect Effects 0.000 description 1
- 239000001301 oxygen Substances 0.000 description 1
- 229910052760 oxygen Inorganic materials 0.000 description 1
- 239000002245 particle Substances 0.000 description 1
- 230000001105 regulatory effect Effects 0.000 description 1
- 230000000630 rising effect Effects 0.000 description 1
- 238000005070 sampling Methods 0.000 description 1
- 239000000126 substance Substances 0.000 description 1
- 238000012549 training Methods 0.000 description 1
- 238000010977 unit operation Methods 0.000 description 1
- 238000009941 weaving Methods 0.000 description 1
- 238000005303 weighing Methods 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
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
-
- 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)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Strategic Management (AREA)
- Human Resources & Organizations (AREA)
- Economics (AREA)
- General Business, Economics & Management (AREA)
- Tourism & Hospitality (AREA)
- Marketing (AREA)
- Manufacturing & Machinery (AREA)
- Primary Health Care (AREA)
- Evolutionary Computation (AREA)
- Evolutionary Biology (AREA)
- Health & Medical Sciences (AREA)
- Computer Vision & Pattern Recognition (AREA)
- General Health & Medical Sciences (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- General Engineering & Computer Science (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Probability & Statistics with Applications (AREA)
- Development Economics (AREA)
- Game Theory and Decision Science (AREA)
- Entrepreneurship & Innovation (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The present invention relates to a kind of manufacturing parameter Optimization Prediction method, apparatus, equipment and storage mediums, including：Obtain the monitoring data of each process in production procedure；The monitoring data are pre-processed；The most strong association chain for indicating that two inter processes influence relationships is built in any two inter process using rule association algorithm, and most association chain is combined with the monitoring data fluctuation status by force by described in, obtains state relation chain；Prediction model is established according to the state relation chain using flexible neural tree algorithm, obtains and exports prediction result.This method can optimize the parameter of critical process according to prediction result, optimize the production procedure of coal-burning boiler by optimizing the parameter of critical process, achievees the effect that energy-saving and emission-reduction, improves economy and production security.
Description
Technical field
The invention belongs to coal-burning boiler production technical fields, and in particular to a kind of manufacturing parameter Optimization Prediction method, apparatus,
Equipment and storage medium.
Background technology
Boiler combustion process is exactly a large amount of thermal energy of a benefit conversion process, boiler high temperature water or steam generation, can
It directly uses, huge numbers of families is entered by pipeline, be applied in people’s lives production, ensure that life production develops in a healthy way,
Good heating is provided for house, air quality is adjusted comprehensively, for all trades and professions such as weaving, chemical industry, papermaking provides application, together
When, it by the burning of boiler, can also realize the conversion of electric energy, mechanical energy etc., realize economic good operation and development, it is seen that pot
Critical role and effect of the stove in production and living.
The parameter for grasping critical process in coal-burning boiler production procedure is particularly significant to the utilization of coal-burning boiler.In order to ensure
The reliability and economy of coal-burning boiler unit operation, there is an urgent need to grasp the parameter and efficiency of critical process in real time online, with
Just parameter effectively optimize and control according to actual conditions.
But the monitoring data that people directly obtain from coal-burning boiler are magnanimity, and usually there is a large amount of noises
Data and mistakes and omissions information, and the relationship of influencing each other of inter process can not directly embody in data, and have and be distributed, is different
Step, discrete characteristic cannot be used directly for big data processing.People be difficult directly obtained from magnanimity monitoring data it is valuable
Regular information, and then the parameter adjustments such as main vapour pressure, oxygen content, pressure fan rotating speed can not be optimized, also just it is unable to reach
Optimization production, energy-saving and emission-reduction improve the purpose of enterprise's production security.
Invention content
It is an object of the present invention to design a kind of manufacturing parameter optimization in view of the above-mentioned drawbacks of the prior art, providing
Prediction technique, device, equipment and storage medium, to solve the above technical problems.
In a first aspect, the embodiment of the present application provides a kind of manufacturing parameter Optimization Prediction method, including：It obtains in production procedure
The monitoring data of each process；The monitoring data are pre-processed；It is built in any two inter process using rule association algorithm
It indicates the most strong association chain of two inter processes influence relationships, and is most associated with chain and the monitoring data fluctuation status knot by force by described in
It closes, obtains state relation chain；Prediction model is established according to the state relation chain using flexible neural tree algorithm, obtains and exports
Prediction result.
With reference to first aspect, in the first embodiment of first aspect, include to monitoring data pretreatment：It will
Monitoring data in same time period merge, and obtain integrated monitor data；It is filled using arithmetic progression completion method or averaging method
Blank monitoring data；Sequential adjustment is carried out to the monitoring data for belonging to different processes using related coefficient curve；Utilize k- mean values
Algorithm obtains the cluster set of the monitoring data.
The first embodiment with reference to first aspect is closed in second of embodiment of first aspect using rule
Join algorithm and builds the most association chain, and the most strong association by described in by force for indicating that two inter processes influence relationships in any two inter process
Chain is combined with the monitoring data fluctuation status, is obtained state relation chain and is included：Utilize rule association algorithm process any two
The cluster data of process obtains the binomial correlation rule between arbitrary two clusters set in different processes；It calculates in different processes and appoints
Meaning two clusters the degree of association between set, and the degree of association between cluster set is converted to the degree of association of inter process；According to described
The correlation rule and the degree of association of two inter processes of meaning, selection meets time series and the maximum cluster data of the degree of association is closed
Connection, construct different inter processes is most associated with by force chain；By the monitoring data fluctuation status of different inter processes and the most strong association chain
In conjunction with obtaining the state relation chain of different inter processes.
Second of embodiment with reference to first aspect utilizes flexible god in the third embodiment of first aspect
It prediction model established according to the state relation chain through tree algorithm obtains and exports prediction result and include：Utilize flexible Neural Tree side
Method establishes prediction model according to the state relation chain and pretreated monitoring data；The production of critical process is inputted into ginseng
Number and production output parameter are input to the prediction model and are iterated operation；It is obtained most preferably by constantly changing iterations
Iterations, and utilize the accuracy of standard mean square difference and matrix labotstory verification prediction.
Second aspect, the embodiment of the present application provide a kind of manufacturing parameter Optimization Prediction device, including：Data capture unit,
It is configured to obtain the monitoring data of each process in production procedure；Pretreatment unit is configured to pre- to the monitoring data
Processing；It is associated with chain building unit, is configured to build two processes of expression in any two inter process using rule association algorithm
Between influence the most association chain by force of relationship, and described will most be associated with chain by force and be combined with the monitoring data fluctuation status, and obtain state
It is associated with chain；Model foundation unit is configured to establish prediction model according to the state relation chain using flexible neural tree algorithm,
It obtains and exports prediction result.
In conjunction with second aspect, in the first embodiment of second aspect, pretreatment unit includes：Merge subelement,
The monitoring data for being configured to be in same time period merge, and obtain integrated monitor data；Subelement is filled, profit is configured to
It is filled in the blanks monitoring data with arithmetic progression completion method or averaging method；Subelement is adjusted, is configured to utilize related coefficient curve
Monitoring data to belonging to different processes carry out sequential adjustment；Subelement is clustered, is configured to obtain institute using k- mean algorithms
State the cluster set of monitoring data.
In conjunction with the first embodiment of second aspect, in second of embodiment of second aspect, it is associated with chain building
Unit includes：Correlation rule subelement is configured to the cluster data using rule association algorithm process any two process, obtains
Binomial correlation rule between arbitrary two clusters set in different processes；Degree of association subelement is configured to calculate different processes
In the degree of association between arbitrary two clusters set, and by cluster gather between the degree of association be converted to the degree of association of inter process；It is most strong to close
Join chain subelement, be configured to the correlation rule and the degree of association according to any two inter process, selection meets time series
And the maximum cluster data of the degree of association is associated, construct different inter processes is most associated with by force chain；State relation chain subelement, matches
It sets for by the monitoring data fluctuation status of different inter processes and the most strong association chain combination, obtaining the state of different inter processes
It is associated with chain.
In conjunction with second of embodiment of second aspect, in the third embodiment of second aspect, model foundation list
Member includes：Subelement is modeled, is configured to using flexible neural tree method, according to the state relation chain and pretreated prison
Measured data establishes prediction model；Operation subelement is configured to the production input parameter of critical process and production output parameter
It is input to the prediction model and is iterated operation；Subelement is verified, is configured to constantly to change iterations and obtains
Best iterations, and utilize the accuracy of standard mean square difference and matrix labotstory verification prediction.
The third aspect, the embodiment of the present application also provide a kind of equipment, including：One or more processors；Memory is used for
The one or more programs of storage, when one or more of programs are executed by one or more of processors so that described
One or more processors execute the method as described in first aspect and first aspect any embodiment.
Fourth aspect, the embodiment of the present application provide a kind of storage medium, and such as first is realized when which is executed by processor
Method described in aspect and first aspect any embodiment.
The beneficial effects of the present invention are,
The application first pre-processes the monitoring data of magnanimity, it is made to can be used for big data processing.And it further utilizes
Rule association algorithm finds out the incidence relation of different inter processes, and establishes prediction mould using flexible Neural Tree method on this basis
Type, to be optimized to the parameter of critical process according to prediction result.This method is by optimizing the parameter of critical process
The production procedure for optimizing coal-burning boiler achievees the effect that energy-saving and emission-reduction, improves economy and production security.
In addition, design principle of the present invention is reliable, and it is simple in structure, there is very extensive application prospect.
It can be seen that compared with prior art, the present invention with substantive distinguishing features outstanding and significantly improving, implementation
Advantageous effect be also obvious.
Description of the drawings
By reading a detailed description of non-restrictive embodiments in the light of the attached drawings below, the application's is other
Feature, objects and advantages will become more apparent upon：
Fig. 1 is the manufacturing parameter Optimization Prediction method flow in coal-burning boiler production procedure provided by the embodiments of the present application
Figure；
Fig. 2 is that a bar state provided by the embodiments of the present application is associated with chain result analysis chart；
Fig. 3 is modeling and forecasting design sketch provided by the embodiments of the present application.
Specific implementation mode
The application is described in further detail with reference to the accompanying drawings and examples.It is understood that this place is retouched
The specific embodiment stated is used only for explaining related invention, rather than the restriction to the invention.It also should be noted that in order to
Convenient for description, is illustrated only in attached drawing and invent relevant part.
It should be noted that in the absence of conflict, the features in the embodiments and the embodiments of the present application can phase
Mutually combination.The application is described in detail below with reference to the accompanying drawings and in conjunction with the embodiments.
The present embodiment provides a kind of manufacturing parameter Optimization Prediction methods, including the following contents：
The acquisition of S1, monitoring data：Has the characteristics that typical flow object in the monitoring data of thermoelectricity production field, entirely
Production technology includes multiple front and back concerned process steps or process, all deploys OPC data acquisition interface in these processes, can incite somebody to action
In real-time detector data storage to database；And different production systems, such as hot-water boiler change water system, auxiliary system, deposit respectively
It in different subsystems, does not interfere, and is finally integrated into total Database between each other, these collected data are for we
Forecast analysis provide the foundation.
The pretreatment of S2, monitoring data：
(1) data integration and data filling,
Data integration：The problem of disperseing for data, data are integrated, chooses all process steps and be in same time piece
The data of section, merging becomes new data.The data obtained at this time contain the information of a large amount of processing flow sequences, are carried out to data
Accurate result can be obtained when processing flow sequence mode excavation.
Control filling：
For the data of a large amount of blank, directly given up to fall；For the data of a small amount of blank, there is the use of visible trend
Arithmetic progression completion method, the use averaging method for not having visible trend are filled.By being filled to clear data, to protect
Demonstrate,prove the stability of data.Prevent influence of the null value to result in later stage modeling and forecasting.In flow object χ, if a certain process
XiContaining excalation value, the data of missing time section are { Xi(tm),Xi(tn) between data, then the missing data contains altogether
There are n-m-1, j-th of missing values in this section are
If the data do not have tendency feature, filled up using averaging method, the data of filling up in this section are
Wherein, Xi(tj) it is Xi(tm) before n-m data, Xi(tk) it is Xi(tn) after n-m data；If vacancy
Value then directly deletes it in the above or below of data；If Xi(tm) front or Xi(tn) subsequent data less than n-m, then from
Xi(tn) below or Xi(tm) front is postponed, and to choose total amount be 2 (n-m) a data, the mean value ensured contains enough information,
And select data volume excessive, it will increase computation complexity；Data volume is very few, and is not enough to include the information of data.
(2) sequential adjusts,
We regard the sampled data of each process as time series (TimeSeries) data, final purpose are
The relationship that influences each other between different processes is calculated, this requires the data between process are corresponded according to influence relationship
's.However, every record in database includes the sampled data of synchronization difference process, this is not to be closed according to influence
It is corresponding, this just needs us according to the relative delay of process to be adjusted, namely according to relative delay pair
Process data carry out Forward or rear shifting.
The data of each process are the time series of a constant duration, namely are arranged according to chronological order
The sampling value set of row.Influence between different processes constitutes between its corresponding two time series
Two time serieses of delay correlation pair, which carry out delay correlation analysis, can acquire the sequential Pearson of two-step
Related coefficient is used for weighing the linearly related degree of two sequences, and definition is as shown in formula (3).
Wherein,WithIt is the average value of sequence sum respectively.The absolute value of related coefficient is bigger, and correlation is stronger；Phase relation
For number closer to 0, correlation is weaker.It is made of Pearson correlation coefficient of two time serieses under the different delays time
Curve, we term it Pearson correlation coefficient curves.Related coefficient curve illustrates two time serieses with delay time
Increase, the variation tendency of correlation.The extreme point of curve namely the point of correlation maximum, usually characterize two time serieses
It is influenced according to the delay time.Prolonging based on Pearson correlation coefficient curve is carried out to the time series data of two-step
Slow correlation analysis draws related coefficient curve, prolongs the corresponding length of delay of related coefficient maximum value as the opposite of two-step
The slow time.To be adjusted to process.
(3) process clusters
Numerous discrete data states will greatly increase the time complexity of subsequent association analysis, and clustering can help
We reduce the discrete state number of data, us is made only can targetedly to represent almost institute with K a small number of classifications
There are data.We are clustered using K-Means algorithms, and cluster is individually carried out for each process namely each process
All there are one cluster number K and K corresponding types, and the cluster number K of different processes may be different.K when for cluster
The determination of value, we use is carried out based on the silhouette coefficient method of condensation degree and separating degree.Each process data are clustered
Later, it will obtain K classification, final we will replace all process data, subsequent association point using this K classification
Analysis will be based on cluster about subsequent process data and carry out.The paralell design of the K-Means algorithms based on data parallel is given below：
Input：The time series data of a certain process, cluster number K
Output：The cluster result of the process data
1) K cluster centre of random initializtion, and it is broadcasted each calculate node；
2) for each data of each subregion, they are referred to cluster centre institute's generation away from nearest neighbours one by one
The cluster classification of table；
3) the sum of the sum of data of same cluster classification and number are calculated for the data of each subregion, and by it
Be collected into driver nodes, the cluster centre of next iteration is recalculated in driver nodes；
4) if cluster centre no longer changes or reached specified iterations, algorithm performs terminate；Otherwise, will
New cluster centre is broadcast to each calculate node, then turns to step 2).
K-means algorithms control iterations using the minimization of object function criterion, for numeric type data, usually adopt
Use Euclidean distance (Euclidean Distance) as object function.Equipped with two data object X1And X2, EUCLID (X1,
X2) it is X1And X2Euclidean distance, E be all objects square error summation, each object contain t variable, then two
Euclidean distance between object is square root sum square of the difference of t variate-value of two objects, i.e.,
Object function form is square error criterion function
Wherein, CiFor arbitrary cluster set, X CiIn object, miFor class CiMean value.Distance metric in object function
Other modes, such as manhatton distance, Chebyshev's distance, Hamming distance, power distance and mahalanobis distance etc. can also be used.It is false
If arbitrary process XiPreferable clustering number mesh be Ki, the cluster set that is obtained after being clustered according to the preferable clustering number mesh.
S3, the discovery for being associated with chain：
The characteristics of by analyzing Apriori Rule Extraction processes, obtains state using it and closes
Join the algorithm of chain.Under the conditions of meeting technological process, using association rule algorithm between the dimension based on Apriori, obtain
To correlation rule between the class of any two process.By rule between any two, numerous associations is obtained.
1) correlation rule between class：According to the demand of flow object, minimum support min_sup, min confidence are set
Min_conf is closed carrying out the dimension based on Apriori (rule association algorithm) to the cluster data of the arbitrary two-step of cluster set
Join rule digging, search for frequent 2- predicates collection, generates the binomial correlation rule between different arbitrary two clusters of process, these rule lists
Show that two of different processes meet the relationship between minimum support and the cluster of min confidence.In boiler parameter, we
Carry out materialization description, preceding paragraph XiConsequent is XjAll cluster data collection common properties give birth to 2 cluster association rules, preceding paragraph Xj
Consequent is XiAll cluster data collection common properties give birth to 1 cluster association rule.
2) process is associated with：If any two process XiAnd XjCluster between common property give birth to ωijRule, wherein any one rules and regulations
It is then ia → jb.Since interest-degree I (ia → jb) sizes are different to the meaning of correlation rule, signed magnitude is used to calculate herein
Preceding paragraph is XiWith consequent for XjCluster between degree of association cij(β), then
cij(β)=Sβ(ia→jb)×I'β(ia→jb),β≤ωij (6)
Wherein, β indicates the β rules in strictly all rules, and
I'β(ia → jb)=Iβ(ia→jb)-1 (7)
By conversion, the association between process is clustered is converted to the association of inter process, obtains the associated relationship of process.
3) it is most associated with chain and state relation chain by force：Any two inter process correlation rule and its association are known at present
Degree, according to the relationship between item before and after the data, selection meets time series L={ X '1,X'2,…,X'nAnd degree of association maximum
Data be associated, a most association chain by force can be constructed, when which illustrates in flow object that all process steps are run on chain
A kind of influence relationship.Finally obtain most strong association chain all in historical data.In order to preferably utilize above-mentioned association analysis
As a result, and make it easier to be used in actual production with aid decision, by the state of process, (for example monitoring data is upper for we
Liter, decline or invariant state) further it is combined with incidence relation, show that the state relation relationship between process, i.e. state are closed
Join chain.
Association analysis：Association analysis (Association Analysis) includes mainly correlation rule (Association
Rule it) generates and is associated with chain/tree (Association Chain/Tree) generation.In association rule mining, we just for
Correlation rule is generated between two two-steps, and association chain/tree is finally generated by the correlation rule of two two-steps.In process cluster numbers
According to upper 2 correlation rules, and the rule generated according to ordered pair when process are found out using Apriori classical correlations rule mining algorithms
Then it is filtered.
State relation：In order to preferably using above-mentioned association analysis as a result, and making it easier to be used in actual production
With aid decision, we by the state (such as up and down or invariant state of detection data) of process further be associated with
System is combined, and obtains the state relation relationship between process, we term it operating status knowledge bases.Operating status knowledge base can
For in the industry auxiliary such as fault detect and early warning, parameter regulation.State relation can use the method counted to obtain, more
Simply.
In the research of this method, it can be found from mass data and the close phase of optimization aim using association rule algorithm
The process of pass is the committed step of system research.But process is numerous in thermoelectricity production, the attribute for participating in calculating is very more, profit
It is too low with traditional association rules mining algorithm acquisition correlation rule efficiency, it calculates knowledge that is complicated and obtaining and is not easy to directly quilt
User understands, therefore it is proposed that the influence relationship of each inter process in producing, pass is indicated in the way of state relation chain
The representation method of connection chain is embodied in boiler parameter, and such as No. 1 furnace gas packet pressure decline leads to No. 1 stove main vapour pressure lower-left drop, into
And lead to No. 1 burner hearth differential pressure lower-left drop, lead to No. 1 stove feed pressure upper left liter.It, can be big by being associated with chain as acquisition
Key node is found in amount production process to be regulated and controled, and the result of target process is quickly directly affected.It is with hot water boiler system
Example is contributed by being associated with link analysis in order to improve boiler hot, can be by increasing primary air fan power, reducing overfire air fan gear
Plate aperture keeps that the amount of delivering coal is constant reaches target.State relation chain shows " rising ", " decline " and " constant " three kinds of states,
And intuitively show very much the front-to-back effect relationship between multiple working procedure states.
By obtaining state relation chain, the strongest most strong association chain of relevance is found out, by with the modeling of next stage, to
The more intuitively relationship between display and adjustment each process.
The foundation of S4, prediction model：
This part primarily directed to working obtained state relation chain and processed data are modeled above.With
FNT (flexible Neural Tree) method obtains modeling and forecasting formula, realizes the prediction to process data.
1) database table that modeling is used has, and by pretreated tables of data, FNT user's table, FNT operating parameters table, surveys
Try tables of data, training data table, association chained list, output result table etc..It is carried out by the result obtained to the processing of data before
Interative computation.
2) for coal-burning boiler parameter, crucial flow and corresponding production input and output parameter are found, by it
It is input in FNT, obtains the output function shaped like formula (8), An and Bn are the ginsengs after two groups of PSO (particle cluster algorithm) optimization
Number, netn are the tree structures after PIPE (pipeline is used for the communication between affiliation process) optimization；Pass through the letter
Several variations to future production carry out forecast analysis.
3) by running FNT algorithms.Constantly modification iterations, obtain best iterations.In returning the result
Standard average variance (NMSE) value, to determine trained effect accuracy.By obtained formula, with matlab (matrix experiments
Room) preliminary figure forecast analysis is carried out, by comparing the final accuracy for determining prediction.
The effect of modeling and forecasting is to carry out science to the association attributes for influencing target process on the basis of state relation chain
It calculates, obtains the variation tendency formula of measuring point data, industrial processes are simulated, the correlation based on current production status
The production status of parameter prediction a period of time in future, to which auxiliary direction enterprise adjusts production procedure parameter, the step for be real
The now key to the application of auxiliary production process prediction adjustment and the most important content of Project-developing.By to second
It is mitogenetic at state relation chain modeled, will be associated with chain in the other processes in addition to last as the defeated of anticipation function
Enter value, and target process, by changing the value of other processes, can predict target process in a timing as function-output
Between after numerical value, so as to be optimized in advance to production procedure.It is embodied in thermoelectricity production process, equally with hot-water boiler
For, previous step has obtained association chain：Primary air fan power rise → overfire air fan baffle opening declines → and the amount of delivering coal is constant
→ heat, which is contributed, to be improved, and in modeling and forecasting, it is function-output, primary air fan power, overfire air fan baffle opening, pot that heat, which is contributed,
The stove amount of delivering coal is function input value, then can obtain the prediction technique as shown in formula (9).
The present embodiment provides a kind of manufacturing parameter Optimization Prediction devices, including：Data capture unit is configured to obtain life
Produce the monitoring data of each process in flow；Pretreatment unit is configured to pre-process the monitoring data；It is associated with chain building
Unit is configured to indicate that two inter processes influence the most strong of relationship in any two inter process structure using rule association algorithm
It is associated with chain, and most association chain is combined with the monitoring data fluctuation status by force by described in, obtains state relation chain；Model foundation list
Member is configured to establish prediction model according to the state relation chain using flexible neural tree algorithm, obtains and export prediction knot
Fruit.
Wherein, pretreatment unit includes：Merge subelement, the monitoring data for being configured to be in same time period are closed
And obtain integrated monitor data；Subelement is filled, is configured to fill in the blanks monitoring using arithmetic progression completion method or averaging method
Data；Subelement is adjusted, is configured to carry out sequential adjustment to the monitoring data for belonging to different processes using related coefficient curve；
Subelement is clustered, is configured to obtain the cluster set of the monitoring data using k- mean algorithms.
Being associated with chain building unit includes：Correlation rule subelement is configured to utilize rule association algorithm process arbitrary two
The cluster data of a process obtains the binomial correlation rule between arbitrary two clusters set in different processes；Degree of association subelement, matches
It sets for calculating the degree of association in different processes between arbitrary two clusters set, and the degree of association between cluster set is converted into process
Between the degree of association；It is most associated with chain subelement by force, is configured to the correlation rule and the degree of association according to any two inter process,
Selection meets time series and the maximum cluster data of the degree of association is associated, and construct different inter processes is most associated with by force chain；Shape
State is associated with chain subelement, is configured to, by the monitoring data fluctuation status of different inter processes and the most strong association chain combination, obtain
To the state relation chain of different inter processes.
Model foundation unit includes：Subelement is modeled, is configured to, using flexible neural tree method, be closed according to the state
Connection chain and pretreated monitoring data establish prediction model；Operation subelement is configured to input the production of critical process
Parameter and production output parameter are input to the prediction model and are iterated operation；Subelement is verified, is configured to not
Disconnected modification iterations obtain best iterations, and verify the accurate of prediction using standard mean square difference and matrix labotstory
Property.
The present embodiment also provides a kind of equipment, including：One or more processors；Memory, for storing one or more
A program, when one or more of programs are executed by one or more of processors so that one or more of places
The method that reason device executes manufacturing parameter Optimization Prediction provided by the present application.
Fourth aspect, the embodiment of the present application provide a kind of storage medium, which realizes the application when being executed by processor
The method of the manufacturing parameter Optimization Prediction of offer.
Disclosed above is only the preferred embodiment of the present invention, but the present invention is not limited to this, any this field
What technical staff can think does not have a creative variation, and without departing from the principles of the present invention made by several improvement and
Retouching, should all be within the scope of the present invention.
Claims (10)
1. a kind of manufacturing parameter Optimization Prediction method, which is characterized in that the method includes：
Obtain the monitoring data of each process in production procedure；
The monitoring data are pre-processed；
Indicate that two inter processes influence the most strong association chain of relationship in any two inter process structure using rule association algorithm, and
Most association chain is combined with the monitoring data fluctuation status by force by described in, obtains state relation chain；
Prediction model is established according to the state relation chain using flexible neural tree algorithm, obtains and exports prediction result.
2. according to the method described in claim 1, it is characterized in that, described include to monitoring data pretreatment：
Monitoring data in same time period are merged, integrated monitor data are obtained；
It is filled in the blanks monitoring data using arithmetic progression completion method or averaging method；
Sequential adjustment is carried out to the monitoring data for belonging to different processes using related coefficient curve；
The cluster set of the monitoring data is obtained using k- mean algorithms.
3. according to the method described in claim 2, it is characterized in that, described utilize rule association algorithm in any two inter process
The most strong association chain for indicating two inter processes influence relationships is built, and is most associated with chain and the monitoring data undulating by force by described in
State combines, and obtaining state relation chain includes：
Using the cluster data of rule association algorithm process any two process, obtain in different processes between arbitrary two clusters set
Binomial correlation rule；
The degree of association between arbitrary two clusters set in different processes is calculated, and the degree of association between cluster set is converted into inter process
The degree of association；
According to the correlation rule and the degree of association of any two inter process, selection meets time series and the degree of association is maximum poly-
Class data are associated, and construct different inter processes is most associated with by force chain；
By the monitoring data fluctuation status of different inter processes and the most strong association chain combination, the state for obtaining different inter processes is closed
Join chain.
4. according to the method described in claim 3, it is characterized in that, described closed using flexible neural tree algorithm according to the state
Connection chain, which establishes prediction model and obtains and export prediction result, includes：
Using flexible neural tree method, prediction model is established according to the state relation chain and pretreated monitoring data；
The production input parameter of critical process and production output parameter are input to the prediction model and are iterated operation；
Best iterations are obtained by constantly changing iterations, and are verified using standard mean square difference and matrix labotstory
The accuracy of prediction.
5. a kind of manufacturing parameter Optimization Prediction device, which is characterized in that described device includes：
Data capture unit is configured to obtain the monitoring data of each process in production procedure；
Pretreatment unit is configured to pre-process the monitoring data；
It is associated with chain building unit, is configured to build two inter processes of expression in any two inter process using rule association algorithm
Influence relationship is most associated with by force chain, and most association chain is combined with the monitoring data fluctuation status by force by described in, obtains state pass
Join chain；
Model foundation unit is configured to establish prediction model according to the state relation chain using flexible neural tree algorithm, obtain
To and export prediction result.
6. device according to claim 5, which is characterized in that the pretreatment unit includes：
Merge subelement, the monitoring data for being configured to be in same time period merge, and obtain integrated monitor data；
Subelement is filled, is configured to fill in the blanks monitoring data using arithmetic progression completion method or averaging method；
Subelement is adjusted, is configured to carry out sequential adjustment to the monitoring data for belonging to different processes using related coefficient curve；
Subelement is clustered, is configured to obtain the cluster set of the monitoring data using k- mean algorithms.
7. device according to claim 6, which is characterized in that the association chain building unit includes：
Correlation rule subelement is configured to the cluster data using rule association algorithm process any two process, obtains not
With the binomial correlation rule between arbitrary two clusters set in process；
Degree of association subelement is configured to calculate the degree of association between arbitrary two clusters set in different processes, and cluster is gathered
Between the degree of association be converted to the degree of association of inter process；
It is most associated with chain subelement by force, is configured to the correlation rule and the degree of association according to any two inter process, chooses full
The sufficient time series and maximum cluster data of the degree of association is associated constructs the most association chain by force of different inter processes；
State relation chain subelement is configured to chain the monitoring data fluctuation status of different inter processes and the most strong association
It closes, obtains the state relation chain of different inter processes.
8. device according to claim 7, which is characterized in that the model foundation unit includes：
Subelement is modeled, is configured to using flexible neural tree method, according to the state relation chain and pretreated monitoring
Data establish prediction model；
Operation subelement is configured to the production input parameter of critical process and production output parameter being input to the prediction mould
Type is simultaneously iterated operation；
Subelement is verified, is configured to constantly to change iterations and obtains best iterations, and utilization standard mean square
The accuracy of difference and matrix labotstory verification prediction.
9. a kind of equipment, which is characterized in that the equipment includes：
One or more processors；
Memory, for storing one or more programs,
When one or more of programs are executed by one or more of processors so that one or more of processors
Execute the method as described in any one of claim 1-4.
10. a kind of computer readable storage medium being stored with computer program, which is characterized in that the program is executed by processor
Methods of the Shi Shixian as described in any one of claim 1-4.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810322649.3A CN108647808B (en) | 2018-04-11 | 2018-04-11 | Production parameter optimization prediction method, device, equipment and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810322649.3A CN108647808B (en) | 2018-04-11 | 2018-04-11 | Production parameter optimization prediction method, device, equipment and storage medium |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN108647808A true CN108647808A (en) | 2018-10-12 |
| CN108647808B CN108647808B (en) | 2022-03-29 |
Family
ID=63746129
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810322649.3A Active CN108647808B (en) | 2018-04-11 | 2018-04-11 | Production parameter optimization prediction method, device, equipment and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108647808B (en) |
Cited By (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109726503A (en) * | 2019-01-12 | 2019-05-07 | 国电联合动力技术有限公司 | Missing data complementing method and device |
| CN109884893A (en) * | 2019-02-28 | 2019-06-14 | 西安理工大学 | Dynamic lag estimation method between a kind of multi-process variable |
| CN110135740A (en) * | 2019-05-20 | 2019-08-16 | 济南大学 | Real time knowledge towards coal-burning boiler flow object finds method and system |
| CN111125082A (en) * | 2019-12-26 | 2020-05-08 | 北京工业大学 | Cement process parameter data analysis method based on association rule mining |
| CN111679634A (en) * | 2020-01-20 | 2020-09-18 | 武汉裕大华纺织有限公司 | Intelligent roving management system |
| CN112990711A (en) * | 2021-03-19 | 2021-06-18 | 云南建投第九建设有限公司 | Aluminum alloy formwork construction monitoring method and system based on site construction |
| CN113063455A (en) * | 2021-03-15 | 2021-07-02 | 上海联影医疗科技股份有限公司 | Detector parameter configuration method, equipment, electronic device and storage medium |
| CN113723723A (en) * | 2020-05-25 | 2021-11-30 | 中国石油化工股份有限公司 | Operation parameter fluctuation path extraction method and device, storage medium and processor |
| CN113722403A (en) * | 2020-05-25 | 2021-11-30 | 中国石油化工股份有限公司 | Abnormal operation data clustering method and device, storage medium and processor |
| CN114896228A (en) * | 2022-04-27 | 2022-08-12 | 西北工业大学 | Industrial data stream cleaning model and method based on multi-stage combination optimization of filtering rules |
| CN116305671A (en) * | 2023-05-23 | 2023-06-23 | 山东伟国板业科技有限公司 | Method and system for monitoring production line of artificial board |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20140143190A1 (en) * | 2012-11-20 | 2014-05-22 | Qualcomm Incorporated | Piecewise linear neuron modeling |
| CN105096008A (en) * | 2015-08-28 | 2015-11-25 | 浙江大学 | Control method of process industry production system |
| US20170004412A1 (en) * | 2015-07-02 | 2017-01-05 | PRA Health Sciences, Inc. | Normalizing Data Sets for Predicting an Attribute of the Data Sets |
| CN106445788A (en) * | 2016-09-30 | 2017-02-22 | 国家电网公司 | Method and device for predicting operating state of information system |
| CN107247995A (en) * | 2016-09-29 | 2017-10-13 | 上海交通大学 | Transmission line of electricity running status association rule mining and Forecasting Methodology based on Bayesian model |
| CN104346442B (en) * | 2014-10-14 | 2017-10-20 | 济南大学 | A kind of Rules extraction method of Process-Oriented object data |
-
2018
- 2018-04-11 CN CN201810322649.3A patent/CN108647808B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20140143190A1 (en) * | 2012-11-20 | 2014-05-22 | Qualcomm Incorporated | Piecewise linear neuron modeling |
| CN104346442B (en) * | 2014-10-14 | 2017-10-20 | 济南大学 | A kind of Rules extraction method of Process-Oriented object data |
| US20170004412A1 (en) * | 2015-07-02 | 2017-01-05 | PRA Health Sciences, Inc. | Normalizing Data Sets for Predicting an Attribute of the Data Sets |
| CN105096008A (en) * | 2015-08-28 | 2015-11-25 | 浙江大学 | Control method of process industry production system |
| CN107247995A (en) * | 2016-09-29 | 2017-10-13 | 上海交通大学 | Transmission line of electricity running status association rule mining and Forecasting Methodology based on Bayesian model |
| CN106445788A (en) * | 2016-09-30 | 2017-02-22 | 国家电网公司 | Method and device for predicting operating state of information system |
Non-Patent Citations (2)
| Title |
|---|
| 付爱芳: "流程对象建模方法的研究与实现", 《中国优秀硕士学位论文全文数据库工程科技Ⅰ辑》 * |
| 宋巧云: "基于电力系统大数据集的知识发现方法的研究与实现", 《中国优秀硕士学位论文全文数据库信息科技辑》 * |
Cited By (16)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109726503A (en) * | 2019-01-12 | 2019-05-07 | 国电联合动力技术有限公司 | Missing data complementing method and device |
| CN109726503B (en) * | 2019-01-12 | 2020-12-18 | 国电联合动力技术有限公司 | Missing data filling method and device |
| CN109884893A (en) * | 2019-02-28 | 2019-06-14 | 西安理工大学 | Dynamic lag estimation method between a kind of multi-process variable |
| CN110135740A (en) * | 2019-05-20 | 2019-08-16 | 济南大学 | Real time knowledge towards coal-burning boiler flow object finds method and system |
| CN111125082B (en) * | 2019-12-26 | 2023-09-22 | 北京工业大学 | Cement process parameter data analysis method based on association rule mining |
| CN111125082A (en) * | 2019-12-26 | 2020-05-08 | 北京工业大学 | Cement process parameter data analysis method based on association rule mining |
| CN111679634A (en) * | 2020-01-20 | 2020-09-18 | 武汉裕大华纺织有限公司 | Intelligent roving management system |
| CN113723723A (en) * | 2020-05-25 | 2021-11-30 | 中国石油化工股份有限公司 | Operation parameter fluctuation path extraction method and device, storage medium and processor |
| CN113722403A (en) * | 2020-05-25 | 2021-11-30 | 中国石油化工股份有限公司 | Abnormal operation data clustering method and device, storage medium and processor |
| CN113063455A (en) * | 2021-03-15 | 2021-07-02 | 上海联影医疗科技股份有限公司 | Detector parameter configuration method, equipment, electronic device and storage medium |
| CN112990711A (en) * | 2021-03-19 | 2021-06-18 | 云南建投第九建设有限公司 | Aluminum alloy formwork construction monitoring method and system based on site construction |
| CN112990711B (en) * | 2021-03-19 | 2023-11-17 | 云南建投第九建设有限公司 | Aluminum alloy template construction monitoring method and system based on site construction |
| CN114896228A (en) * | 2022-04-27 | 2022-08-12 | 西北工业大学 | Industrial data stream cleaning model and method based on multi-stage combination optimization of filtering rules |
| CN114896228B (en) * | 2022-04-27 | 2024-04-05 | 西北工业大学 | Industrial data stream cleaning model and method based on filtering rule multistage combination optimization |
| CN116305671A (en) * | 2023-05-23 | 2023-06-23 | 山东伟国板业科技有限公司 | Method and system for monitoring production line of artificial board |
| CN116305671B (en) * | 2023-05-23 | 2023-10-20 | 山东伟国板业科技有限公司 | Method and system for monitoring production line of artificial board |
Also Published As
| Publication number | Publication date |
|---|---|
| CN108647808B (en) | 2022-03-29 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108647808A (en) | A kind of manufacturing parameter Optimization Prediction method, apparatus, equipment and storage medium | |
| CN106709662B (en) | Power equipment operation condition division method | |
| CN106094744B (en) | Based on the determination method of thermoelectricity factory owner's operating parameter desired value of association rule mining | |
| Lu et al. | Performance predictions of ground source heat pump system based on random forest and back propagation neural network models | |
| CN110532674A (en) | A kind of coal-fired power station boiler fire box temperature measurement method | |
| WO2023024433A1 (en) | Gas-steam combined cycle generator set operation adjustment and control system, and adjustment and control method | |
| CN111275367A (en) | Regional comprehensive energy system energy efficiency state evaluation method | |
| Yang et al. | An efficient evolutionary approach to parameter identification in a building thermal model | |
| CN113837464A (en) | Load prediction method of cogeneration boiler based on CNN-LSTM-Attention | |
| CN114218292A (en) | Multi-element time sequence similarity retrieval method | |
| CN110516944A (en) | A kind of power distribution network multistage typical case's Run-time scenario generation method | |
| CN113221467B (en) | Turbine gas-thermal performance uncertainty visual analysis method and system | |
| CN110207094A (en) | IQGA-SVR boiler heating surface fouling characteristics discrimination method based on principal component analysis | |
| CN112152840A (en) | Sensor deployment method and system based on BIM and analog simulation | |
| CN109919401B (en) | Multi-dimensional energy efficiency analysis method of multi-energy complementary system | |
| CN112270449B (en) | Industrial system time delay determination and controlled quantity prediction method based on time correlation | |
| Lingqing et al. | Detection method for power theft based on SOM neural network and K-means clustering algorithm | |
| CN111582588A (en) | Building energy consumption prediction method based on triple convolution fusion GRU | |
| CN110135740A (en) | Real time knowledge towards coal-burning boiler flow object finds method and system | |
| Antonio et al. | Feature Selection Technique Impact for Internet Traffic Classification Using Naï ve Bayesian | |
| CN109931709A (en) | Oil field heating furnace energy conservation adjusting method and system | |
| CN109388884A (en) | A kind of generalized regression extreme value response phase method calculating coupling leaf dish fatigue life | |
| CN115390448A (en) | Visual analysis method and system for control strategy of coal-fired power plant | |
| CN114970766A (en) | Power station unit operation parameter reference value obtaining method based on linear fitting | |
| CN114565209A (en) | Process industry energy consumption state evaluation method based on clustering |
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 |