CN108647808B - Production parameter optimization prediction method, device, equipment and storage medium - Google Patents
Production parameter optimization prediction method, device, equipment and storage medium Download PDFInfo
- Publication number
- CN108647808B CN108647808B CN201810322649.3A CN201810322649A CN108647808B CN 108647808 B CN108647808 B CN 108647808B CN 201810322649 A CN201810322649 A CN 201810322649A CN 108647808 B CN108647808 B CN 108647808B
- Authority
- CN
- China
- Prior art keywords
- association
- monitoring data
- chain
- data
- state
- 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.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 195
- 238000004519 manufacturing process Methods 0.000 title claims abstract description 58
- 238000003860 storage Methods 0.000 title claims abstract description 8
- 238000005457 optimization Methods 0.000 title claims description 15
- 230000008569 process Effects 0.000 claims abstract description 99
- 238000012544 monitoring process Methods 0.000 claims abstract description 64
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 38
- 230000001537 neural effect Effects 0.000 claims abstract description 15
- 238000007781 pre-processing Methods 0.000 claims abstract description 12
- 238000012935 Averaging Methods 0.000 claims description 7
- 239000011159 matrix material Substances 0.000 claims description 6
- 238000012545 processing Methods 0.000 claims description 5
- 238000013459 approach Methods 0.000 claims description 4
- 238000004891 communication Methods 0.000 claims description 3
- 238000010276 construction Methods 0.000 claims description 3
- 239000002245 particle Substances 0.000 claims description 3
- 238000004590 computer program Methods 0.000 claims 1
- 230000000694 effects Effects 0.000 abstract description 3
- 230000006870 function Effects 0.000 description 13
- 238000010219 correlation analysis Methods 0.000 description 7
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 6
- 238000004458 analytical method Methods 0.000 description 5
- 238000012098 association analyses Methods 0.000 description 4
- 238000005065 mining Methods 0.000 description 4
- 230000008901 benefit Effects 0.000 description 3
- 238000004364 calculation method Methods 0.000 description 3
- 238000006243 chemical reaction Methods 0.000 description 3
- 239000003245 coal Substances 0.000 description 3
- 238000005070 sampling Methods 0.000 description 3
- 239000000126 substance Substances 0.000 description 3
- 230000009286 beneficial effect Effects 0.000 description 2
- 230000008859 change Effects 0.000 description 2
- 238000002485 combustion reaction Methods 0.000 description 2
- 238000013461 design Methods 0.000 description 2
- 238000001514 detection method Methods 0.000 description 2
- 238000011161 development Methods 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 239000006185 dispersion Substances 0.000 description 2
- 230000010354 integration Effects 0.000 description 2
- 238000005192 partition Methods 0.000 description 2
- 230000009467 reduction Effects 0.000 description 2
- 238000011160 research Methods 0.000 description 2
- 230000000630 rising effect Effects 0.000 description 2
- 238000012549 training Methods 0.000 description 2
- 238000005054 agglomeration Methods 0.000 description 1
- 230000002776 aggregation Effects 0.000 description 1
- 230000001174 ascending effect Effects 0.000 description 1
- QVGXLLKOCUKJST-UHFFFAOYSA-N atomic oxygen Chemical compound [O] QVGXLLKOCUKJST-UHFFFAOYSA-N 0.000 description 1
- 230000003247 decreasing effect Effects 0.000 description 1
- 238000009826 distribution Methods 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 238000001914 filtration Methods 0.000 description 1
- 230000002068 genetic effect Effects 0.000 description 1
- 238000010438 heat treatment Methods 0.000 description 1
- 238000009776 industrial production Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 229910052760 oxygen Inorganic materials 0.000 description 1
- 239000001301 oxygen Substances 0.000 description 1
- 238000011897 real-time detection Methods 0.000 description 1
- 238000000926 separation method Methods 0.000 description 1
- 238000007619 statistical method Methods 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
- 239000004753 textile Substances 0.000 description 1
Images
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 invention relates to a method, a device, equipment and a storage medium for optimizing and predicting production parameters, wherein the method comprises the following steps: acquiring monitoring data of each procedure in the production flow; preprocessing the monitoring data; constructing a strongest association chain representing the influence relation between two processes between any two processes by using a rule association algorithm, and combining the strongest association chain with the fluctuation state of the monitoring data to obtain a state association chain; and establishing a prediction model according to the state association chain by using a flexible neural tree algorithm to obtain and output a prediction result. The method can optimize the parameters of the key working procedures according to the prediction results, further optimize the production flow of the coal-fired boiler by optimizing the parameters of the key working procedures, and achieve the effects of saving energy, reducing emission and improving economy and production safety.
Description
Technical Field
The invention belongs to the technical field of coal-fired boiler production, and particularly relates to a production parameter optimization prediction method, a device, equipment and a storage medium.
Background
The boiler combustion process is a benefit conversion process, a large amount of heat energy generated by high-temperature water or steam of the boiler can be directly used, enters thousands of households through a pipeline, is applied to life and production of people, ensures the healthy development of life and production, provides good heating for houses, comprehensively adjusts the air quality, and provides application for various industries such as textile, chemical industry, papermaking and the like, meanwhile, the conversion of electric energy, mechanical energy and the like can be realized through the combustion of the boiler, the good operation and development of economy are realized, and the important position and the function of the boiler in production and life can be seen.
The control of the parameters of key procedures in the production flow of the coal-fired boiler is very important for the utilization of the coal-fired boiler. In order to ensure the reliability and economy of the operation of the coal-fired boiler unit, the parameters and efficiency of key processes need to be mastered online in real time, so that the parameters can be effectively optimized and controlled according to actual conditions.
However, the monitoring data directly obtained from the coal-fired boiler is massive, and usually has a large amount of noise data and missing information, and the mutual influence relationship between the processes cannot be directly reflected in the data, and the data has the characteristics of distribution, asynchrony and dispersion, and cannot be directly used for large data processing. People can hardly obtain valuable regularity information from mass monitoring data directly, and further can not adjust and optimize parameters such as main steam pressure, oxygen content, blower rotating speed and the like, and can not achieve the purposes of optimizing production, saving energy, reducing emission and improving production safety of enterprises.
Disclosure of Invention
The present invention is directed to a method, an apparatus, a device and a storage medium for optimizing and predicting production parameters, so as to solve the above technical problems.
In a first aspect, an embodiment of the present application provides a production parameter optimization prediction method, including: acquiring monitoring data of each procedure in the production flow; preprocessing the monitoring data; constructing a strongest association chain representing the influence relation between two processes between any two processes by using a rule association algorithm, and combining the strongest association chain with the fluctuation state of the monitoring data to obtain a state association chain; and establishing a prediction model according to the state association chain by using a flexible neural tree algorithm to obtain and output a prediction result.
With reference to the first aspect, in a first implementation manner of the first aspect, the preprocessing the monitoring data includes: merging the monitoring data in the same time period to obtain integrated monitoring data; filling blank monitoring data by using an arithmetic progression filling method or an averaging method; carrying out time sequence adjustment on monitoring data belonging to different procedures by utilizing a correlation coefficient curve; and acquiring the clustering set of the monitoring data by using a k-means algorithm.
With reference to the first implementation manner of the first aspect, in a second implementation manner of the first aspect, constructing a strongest association chain representing an influence relationship between two processes between any two processes by using a rule association algorithm, and combining the strongest association chain with the fluctuation state of the monitoring data to obtain a state association chain includes: processing the clustering data of any two working procedures by using a rule association algorithm to obtain a binomial association rule between any two clustering sets in different working procedures; calculating the association degree between any two cluster sets in different working procedures, and converting the association degree between the cluster sets into the association degree between the working procedures; according to the association rule and the association degree between any two working procedures, selecting the clustered data which meet the time sequence and have the maximum association degree for association, and constructing the strongest association chain between different working procedures; and combining the monitoring data fluctuation states among different processes with the strongest association chain to obtain the state association chain among different processes.
With reference to the second implementation manner of the first aspect, in a third implementation manner of the first aspect, establishing a prediction model according to the state association chain by using a flexible neural tree algorithm to obtain and output a prediction result includes: establishing a prediction model according to the state association chain and the preprocessed monitoring data by using a flexible neural tree method; inputting production input parameters and production output parameters of key procedures into the prediction model and performing iterative operation; and continuously modifying the iteration times to obtain the optimal iteration times, and verifying the accuracy of prediction by using the standard mean square deviation value and the matrix laboratory.
In a second aspect, an embodiment of the present application provides a production parameter optimization prediction apparatus, including: the data acquisition unit is configured for acquiring monitoring data of each process in the production flow; the preprocessing unit is configured to preprocess the monitoring data; the association chain construction unit is configured to construct a strongest association chain representing the influence relationship between two processes between any two processes by using a rule association algorithm, and combine the strongest association chain with the monitoring data fluctuation state to obtain a state association chain; and the model establishing unit is configured for establishing a prediction model according to the state association chain by using a flexible neural tree algorithm to obtain and output a prediction result.
With reference to the second aspect, in a first embodiment of the second aspect, the pre-processing unit comprises: the merging subunit is configured to merge the monitoring data in the same time period to obtain integrated monitoring data; the filling subunit is configured to fill the blank monitoring data by using an arithmetic progression filling method or an averaging method; the adjusting subunit is configured to perform time sequence adjustment on the monitoring data belonging to different procedures by using the correlation coefficient curve; and the clustering subunit is configured to acquire a clustering set of the monitoring data by using a k-means algorithm.
With reference to the first implementation manner of the second aspect, in a second implementation manner of the second aspect, the association chain constructing unit includes: the association rule subunit is configured to process the clustering data of any two working procedures by using a rule association algorithm to obtain a binomial association rule between any two clustering sets in different working procedures; the association degree subunit is configured to calculate association degrees between any two cluster sets in different processes, and convert the association degrees between the cluster sets into the association degrees between the processes; the strongest association chain subunit is configured to select the clustered data which meet the time sequence and have the maximum association degree for association according to the association rule and the association degree between any two working procedures, and construct the strongest association chain between different working procedures; and the state association chain subunit is configured to combine the monitoring data fluctuation states among different processes with the strongest association chain to obtain the state association chain among the different processes.
With reference to the second embodiment of the second aspect, in a third embodiment of the second aspect, the model establishing unit includes: a modeling subunit configured to establish a prediction model from the state association chain and the preprocessed monitoring data using a flexible neural tree approach; the operation subunit is configured to input production input parameters and production output parameters of the key process to the prediction model and perform iterative operation; and the checking subunit is configured to obtain the optimal iteration times by continuously modifying the iteration times, and check the accuracy of prediction by using the standard mean square deviation value and the matrix laboratory.
In a third aspect, an embodiment of the present application further provides an apparatus, including: one or more processors; memory for storing one or more programs which, when executed by the one or more processors, cause the one or more processors to perform the method according to any of the embodiments of the first aspect and the first aspect.
In a fourth aspect, an embodiment of the present application provides a storage medium, and the program, when executed by a processor, implements the method according to any one of the embodiments of the first aspect and the first aspect.
The beneficial effect of the invention is that,
the method and the device have the advantages that massive monitoring data are preprocessed, so that the method and the device can be used for big data processing. And further finding out the association relation among different procedures by using a rule association algorithm, and establishing a prediction model by using a flexible neural tree method on the basis, thereby optimizing the parameters of the key procedures according to the prediction result. The method optimizes the production flow of the coal-fired boiler by optimizing the parameters of the key working procedures, and achieves the effects of saving energy, reducing emission, and improving economy and production safety.
In addition, the invention has reliable design principle, simple structure and very wide application prospect.
Therefore, compared with the prior art, the invention has prominent substantive features and remarkable progress, and the beneficial effects of the implementation are also obvious.
Drawings
Other features, objects and advantages of the present application will become more apparent upon reading of the following detailed description of non-limiting embodiments thereof, made with reference to the accompanying drawings in which:
FIG. 1 is a flow chart of a method for optimizing and predicting production parameters in a production process of a coal-fired boiler according to an embodiment of the present disclosure;
FIG. 2 is a block diagram of a bar-shaped association link result analysis according to an embodiment of the present disclosure;
fig. 3 is a diagram of a modeled predicted effect provided by an embodiment of the present application.
Detailed Description
The present application will be described in further detail with reference to the following drawings and examples. It is to be understood that the specific embodiments described herein are merely illustrative of the relevant invention and not restrictive of the invention. It should be noted that, for convenience of description, only the portions related to the present invention are shown in the drawings.
It should be noted that the embodiments and features of the embodiments in the present application may be combined with each other without conflict. The present application will be described in detail below with reference to the embodiments with reference to the attached drawings.
The embodiment provides a production parameter optimization prediction method, which comprises the following steps:
s1, acquisition of monitoring data: monitoring data in the field of thermoelectric production has the characteristic of a typical flow object, the whole production process comprises a plurality of front and back related processes or processes, an OPC data acquisition interface is deployed in the processes, and real-time detection data can be stored in a database; different production systems, such as a hot water boiler, a chemical water system and an auxiliary system, are respectively arranged in different subsystems, are not interfered with each other, and are finally integrated into a total database, and the acquired data provide a basis for prediction analysis.
S2, preprocessing of monitoring data:
(1) the data integration and the data filling are carried out,
data integration: aiming at the problem of data dispersion, data are integrated, data of all processes in the same time segment are selected and combined into new data. The data obtained at this time contains a large amount of information of the process flow sequence, and a relatively accurate result can be obtained when the process flow sequence mode mining is carried out on the data.
And (3) controlling filling:
for a large amount of blank data, directly discarding the blank data; for data with a small amount of blank, an arithmetic series filling method is adopted with obvious trend, and an averaging method is adopted for filling without obvious trend. And the blank data is filled, so that the stability of the data is ensured. And the influence of the space-time value on the result is prevented from being predicted in the later modeling. In the flow object chi, if a certain process XiData containing partial missing values and missing time periods are { Xi(tm),Xi(tn) The data in between, then the missing data contains n-m-1 in total, the jth missing value in the segment is
If the data does not have the trend characteristic, the data is filled by using an averaging method, and the filled data in the segment is
Wherein, Xi(tj) Is Xi(tm) Previous n-m data, Xi(tk) Is Xi(tn) The next n-m data; if the vacancy value is in front of or behind the data, directly deleting the vacancy value; if Xi(tm) Front face or Xi(tn) If the number of the following data is less than n-m, the number of the following data is Xi(tn) Rear or Xi(tm) The total amount of the forward delay selection is 2(n-m) data, the obtained average value is ensured to contain enough information, and the calculation complexity is increased due to excessive selection data; the data amount is too small and is not enough to contain the information of the data.
(2) The time sequence is adjusted, and the time sequence is adjusted,
the sampling data of each process is regarded as time series (TimeSeries) data, and the final purpose is to calculate the influence relationship among different processes, which requires that the data among the processes are corresponding to each other according to the influence relationship. However, each record in the database contains the sampling data of different processes at the same time, which does not correspond to the influence relationship, and therefore, the adjustment needs to be performed according to the relative delay time of the processes, that is, the process data is moved forward or backward according to the relative delay time.
The data of each procedure is a time sequence with equal time intervals, namely a sampling value set arranged according to the time sequence. The influence between different processes constituting the influence between their corresponding two time sequences
Delay correlation two time series are subjected to delay correlation analysis, so that a time sequence Pearson correlation coefficient of two procedures can be obtained to measure the linear correlation degree of the two sequences, and the definition of the time sequence is shown as an equation (3).
Wherein the content of the first and second substances,
and
are the average of the sequence sums, respectively. The larger the absolute value of the correlation coefficient is, the stronger the correlation is; the closer the correlation coefficient is to 0, the weaker the correlation. The curve formed by the Pearson correlation coefficients of the two time series at different delay times is called a Pearson correlation coefficient curve. The correlation coefficient curve shows the trend of the correlation of the two time series along with the increase of the delay time. The extreme points of the curve, i.e. the points at which the correlation is greatest, generally characterize the influence of the two time series according to the delay time. And performing delay correlation analysis based on a Pearson correlation coefficient curve on the time series data of the two processes, drawing a correlation coefficient curve, and taking a delay value corresponding to the maximum value of the correlation coefficient as the relative delay time of the two processes. Thereby adjusting the process.
(3) Process clustering
The numerous discrete data states will greatly increase the time complexity of subsequent association analysis, and clustering can help us to reduce the number of discrete states of data, so that we can represent almost all data with a small number of K categories in a targeted manner. The clustering is carried out by using a K-Means algorithm, the clustering is carried out independently for each procedure, namely each procedure has a clustering number K and K types corresponding to the clustering number K, and the clustering numbers K of different procedures can be different. For the determination of K-value in clustering, we use the contour coefficient method based on the degree of agglomeration and degree of separation. After each process data is clustered, K categories are obtained, and finally, all the process data are replaced by the K categories, and the subsequent association analysis is performed based on the process data after clustering. The parallelization design of the K-Means algorithm based on data parallelization is given as follows:
inputting: time series data of a certain process, clustering number K
And (3) outputting: clustering result of the process data
1) Randomly initializing K clustering centers and broadcasting the K clustering centers to each computing node;
2) for each data of each partition, classifying the data into a cluster class represented by a cluster center closest to the data one by one;
3) for the data of each partition, calculating the sum of the data of the same clustering class and the sum of the number, collecting the data of the same clustering class and the sum of the number to a driver node, and recalculating a clustering center of the next iteration in the driver node;
4) if the clustering center is not changed any more or the specified iteration times are reached, the algorithm execution is ended; otherwise, broadcasting the new cluster center to each computing node, and then turning to step 2).
The K-means algorithm uses an objective function minimization criterion to control the number of iterations, and for numerical data, generally uses Euclidean Distance (Euclidean Distance) as an objective function. Is provided with two data objects X1And X2，EUCLID(X1,X2) Is X1And X2E is the sum of the squared errors of all objects, each object containing t variables, the euclidean distance between two objects is then the square root of the sum of the squares of the differences between the t variable values of the two objects, i.e. the sum of the squared errors of the two objects
The objective function is in the form of a squared error criterion function
Wherein, CiIs an arbitrary cluster, X is CiObject of (1), miIs of class CiIs measured. The distance measure in the objective function may also take other forms, such as manhattan distance, chebyshev distance, hamming distance, power distance, mahalanobis distance, and the like. Suppose an arbitrary process XiThe optimal number of clusters is KiAnd clustering according to the optimal clustering number to obtain a clustering set.
S3, discovery of association chain:
by analyzing the characteristics of the Apriori rule extraction process, the state relation is obtained by utilizing the characteristics
And (4) a linked algorithm. And obtaining the inter-class association rule of any two working procedures by using an Apriori-based inter-dimensional association rule algorithm under the condition of meeting the process flow conditions. Through the rules between each two, a plurality of associations are obtained.
1) Inter-class association rules: according to the requirements of the process object, a minimum support degree min _ sup and a minimum confidence degree min _ conf are set, clustering data of any two processes of the cluster set are mined based on an Apriori (rule association algorithm) inter-dimensional association rule, a frequent 2-predicate set is searched, and two association rules between any two clusters of different processes are generated, wherein the rules represent the relationship between two clusters of different processes meeting the minimum support degree and the minimum confidence degree. In the boiler parameters, we make a concrete description, the former term is XiThe latter term being XjAll the clustering data sets generate 2 clustering association rules together, the former item is XjThe latter term being XiGenerates 1 cluster association rule in total.
2) Step (3) association: setting two arbitrary procedures XiAnd XjCo-generate omega between clusters ofijThe rule, wherein any rule is ia → jb. Because the interest level I (ia → jb) has different meanings to the association rule, the signed numerical value is adopted to calculate the preceding item as XiAnd the latter term is XjDegree of association c between clusters ofij(beta) then
cij(β)＝Sβ(ia→jb)×I'β(ia→jb),β≤ωij (6)
Wherein β represents the β -th rule of all rules, and
I'β(ia→jb)＝Iβ(ia→jb)-1 (7)
through conversion, the association between the process clusters is converted into the association between the processes, and the process association relationship is obtained.
3) Strongest association chain and state association chain: currently, the association rule and the association degree between any two processes are known, and the data are selected to satisfy the time sequence L ═ { X'1,X'2,…,X'nAnd associating the data with the maximum association degree, so as to construct a strongest association chain, wherein the chain represents an influence relation of all processes in the process object during operation. And finally obtaining all strongest association chains in the historical data. In order to better utilize the correlation analysis results and make the correlation analysis results easy to be used in actual production to assist decision making, the states of the processes (such as rising, falling or unchanged states of monitoring data) are further combined with the correlation, and the state correlation between the processes, namely a state correlation chain, is obtained.
Correlation analysis: association Analysis (Association Analysis) mainly includes Association Rule (Association Rule) generation and Association Chain/Tree (Association Chain/Tree) generation. In association rule mining, association rules are generated only between every two processes, and finally association chains/trees are generated according to the association rules of every two processes. And finding out 2 association rules by using an Apriori classical association rule mining algorithm on the process clustering data, and filtering the generated rules according to the process time sequence.
And (3) state association: in order to better utilize the correlation analysis results and make the correlation analysis results easy to be used in actual production to assist decision making, the states of the processes (such as rising, falling or unchanged states of detection data) are further combined with the correlation to obtain the state correlation between the processes, which is called as an operation state knowledge base. The operation state knowledge base can be used for industrial assistance such as fault detection and early warning, parameter adjustment and the like. The state association can be obtained by using a statistical method, and is simple.
In the research of the method, the process closely related to the optimization target can be found from mass data by using the association rule algorithm, and the method is a key step of system research. However, the number of working procedures in thermoelectric production is large, the attributes participating in calculation are very many, the efficiency of obtaining the association rules by using a traditional association rule mining algorithm is too low, the calculation is complex, and the obtained knowledge is not easy to be directly understood by a user, so that a state association chain mode is used for representing the influence relation among all working procedures in production, and a representation method of the association chain is embodied in boiler parameters, for example, the pressure of a furnace 1 air bag is reduced to cause the left reduction of the main steam pressure of the furnace 1, and further, the left reduction of the differential pressure of a furnace 1 is caused, and the left increase of the water supply pressure of the furnace 1 is caused. By obtaining the association chain, key nodes can be found in the mass production process for regulation and control, and the result of the target process is quickly and directly influenced. Taking a hot water boiler system as an example, through the analysis of the association chain, in order to improve the heat output of the boiler, the aim can be achieved by increasing the power of a primary fan, reducing the opening of a baffle of a secondary fan and keeping the coal feeding quantity unchanged. The state association chain displays three states of ascending, descending and unchanging, and intuitively displays the front-back influence relation among a plurality of process states.
And finding out the strongest correlation chain with the strongest correlation by obtaining the state correlation chain, and more intuitively displaying and adjusting the relation among the working procedures by modeling at the following stages.
S4, establishing a prediction model:
this section is mainly to model the state association chain and processed data from the above work. And obtaining a modeling prediction formula by using a Flexible Neural Tree (FNT) method to realize the prediction of the process data.
1) The database tables used for modeling comprise a preprocessed data table, an FNT user table, an FNT operation parameter table, a test data table, a training data table, an association linked list, an output result table and the like. Iterative operations are performed on results obtained by processing of previous data.
2) For the parameters of the coal-fired boiler, finding out a key process and corresponding production input parameters and output parameters, inputting the key process and the corresponding production input parameters and output parameters into FNT to obtain An output function shaped as a formula (8), wherein An and Bn are parameters optimized by two groups of PSO (particle swarm optimization), and netn is a tree structure optimized by PIPE (pipelines used for communication between processes with genetic relationship); the function is used for carrying out predictive analysis on the change of future production.
3) By running the FNT algorithm. And continuously modifying the iteration times to obtain the optimal iteration times. The effectiveness accuracy of the training is determined by the value of the standard mean square error (NMSE) in the returned results. By the obtained formula, matlab (matrix laboratory) is used for preliminary image prediction analysis, and the accuracy of prediction is finally determined by comparison.
The modeling prediction has the functions of scientifically calculating relevant attributes influencing a target process on the basis of the state association chain, acquiring a change trend formula of measured point data, simulating an industrial production process, and predicting a production state in a future period of time on the basis of relevant parameters of a current production state, so as to assist in guiding an enterprise to adjust production process parameters. By modeling the state association chain generated by the second part, other processes except the last item in the association chain are used as input values of the prediction function, the target process is used as a function output value, and the values of the other processes are modified, so that the value of the target process after a certain time can be predicted, and the production process can be optimized in advance. In the thermoelectric production process, the hot water boiler is taken as an example, and the above step obtains a related chain: the method comprises the steps of increasing primary fan power → decreasing secondary fan baffle opening → keeping coal feeding quantity constant → improving thermal output, wherein in modeling prediction, the thermal output is a function output value, and the primary fan power, the secondary fan baffle opening and the boiler coal feeding quantity are function input values, so that the prediction method shown in the formula (9) can be obtained.
The embodiment provides a production parameter optimization prediction device, which comprises: the data acquisition unit is configured for acquiring monitoring data of each process in the production flow; the preprocessing unit is configured to preprocess the monitoring data; the association chain construction unit is configured to construct a strongest association chain representing the influence relationship between two processes between any two processes by using a rule association algorithm, and combine the strongest association chain with the monitoring data fluctuation state to obtain a state association chain; and the model establishing unit is configured for establishing a prediction model according to the state association chain by using a flexible neural tree algorithm to obtain and output a prediction result.
Wherein, the preprocessing unit includes: the merging subunit is configured to merge the monitoring data in the same time period to obtain integrated monitoring data; the filling subunit is configured to fill the blank monitoring data by using an arithmetic progression filling method or an averaging method; the adjusting subunit is configured to perform time sequence adjustment on the monitoring data belonging to different procedures by using the correlation coefficient curve; and the clustering subunit is configured to acquire a clustering set of the monitoring data by using a k-means algorithm.
The association chain building unit comprises: the association rule subunit is configured to process the clustering data of any two working procedures by using a rule association algorithm to obtain a binomial association rule between any two clustering sets in different working procedures; the association degree subunit is configured to calculate association degrees between any two cluster sets in different processes, and convert the association degrees between the cluster sets into the association degrees between the processes; the strongest association chain subunit is configured to select the clustered data which meet the time sequence and have the maximum association degree for association according to the association rule and the association degree between any two working procedures, and construct the strongest association chain between different working procedures; and the state association chain subunit is configured to combine the monitoring data fluctuation states among different processes with the strongest association chain to obtain the state association chain among the different processes.
The model building unit includes: a modeling subunit configured to establish a prediction model from the state association chain and the preprocessed monitoring data using a flexible neural tree approach; the operation subunit is configured to input production input parameters and production output parameters of the key process to the prediction model and perform iterative operation; and the checking subunit is configured to obtain the optimal iteration times by continuously modifying the iteration times, and check the accuracy of prediction by using the standard mean square deviation value and the matrix laboratory.
The present embodiment also provides an apparatus, including: one or more processors; a memory for storing one or more programs that, when executed by the one or more processors, cause the one or more processors to perform the method for production parameter optimization prediction provided herein.
In a fourth aspect, an embodiment of the present application provides a storage medium, and the program is executed by a processor to implement the method for optimizing and predicting the production parameter provided in the present application.
The above disclosure is only for the preferred embodiments of the present invention, but the present invention is not limited thereto, and any non-inventive changes that can be made by those skilled in the art and several modifications and amendments made without departing from the principle of the present invention shall fall within the protection scope of the present invention.
Claims (8)
1. A method for optimizing and predicting production parameters, the method comprising:
acquiring monitoring data of each procedure in the production flow;
preprocessing the monitoring data;
constructing a strongest association chain representing the influence relation between two processes between any two processes by using a rule association algorithm, and combining the strongest association chain with the fluctuation state of the monitoring data to obtain a state association chain;
establishing a prediction model according to the state association chain by using a flexible neural tree algorithm to obtain and output a prediction result;
the step of establishing a prediction model according to the state association chain by using a flexible neural tree algorithm to obtain and output a prediction result comprises the following steps:
establishing a predictive model from the state association chain and the preprocessed monitoring data using a flexible neural tree approach
Wherein An and Bn are parameters optimized by two groups of particle swarm optimization, netnThe method is a tree structure state association chain after pipeline communication optimization;
inputting production input parameters and production output parameters of key procedures into the prediction model and performing iterative operation;
and continuously modifying the iteration times to obtain the optimal iteration times, and verifying the accuracy of prediction by using the standard mean square deviation value and the matrix laboratory.
2. The method of claim 1, wherein the pre-processing the monitoring data comprises:
merging the monitoring data in the same time period to obtain integrated monitoring data;
filling blank monitoring data by using an arithmetic progression filling method or an averaging method;
carrying out time sequence adjustment on monitoring data belonging to different procedures by utilizing a correlation coefficient curve;
and acquiring the clustering set of the monitoring data by using a k-means algorithm.
3. The method of claim 2, wherein the using a rule association algorithm to construct a strongest association chain between any two processes, the strongest association chain representing an influence relationship between the two processes, and combining the strongest association chain with the fluctuation state of the monitoring data to obtain a state association chain comprises:
processing the clustering data of any two working procedures by using a rule association algorithm to obtain a binomial association rule between any two clustering sets in different working procedures;
calculating the association degree between any two cluster sets in different working procedures, and converting the association degree between the cluster sets into the association degree between the working procedures;
according to the association rule and the association degree between any two working procedures, selecting the clustered data which meet the time sequence and have the maximum association degree for association, and constructing the strongest association chain between different working procedures;
and combining the monitoring data fluctuation states among different processes with the strongest association chain to obtain the state association chain among different processes.
4. A production parameter optimization prediction apparatus, the apparatus comprising:
the data acquisition unit is configured for acquiring monitoring data of each process in the production flow;
the preprocessing unit is configured to preprocess the monitoring data;
the association chain construction unit is configured to construct a strongest association chain representing the influence relationship between two processes between any two processes by using a rule association algorithm, and combine the strongest association chain with the monitoring data fluctuation state to obtain a state association chain;
the model establishing unit is configured to utilize a flexible neural tree algorithm to establish a prediction model according to the state association chain, and obtain and output a prediction result;
the model building unit includes:
a modeling subunit configured to build a predictive model from the state association chain and the preprocessed monitoring data using a flexible neural tree approach
Wherein An and Bn are parameters optimized by two groups of particle swarm optimization, netnThe method is a tree structure state association chain after pipeline communication optimization;
the operation subunit is configured to input production input parameters and production output parameters of the key process to the prediction model and perform iterative operation;
and the checking subunit is configured to obtain the optimal iteration times by continuously modifying the iteration times, and check the accuracy of prediction by using the standard mean square deviation value and the matrix laboratory.
5. The apparatus of claim 4, wherein the pre-processing unit comprises:
the merging subunit is configured to merge the monitoring data in the same time period to obtain integrated monitoring data;
the filling subunit is configured to fill the blank monitoring data by using an arithmetic progression filling method or an averaging method;
the adjusting subunit is configured to perform time sequence adjustment on the monitoring data belonging to different procedures by using the correlation coefficient curve;
and the clustering subunit is configured to acquire a clustering set of the monitoring data by using a k-means algorithm.
6. The apparatus of claim 5, wherein the association chain constructing unit comprises:
the association rule subunit is configured to process the clustering data of any two working procedures by using a rule association algorithm to obtain a binomial association rule between any two clustering sets in different working procedures;
the association degree subunit is configured to calculate association degrees between any two cluster sets in different processes, and convert the association degrees between the cluster sets into the association degrees between the processes;
the strongest association chain subunit is configured to select the clustered data which meet the time sequence and have the maximum association degree for association according to the association rule and the association degree between any two working procedures, and construct the strongest association chain between different working procedures;
and the state association chain subunit is configured to combine the monitoring data fluctuation states among different processes with the strongest association chain to obtain the state association chain among the different processes.
7. An apparatus, characterized in that the apparatus comprises:
one or more processors;
a memory for storing one or more programs,
the one or more programs, when executed by the one or more processors, cause the one or more processors to perform the method recited in any of claims 1-3.
8. A computer-readable storage medium, in which a computer program is stored which, when being executed by a processor, carries out the method according to any one of claims 1-3.
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 CN108647808A (en) | 2018-10-12 |
| CN108647808B true 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) |
Families Citing this family (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| 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 |
| CN109884893B (en) * | 2019-02-28 | 2021-09-10 | 西安理工大学 | Multi-process variable dynamic time lag estimation method |
| 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 |
| CN111679634A (en) * | 2020-01-20 | 2020-09-18 | 武汉裕大华纺织有限公司 | Intelligent roving management system |
| 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 |
| 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 |
| CN113063455B (en) * | 2021-03-15 | 2023-04-07 | 上海联影医疗科技股份有限公司 | Detector parameter configuration method, equipment, electronic device and storage medium |
| 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 |
| 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 |
| 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 |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105096008A (en) * | 2015-08-28 | 2015-11-25 | 浙江大学 | Control method of process industry production system |
| 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 |
Family Cites Families (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20140143191A1 (en) * | 2012-11-20 | 2014-05-22 | Qualcomm Incorporated | Piecewise linear neuron modeling |
| US10748654B2 (en) * | 2015-07-02 | 2020-08-18 | PRA Health Sciences, Inc. | Normalizing data sets for predicting an attribute of the data sets |
-
2018
- 2018-04-11 CN CN201810322649.3A patent/CN108647808B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104346442B (en) * | 2014-10-14 | 2017-10-20 | 济南大学 | A kind of Rules extraction method of Process-Oriented object data |
| 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 |
|---|
| 基于电力系统大数据集的知识发现方法的研究与实现;宋巧云;《中国优秀硕士学位论文全文数据库信息科技辑》;20160615;全文 * |
| 流程对象建模方法的研究与实现;付爱芳;《中国优秀硕士学位论文全文数据库工程科技Ⅰ辑》;20110415;全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN108647808A (en) | 2018-10-12 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108647808B (en) | Production parameter optimization prediction method, device, equipment and storage medium | |
| CN108494021B (en) | Stability evaluation and static control method of electricity-heat-gas comprehensive energy system | |
| CN106709662B (en) | Power equipment operation condition division method | |
| CN100456281C (en) | Data division apparatus, data division method | |
| CN109992921B (en) | On-line soft measurement method and system for thermal efficiency of boiler of coal-fired power plant | |
| WO2023024433A1 (en) | Gas-steam combined cycle generator set operation adjustment and control system, and adjustment and control method | |
| CN101802728A (en) | Gaseous emission is carried out system and method based on the virtual sensing of experience set | |
| CN111260138B (en) | Dynamic optimization method for thermal power unit energy consumption by weighting and self-adaptive parallel association rule | |
| CN110532674A (en) | A kind of coal-fired power station boiler fire box temperature measurement method | |
| CN111006240B (en) | Biomass boiler furnace temperature and load prediction method | |
| CN115587433A (en) | State condition based unit consumption difference analysis optimization method and system | |
| CN112417764A (en) | K nearest neighbor regression prediction method for boiler special equipment steam flow prediction | |
| CN112503400A (en) | Multi-objective optimization arrangement method for pressure measuring points of water supply pipe network | |
| CN112365094A (en) | Method, computer device and storage medium for improving accuracy of prediction system | |
| CN113515891B (en) | Emulsion explosive quality prediction and optimization method | |
| CN112765746B (en) | Turbine blade top gas-thermal performance uncertainty quantification system based on polynomial chaos | |
| Cheng et al. | Efficient adaptive sampling methods based on deviation analysis for on-machine inspection | |
| Thota et al. | Analysis of feature selection techniques for prediction of boiler efficiency in case of coal based power plant using real time data | |
| CN113250768A (en) | Thermoelectric load optimization method for cogeneration heat supply unit | |
| JP4857228B2 (en) | Plant operation optimization device | |
| CN110320803B (en) | Method for calculating optimal load-rise rate of ultra-supercritical coal-fired unit | |
| CN114970766A (en) | Power station unit operation parameter reference value obtaining method based on linear fitting | |
| CN116029433A (en) | Energy efficiency reference value judging method, system, equipment and medium based on grey prediction | |
| CN109388884A (en) | A kind of generalized regression extreme value response phase method calculating coupling leaf dish fatigue life | |
| JP2010146068A (en) | Control device for plant, and control device of thermal power generation plant |
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 |