CN107544450B - Process industry network model construction method and system based on data - Google Patents
Process industry network model construction method and system based on data Download PDFInfo
- Publication number
- CN107544450B CN107544450B CN201710941085.7A CN201710941085A CN107544450B CN 107544450 B CN107544450 B CN 107544450B CN 201710941085 A CN201710941085 A CN 201710941085A CN 107544450 B CN107544450 B CN 107544450B
- Authority
- CN
- China
- Prior art keywords
- production equipment
- sensor
- event
- data
- numerical value
- 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
Classifications
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
- Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
- Y02P90/02—Total factory control, e.g. smart factories, flexible manufacturing systems [FMS] or integrated manufacturing systems [IMS]
Landscapes
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses process industry network model construction methods and system based on data, include: setting back end: flexible job shop includes several production equipments, each production equipment is referred to as flow nodes, several sensors are disposed in each production equipment, each sensor is referred to as back end；Data prediction: the data that each sensor is acquired are stored into corresponding bivariate table, will be compared before and after numerical value collected, are+1,0 and -1 by subsequent data modification；Side is set for selected back end: by+1,0 and -1 input value as Apriori algorithm, calculating separately the incidence relation between the incidence relation and sensor in successive different production equipments in the same production equipment between different sensors；If sensor is related, side is just set between back end；The weight on side between back end is set；Network is attached, the process industry network model based on data is obtained.
Description
Technical field
The present invention relates to process industry network model construction methods and system based on data.
Background technique
Process industry refers to as the industry such as electric power, metallurgy, chemical industry, medicine, they the characteristics of be continuity.Process industry
Production scheduling target includes economic indicator and performance indicator, is finally presented as that cost is minimum or profit is maximum.
Present process industry is still faced with many problems, with the development of productivity and science and technology, the life of process industry
The continuous expansion of production scale, production link also become more and more accurate and complicated.The production efficiency for how improving process industry, subtracts
Few energy consumption and waste discharge have become process industry urgent problem.However due to the complexity of modern process industry
Initial stage industrial age is much surmounted, we are difficult to make flow scheduling be optimal state according to traditional experiential operating.
It is how higher in the production process of process industry due to the limitation of the limitation of resource, especially non-renewable resources
Effect ground is significantly larger than previous using the significance level of resource.Meanwhile to sacrifice the promotion utilization of resources of the product quality as cost
Rate is worthless.The utilization rate for improving resource plays very big effect to cost is reduced.Compared with use it is more advanced production set
It is standby, it improves outside existing production technology, it is lower to optimize existing process cost, it is easier to realize.
With the continuous improvement and production-scale expansion required product quality, the production procedure of product is also constantly
It improves, while the complexity of process is also being continuously increased.Therefore the control difficulty of process is also increased with it.In quick-reading flow sheets
The management and optimization method applied in small-scale production expose disadvantage.Traditional data processing method can not meet at this stage
The mass data generated in production process.
With the arrival in industrial 4.0 epoch, industrial production starts to change to intelligence manufacture.Modern industry increasingly relies on
Data.Meanwhile the data volume generated in industrial production is but also fundamental change has occurred in industrial data.Traditional data processing side
Formula is no longer desirable for the processing of big data.
In task-resource flow network model of industrial flow, a certain task of node on behalf, while representing between two tasks
Resource transfer.For example, processing tasks include raw material or semi-finished product by handling a variety of chargings, corresponding discharging is generated, and will
It passes to other tasks of downstream.
Task resource network model has truly reacted each production unit in production process in entire production link
Status, the important node found in the simplification network of generation are consistent with the important link in actual production, are one successful
Model.
However, the model still remains certain deficiency, such as:
(1) what existing network model reflected is the association in production process between physical node, rather than physical node institute
The association between data for including；
(2) existing network model cannot react the dynamic process of production；Production link in process industry is not one one-tenth
Constant, once changing production plan, then the weight on side will change in network, and existing network cannot react new production
Plan；
(3) existing network model cannot react the utilization of the energy with refining.In actual production process, in addition to resource
Flowing, the energy utilization be also one cannot be neglected link, the prior art is optimized just for resource flow
It as a result may be locally optimal solution.
Summary of the invention
In order to solve the deficiencies in the prior art, the present invention provides the process industry network model building sides based on data
Method, farthest using the data generated in production, model can digitize entire production process, construct one by counting
The network of word composition, connection of the internetwork relationship between different data, so that more convenient make optimization for parameters.
Process industry network model construction method based on data, comprising:
Step (1): setting back end: flexible job shop includes several production equipments, and each production equipment is referred to as
For flow nodes, several sensors are disposed in each production equipment, each sensor is referred to as back end；
Step (2): data prediction: the data that each sensor is acquired are stored into corresponding bivariate table, will be adopted
Compare before and after the numerical value of collection, is+1,0 and -1 by subsequent data modification；Wherein ,+1 increase event is indicated, 0 indicates constant thing
Part, -1 indicates diminution event；
Step (3): side is set for selected back end: increase that step (2) is obtained, constant and reduce event and make
For the input value of Apriori algorithm, the incidence relation in the same production equipment between different sensors and front and back are calculated separately
Incidence relation in the different production equipments of linking between sensor；If sensor is related, just it is arranged between back end
Otherwise side is not provided with side；
Step (4): the weight on side between setting back end；
Step (5): according to the side and data section being arranged between back end, back end selected by step (1)-(4)
The weight on side, is attached network, obtains the process industry network model based on data between point.
The step of step (1) are as follows:
Using the operating state data for the sensor collecting flowchart node being deployed in each production equipment, flexible work is obtained
The operating state data set A={ A of all production equipments in industry workshop1,A2,A3...Ai...An, wherein Ai={ ai1,ai2,
ai3...aij...aimBe i-th of production equipment operating state data set；I indicates i-th of production equipment, the value model of i
Enclosing is 1~n, and n is positive integer, and n indicates the sum of production equipment；J indicates j-th of sensor, and the value range of j is 1~m, and m is
Positive integer, m indicate the sum for the sensor installed in corresponding production equipment, aijIndicate j-th of sensor in i-th of production equipment
The data of acquisition；
The data a that j-th of sensor acquires in i-th of production equipmentijIt include: data acquisition time tijWith acquisition numerical value
The step of step (2) are as follows:
The data that each sensor is acquired are stored into corresponding bivariate table, and bivariate table first is classified as data acquisition time
tij, bivariate table second is classified as the corresponding acquisition numerical value of data acquisition time
In the acquisition numerical value of the bivariate table of each sensorColumn vector in, by each acquisition numerical value with it is previous
The numerical value of item acquisition is compared,
If the numerical value currently acquired is greater than the numerical value of previous item acquisition, the numerical value currently acquired is revised as+1；
If the numerical value currently acquired is equal to the numerical value of previous item acquisition, the numerical value currently acquired is revised as 0；
If the numerical value currently acquired is less than the numerical value of previous item acquisition, the numerical value currently acquired is revised as -1；
Finally, first numerical value is revised as 0；
Wherein ,+1 increase event is indicated, 0 indicates invariant event, and -1 indicates diminution event；
To which it only includes increase that the column vector of all the sensors acquisition numerical value, which becomes, is reduced, constant three kinds of events
Event set.
The step (2):
Gather the event set of all the sensors of each production equipment as one；
Assuming that AiFor the set of the data of all the sensors acquisition of i-th production equipment；Ai+1It is set for the production of i+1 platform
The set of the data of standby all the sensors acquisition；I-th production equipment belongs to i+1 platform production equipment in production procedure
Adjacent production equipment；I-th production equipment is located at the upstream of i+1 platform production equipment in production procedure.
Incidence relation step in the same production equipment of calculating of the step (3) between different sensors are as follows:
By the thing of the event set of the 1st sensor of i-th production equipment and the 2nd sensor of i-th production equipment
Part collection, is input in Apriori algorithm, export i-th production equipment the 1st sensor and i-th production equipment the 2nd
The association status of a sensor；
By the thing of the event set of the 1st sensor of i-th production equipment and the 3rd sensor of i-th production equipment
Part collection, is input in Apriori algorithm, exports the event set and i-th production of the 1st sensor of i-th production equipment
The association status of 3rd sensor of equipment；
By the thing of the event set of the 1st sensor of i-th production equipment and the 4th sensor of i-th production equipment
Part collection, is input in Apriori algorithm, exports the event set and i-th production of the 1st sensor of i-th production equipment
The association status of 4th sensor of equipment；
And so on, obtain i-th production equipment the 1st sensor and i-th production equipment in addition to the 1st senses
The incidence relation between other all the sensors other than device, if association, is arranged side between two back end, otherwise
It is not provided with side；
And so on, obtain the incidence relation between all the sensors of i-th production equipment itself；
And so on, it obtains between the sensor of each production equipment itself and the other sensors of production equipment itself
Incidence relation.
The calculating step of incidence relation in the successive different production equipments of the step (3) between sensor
Are as follows:
By the thing of the event set of the 1st sensor of i-th production equipment and each sensor of i+1 platform production equipment
Part collection, is input in Apriori algorithm, exports the 1st sensor and i+1 platform production equipment of i-th production equipment
The association status of each sensor；
By the thing of the event set of the 2nd sensor of i-th production equipment and each sensor of i+1 platform production equipment
Part collection, is input in Apriori algorithm, exports the 2nd sensor and i+1 platform production equipment of i-th production equipment
The association status of each sensor；
By the thing of the event set of the 3rd sensor of i-th production equipment and each sensor of i+1 platform production equipment
Part collection, is input in Apriori algorithm, exports the 3rd sensor and i+1 platform production equipment of i-th production equipment
The association status of each sensor；
And so on, obtain i-th production equipment all the sensors and i+1 platform production equipment each sensor it
Between incidence relation, if association, side is set between two back end, is otherwise not provided with side；
And so on, obtain the incidence relation between the sensor of the adjacent production equipment of any two.
The associated judgment criteria of the step (3) is:
Assuming that judging that associated two sensors are j-th of sensor and the life of i+1 platform in i-th production equipment respectively
Produce first of sensor in equipment；
Judge j-th of sensors A in i-th production equipmentijAcquire the state-event M at the column vector current time of numerical value
With first of sensors A under synchronization in i+1 platform production equipment(i+1)lAcquire the state-event N of the column vector of numerical value；Institute
Stating state-event includes :+1,0 or -1；
If M=1 and N=1, then it represents that AijAcquire numerical value raising and A(i+1)lAcquire the raised event of numerical value, event it is general
Rate is expressed as P (1,1)；
If M=0 and N=0, then it represents that AijAcquire that numerical value is constant and A(i+1)lAcquire the constant event of numerical value, event it is general
Rate is expressed as P (0,0)；
If M=-1 and N=-1, then it represents that AijAcquire numerical value reduction and A(i+1)lThe event that numerical value reduces is acquired, event
Probability is expressed as P (- 1, -1)；
If M=1 and N=-1, then it represents that AijAcquire numerical value raising and A(i+1)lThe event that numerical value reduces is acquired, event
Probability is expressed as P (1, -1)；
If M=-1 and N=1, then it represents that AijAcquire numerical value reduction and A(i+1)lThe raised event of numerical value is acquired, event
Probability is expressed as P (- 1,1)；
If M=1 and N=0, then it represents that AijAcquire numerical value raising and A(i+1)lAcquire the constant event of numerical value, event it is general
Rate is expressed as P (1,0)；
If M=0 and N=1, then it represents that AijAcquire that numerical value is constant and A(i+1)lAcquire the raised event of numerical value, event it is general
Rate is expressed as P (0,1)；
If M=0 and N=-1, then it represents that AijAcquire that numerical value is constant and A(i+1)lThe event that numerical value reduces is acquired, event
Probability is expressed as P (0, -1)；
If M=-1 and N=0, then it represents that AijAcquire numerical value reduction and A(i+1)lThe constant event of numerical value is acquired, event
Probability is expressed as P (- 1,0)；
If two groups of data are related, there are three types of situations, be respectively be positively correlated (with increasing with subtracting), negatively correlated (shifting),
One group of data does not change and another group of data also do not change；The relevant probability of happening of data is respectively: P (1,1) ∪ P
(0,0) ∪ P (- 1, -1) and P (1, -1) ∪ P (0,0) ∪ P (- 1,1)；
If two groups of data are unrelated, one group of data changes, and another group of data do not change, the incoherent thing of data
Part probability is P (1,0) ∪ P (0,1) ∪ P (0, -1) ∪ P (- 1,0)；
Support s (M, N)=P (M ∪ N)；
Wherein, P (M ∪ N) indicates the percentage of the total event of state-event M and N concurrent Zhan；
Confidence level c (M, N)=P (N | M)；
Wherein, when P (N | M) indicates that state-event M occurs, probability that state-event N also occurs；
Finally, A is calculated using Apriori algorithm, the association situation of B:
Wherein minimum support is 10%, min confidence 75%, if support s (M, N) and confidence level c (M, N) are
It is then Qiang Guanlian, there are sides by A, B greater than the minimum support and min confidence pre-defined.
The step of step (4) are as follows:
It is inscribed when calculating each, sensors A(i+1)lAcquire numerical valueWith sensors AijThe numerical value of acquisitionRatio；
The ratio adduction inscribed when will be all, then average；Using the average value as the power on side between back end
Weight.
In order to solve the deficiencies in the prior art, the present invention also provides the process industry network models based on data to construct system
System, farthest using the data generated in production, model can digitize entire production process, construct one by counting
The network of word composition, connection of the internetwork relationship between different data, so that more convenient make optimization for parameters.
Process industry network model based on data constructs system, comprising: memory, processor and is stored in memory
Computer instruction that is upper and running on a processor, the computer instruction execute following steps when being run by processor:
Step (1): setting back end: flexible job shop includes several production equipments, and each production equipment is referred to as
For flow nodes, several sensors are disposed in each production equipment, each sensor is referred to as back end；
Step (2): data prediction: the data that each sensor is acquired are stored into corresponding bivariate table, will be adopted
Compare before and after the numerical value of collection, is+1,0 and -1 by subsequent data modification；Wherein ,+1 increase event is indicated, 0 indicates constant thing
Part, -1 indicates diminution event；
Step (3): side is set for selected back end: increase that step (2) is obtained, constant and reduce event and make
For the input value of Apriori algorithm, the incidence relation in the same production equipment between different sensors and front and back are calculated separately
Incidence relation in the different production equipments of linking between sensor；If sensor is related, just it is arranged between back end
Otherwise side is not provided with side；
Step (4): the weight on side between setting back end；
Step (5): according to the side and data section being arranged between back end, back end selected by step (1)-(4)
The weight on side, is attached network, obtains the process industry network model based on data between point.
A kind of computer readable storage medium is stored thereon with computer instruction, and the computer instruction is by processor
When operation, following steps are completed:
Step (1): setting back end: flexible job shop includes several production equipments, and each production equipment is referred to as
For flow nodes, several sensors are disposed in each production equipment, each sensor is referred to as back end；
Step (2): data prediction: the data that each sensor is acquired are stored into corresponding bivariate table, will be adopted
Compare before and after the numerical value of collection, is+1,0 and -1 by subsequent data modification；Wherein ,+1 increase event is indicated, 0 indicates constant thing
Part, -1 indicates diminution event；
Step (3): side is set for selected back end: increase that step (2) is obtained, constant and reduce event and make
For the input value of Apriori algorithm, the incidence relation in the same production equipment between different sensors and front and back are calculated separately
Incidence relation in the different production equipments of linking between sensor；If sensor is related, just it is arranged between back end
Otherwise side is not provided with side；
Step (4): the weight on side between setting back end；
Step (5): according to the side and data section being arranged between back end, back end selected by step (1)-(4)
The weight on side, is attached network, obtains the process industry network model based on data between point.
Compared with prior art, the beneficial effects of the present invention are:
1 in task-resource network, we can accurately find out the important node in network, these nodes often have
Apparent structure feature, such as the degree of node is bigger, these nodes play important role, Wo Men in a network
It generally requires to put into more energy to these nodes in production process.However, task resource network asks multiple-objection optimization
It inscribes inflexible.
2 in the complex network model based on data, node be made of data, and the side of network then reflect data it
Between connection, using the model, we can will be the optimization to data to the transformation of process, and advantage mainly has following
Several points:
Shield complicated production link: model only reflects the connection between creation data, and no longer pays close attention to these data categories
In which link, the optimization problem to data will be converted to the optimization problem of process.
Utmostly utilize creation data: industrial data generate in the way of and data type and internet or transportation network
The gap data of equal generations is larger, and industrial big data is often pure data, and the redundancy of data is higher, therefore is being handled
Great care is wanted to the choice of data when industrial big data, which eliminates the mistake of artificial screening data in modeling process
Journey avoids the loss of significant data.
Optimize more accurate: in production process, some production link may include several item datas, in previous optimization side
In case, we artificially filter out the high parameter of different degree and abandon the low parameter of different degree often by the help of domain expert.
While reducing Optimization Work amount, mass data is also had lost, so that the result and optimal result of optimization have certain deviation.
In the model, our optimization object is no longer some production unit, but certain item data, therefore optimization specific aim is stronger,
Accuracy is higher.
Detailed description of the invention
The accompanying drawings constituting a part of this application is used to provide further understanding of the present application, and the application's shows
Meaning property embodiment and its explanation are not constituted an undue limitation on the present application for explaining the application.
Fig. 1 is flow chart of the invention；
Relationship of the Fig. 2 between back end and process entities；
Specific embodiment
It is noted that following detailed description is all illustrative, it is intended to provide further instruction to the application.Unless another
It indicates, all technical and scientific terms used herein has usual with the application person of an ordinary skill in the technical field
The identical meanings of understanding.
It should be noted that term used herein above is merely to describe specific embodiment, and be not intended to restricted root
According to the illustrative embodiments of the application.As used herein, unless the context clearly indicates otherwise, otherwise singular
Also it is intended to include plural form, additionally, it should be understood that, when in the present specification using term "comprising" and/or " packet
Include " when, indicate existing characteristics, step, operation, device, component and/or their combination.
As shown in Figure 1, the process industry network model construction method based on data, comprising:
Step (1): setting back end: flexible job shop includes several production equipments, and each production equipment is referred to as
For flow nodes, several sensors are disposed in each production equipment, each sensor is referred to as back end；
Using the operating state data for the sensor collecting flowchart node being deployed in each production equipment, flexible work is obtained
The operating state data set A={ A of all production equipments in industry workshop1,A2,A3...Ai...An, wherein Ai={ ai1,ai2,
ai3...aij...aimBe i-th of production equipment operating state data set；I indicates i-th of production equipment, the value model of i
Enclosing is 1~n, and n is positive integer, and n indicates the sum of production equipment；J indicates j-th of sensor, and the value range of j is 1~m, and m is
Positive integer, m indicate the sum for the sensor installed in corresponding production equipment, aijIndicate j-th of sensor in i-th of production equipment
The data of acquisition；
The data a that j-th of sensor acquires in i-th of production equipmentijIt include: data acquisition time tijWith acquisition numerical valueSuch as: prepared by batch may be comprising multi-group datas such as temperature, wind speed, oxygen concentrations in node.
Step (2): data prediction: the data that each sensor is acquired are stored into corresponding bivariate table, will be adopted
Compare before and after the numerical value of collection, is+1,0 and -1 by subsequent data modification；Wherein ,+1 increase event is indicated, 0 indicates constant thing
Part, -1 indicates diminution event；
The data that each sensor is acquired are stored into corresponding bivariate table, and bivariate table first is classified as data acquisition time
tij, bivariate table second is classified as the corresponding acquisition numerical value of data acquisition time
In the acquisition numerical value of the bivariate table of each sensorColumn vector in, by each acquisition numerical value with it is previous
The numerical value of item acquisition is compared,
If the numerical value currently acquired is greater than the numerical value of previous item acquisition, the numerical value currently acquired is revised as+1；
If the numerical value currently acquired is equal to the numerical value of previous item acquisition, the numerical value currently acquired is revised as 0；
If the numerical value currently acquired is less than the numerical value of previous item acquisition, the numerical value currently acquired is revised as -1；
Finally, first numerical value is revised as 0；
Wherein ,+1 increase event is indicated, 0 indicates invariant event, and -1 indicates diminution event；
To which it only includes increase that the column vector of all the sensors acquisition numerical value, which becomes, is reduced, constant three kinds of events
Event set.
Gather the event set of all the sensors of each production equipment as one；
Assuming that AiFor the set of the data of all the sensors acquisition of i-th production equipment；Ai+1It is set for the production of i+1 platform
The set of the data of standby all the sensors acquisition；I-th production equipment belongs to i+1 platform production equipment in production procedure
Adjacent production equipment；I-th production equipment is located at the upstream of i+1 platform production equipment in production procedure.
Step (3): side is set for selected back end: increase that step (2) is obtained, constant and reduce event and make
For the input value of Apriori algorithm, the incidence relation in the same production equipment between different sensors and front and back are calculated separately
Incidence relation in the different production equipments of linking between sensor；If sensor is related, just it is arranged between back end
Otherwise side is not provided with side；
Calculate the incidence relation step in the same production equipment between different sensors are as follows:
By the thing of the event set of the 1st sensor of i-th production equipment and the 2nd sensor of i-th production equipment
Part collection, is input in Apriori algorithm, export i-th production equipment the 1st sensor and i-th production equipment the 2nd
The association status of a sensor；
By the thing of the event set of the 1st sensor of i-th production equipment and the 3rd sensor of i-th production equipment
Part collection, is input in Apriori algorithm, exports the event set and i-th production of the 1st sensor of i-th production equipment
The association status of 3rd sensor of equipment；
By the thing of the event set of the 1st sensor of i-th production equipment and the 4th sensor of i-th production equipment
Part collection, is input in Apriori algorithm, exports the event set and i-th production of the 1st sensor of i-th production equipment
The association status of 4th sensor of equipment；
And so on, obtain i-th production equipment the 1st sensor and i-th production equipment in addition to the 1st senses
The incidence relation between other all the sensors other than device, if association, is arranged side between two back end, otherwise
It is not provided with side；
And so on, obtain the incidence relation between all the sensors of i-th production equipment itself；
And so on, it obtains between the sensor of each production equipment itself and the other sensors of production equipment itself
Incidence relation.
Such as: there are incidence relations between the temperature sensor and humidity sensor of First production equipment；
The calculating step of incidence relation in the successive different production equipments between sensor are as follows:
By the thing of the event set of the 1st sensor of i-th production equipment and each sensor of i+1 platform production equipment
Part collection, is input in Apriori algorithm, exports the 1st sensor and i+1 platform production equipment of i-th production equipment
The association status of each sensor；
By the thing of the event set of the 2nd sensor of i-th production equipment and each sensor of i+1 platform production equipment
Part collection, is input in Apriori algorithm, exports the 2nd sensor and i+1 platform production equipment of i-th production equipment
The association status of each sensor；
By the thing of the event set of the 3rd sensor of i-th production equipment and each sensor of i+1 platform production equipment
Part collection, is input in Apriori algorithm, exports the 3rd sensor and i+1 platform production equipment of i-th production equipment
The association status of each sensor；
And so on, obtain i-th production equipment all the sensors and i+1 platform production equipment each sensor it
Between incidence relation, if association, side is set between two back end, is otherwise not provided with side；
And so on, obtain the incidence relation between the sensor of the adjacent production equipment of any two.
Such as: the pass on the temperature sensor and second production equipment in First production equipment between baroceptor
Connection relationship.
The associated judgment criteria is:
Assuming that judging that associated two sensors are j-th of sensor and the life of i+1 platform in i-th production equipment respectively
Produce first of sensor in equipment；
Judge j-th of sensors A in i-th production equipmentijAcquire the state-event M at the column vector current time of numerical value
With first of sensors A under synchronization in i+1 platform production equipment(i+1)lAcquire the state-event N of the column vector of numerical value；Institute
Stating state-event includes :+1,0 or -1；
If M=1 and N=1, then it represents that AijAcquire numerical value raising and A(i+1)lAcquire the raised event of numerical value, event it is general
Rate is expressed as P (1,1)；
If M=0 and N=0, then it represents that AijAcquire that numerical value is constant and A(i+1)lAcquire the constant event of numerical value, event it is general
Rate is expressed as P (0,0)；
If M=-1 and N=-1, then it represents that AijAcquire numerical value reduction and A(i+1)lThe event that numerical value reduces is acquired, event
Probability is expressed as P (- 1, -1)；
If M=1 and N=-1, then it represents that AijAcquire numerical value raising and A(i+1)lThe event that numerical value reduces is acquired, event
Probability is expressed as P (1, -1)；
If M=-1 and N=1, then it represents that AijAcquire numerical value reduction and A(i+1)lThe raised event of numerical value is acquired, event
Probability is expressed as P (- 1,1)；
If M=1 and N=0, then it represents that AijAcquire numerical value raising and A(i+1)lAcquire the constant event of numerical value, event it is general
Rate is expressed as P (1,0)；
If M=0 and N=1, then it represents that AijAcquire that numerical value is constant and A(i+1)lAcquire the raised event of numerical value, event it is general
Rate is expressed as P (0,1)；
If M=0 and N=-1, then it represents that AijAcquire that numerical value is constant and A(i+1)lThe event that numerical value reduces is acquired, event
Probability is expressed as P (0, -1)；
If M=-1 and N=0, then it represents that AijAcquire numerical value reduction and A(i+1)lThe constant event of numerical value is acquired, event
Probability is expressed as P (- 1,0)；
If two groups of data are related, there are three types of situations, be respectively be positively correlated (with increasing with subtracting), negatively correlated (shifting),
One group of data does not change and another group of data also do not change；The relevant probability of happening of data is respectively: P (1,1) ∪ P
(0,0) ∪ P (- 1, -1) and P (1, -1) ∪ P (0,0) ∪ P (- 1,1)；
If two groups of data are unrelated, one group of data changes, and another group of data do not change, the incoherent thing of data
Part probability is P (1,0) ∪ P (0,1) ∪ P (0, -1) ∪ P (- 1,0)；
Support s (M, N)=P (M ∪ N)；
Wherein, P (M ∪ N) indicates the percentage of the total event of state-event M and N concurrent Zhan；
Confidence level c (M, N)=P (N | M)；
Wherein, when P (N | M) indicates that state-event M occurs, probability that state-event N also occurs；
Finally, A is calculated using Apriori algorithm, the association situation of B:
Wherein minimum support is 10%, min confidence 75%, if support s (M, N) and confidence level c (M, N) are
It is then Qiang Guanlian, there are sides by A, B greater than the minimum support and min confidence pre-defined.
Step (4): the weight on side between setting back end；
It is inscribed when calculating each, sensors A(i+1)lAcquire numerical valueWith sensors AijThe numerical value of acquisitionRatio；
The ratio adduction inscribed when will be all, then average；Using the average value as the power on side between back end
Weight.
Step (5): according to the side and data section being arranged between back end, back end selected by step (1)-(4)
The weight on side, is attached network, obtains the process industry network model based on data between point.
Side between back end is directed edge, and direction is based on production procedure, and side can only be by superior node or section at the same level
Point is directed toward junior or the brother of node, cannot be directed toward superior node by downstream site.Relationship between back end and process entities
As shown in Figure 2.
The foregoing is merely preferred embodiment of the present application, are not intended to limit this application, for the skill of this field
For art personnel, various changes and changes are possible in this application.Within the spirit and principles of this application, made any to repair
Change, equivalent replacement, improvement etc., should be included within the scope of protection of this application.
Claims (10)
1. the process industry network model construction method based on data, characterized in that include:
Step (1): setting back end: flexible job shop includes several production equipments, and each production equipment is referred to as to flow
Cheng Jiedian disposes several sensors in each production equipment, and each sensor is referred to as back end；
Step (2): data prediction: the data that each sensor is acquired are stored into corresponding bivariate table, will be collected
Compare before and after numerical value, is+1,0 and -1 by subsequent data modification；Wherein ,+1 expression increase event, 0 expression invariant event, -1
Indicate diminution event；
Step (3): side is set for selected back end: increase that step (2) is obtained, it is constant and reduce event as
The input value of Apriori algorithm calculates separately incidence relation and front and back in the same production equipment between different sensors and holds in the mouth
Incidence relation in the different production equipments connect between sensor；If sensor is related, side is just set between back end,
Otherwise it is not provided with side；
Step (4): the weight on side between setting back end；
Step (5): according to the side and back end being arranged between back end, back end selected by step (1)-(4) it
Between side weight, network is attached, the process industry network model based on data is obtained.
2. the process industry network model construction method based on data as described in claim 1, characterized in that the step
(1) the step of are as follows:
Using the operating state data for the sensor collecting flowchart node being deployed in each production equipment, flexible job vehicle is obtained
The operating state data set A={ A of interior all production equipments1,A2,A3...Ai...An, wherein Ai={ ai1,ai2,
ai3...aij...aimBe i-th of production equipment operating state data set；I indicates i-th of production equipment, the value model of i
Enclosing is 1~n, and n is positive integer, and n indicates the sum of production equipment；J indicates j-th of sensor, and the value range of j is 1~m, and m is
Positive integer, m indicate the sum for the sensor installed in corresponding production equipment, aijIndicate j-th of sensor in i-th of production equipment
The data of acquisition；
The data a that j-th of sensor acquires in i-th of production equipmentijIt include: data acquisition time tijWith acquisition numerical value
3. the process industry network model construction method based on data as described in claim 1, characterized in that the step
(2) the step of are as follows:
The data that each sensor is acquired are stored into corresponding bivariate table, and bivariate table first is classified as data acquisition time tij,
Bivariate table second is classified as the corresponding acquisition numerical value of data acquisition time
In the acquisition numerical value of the bivariate table of each sensorColumn vector in, by each acquire numerical value adopted with previous item
The numerical value of collection is compared,
If the numerical value currently acquired is greater than the numerical value of previous item acquisition, the numerical value currently acquired is revised as+1；
If the numerical value currently acquired is equal to the numerical value of previous item acquisition, the numerical value currently acquired is revised as 0；
If the numerical value currently acquired is less than the numerical value of previous item acquisition, the numerical value currently acquired is revised as -1；
Finally, first numerical value is revised as 0；
Wherein ,+1 increase event is indicated, 0 indicates invariant event, and -1 indicates diminution event；
To which it only includes increase that the column vector of all the sensors acquisition numerical value, which becomes, is reduced, the event of constant three kinds of events
Collection.
4. the process industry network model construction method based on data as described in claim 1, characterized in that the step
(2):
Gather the event set of all the sensors of each production equipment as one；
Assuming that AiFor the set of the data of all the sensors acquisition of i-th production equipment；Ai+1For the institute of i+1 platform production equipment
The set for the data for thering is sensor to acquire；I-th production equipment and i+1 platform production equipment belong to adjacent life in production procedure
Produce equipment；I-th production equipment is located at the upstream of i+1 platform production equipment in production procedure.
5. the process industry network model construction method based on data as described in claim 1, characterized in that the step
(3) the incidence relation step in the same production equipment of calculating between different sensors are as follows:
By the event set of the event set of the 1st sensor of i-th production equipment and the 2nd sensor of i-th production equipment,
It is input in Apriori algorithm, exports the 1st sensor of i-th production equipment and the 2nd biography of i-th production equipment
The association status of sensor；
By the event set of the event set of the 1st sensor of i-th production equipment and the 3rd sensor of i-th production equipment,
It is input in Apriori algorithm, exports the event set of the 1st sensor of i-th production equipment and i-th production equipment
The association status of 3rd sensor；
By the event set of the event set of the 1st sensor of i-th production equipment and the 4th sensor of i-th production equipment,
It is input in Apriori algorithm, exports the event set of the 1st sensor of i-th production equipment and i-th production equipment
The association status of 4th sensor；
And so on, obtain i-th production equipment the 1st sensor and i-th production equipment in addition to the 1st sensor with
Incidence relation between other outer all the sensors, if association, is arranged side between two back end, does not otherwise set
Set side；
And so on, obtain the incidence relation between all the sensors of i-th production equipment itself；
And so on, obtain the pass between the sensor of each production equipment itself and the other sensors of production equipment itself
Connection relationship.
6. the process industry network model construction method based on data as described in claim 1, characterized in that the step
(3) the calculating step of the incidence relation in successive different production equipments between sensor are as follows:
By the event of the event set of the 1st sensor of i-th production equipment and each sensor of i+1 platform production equipment
Collection, is input in Apriori algorithm, exports the every of the 1st sensor of i-th production equipment and i+1 platform production equipment
The association status of a sensor；
By the event of the event set of the 2nd sensor of i-th production equipment and each sensor of i+1 platform production equipment
Collection, is input in Apriori algorithm, exports the every of the 2nd sensor of i-th production equipment and i+1 platform production equipment
The association status of a sensor；
By the event of the event set of the 3rd sensor of i-th production equipment and each sensor of i+1 platform production equipment
Collection, is input in Apriori algorithm, exports the every of the 3rd sensor of i-th production equipment and i+1 platform production equipment
The association status of a sensor；
And so on, it obtains between all the sensors of i-th production equipment and each sensor of i+1 platform production equipment
Incidence relation, if association, is arranged side between two back end, is otherwise not provided with side；
And so on, obtain the incidence relation between the sensor of the adjacent production equipment of any two.
7. the process industry network model construction method based on data as described in claim 1, characterized in that the step
(3) associated judgment criteria is:
Assuming that judging that associated two sensors are that j-th of sensor in i-th production equipment and the production of i+1 platform are set respectively
Standby first upper of sensor；
Judge j-th of sensors A in i-th production equipmentijAcquire the state-event M at the column vector current time of numerical value and same
First of sensors A in i+1 platform production equipment is inscribed for the moment(i+1)lAcquire the state-event N of the column vector of numerical value；The thing
Part state includes :+1,0 or -1；
If M=1 and N=1, then it represents that AijAcquire numerical value raising and A(i+1)lAcquire the raised event of numerical value, the probability tables of event
It is shown as P (1,1)；
If M=0 and N=0, then it represents that AijAcquire that numerical value is constant and A(i+1)lAcquire the constant event of numerical value, the probability tables of event
It is shown as P (0,0)；
If M=-1 and N=-1, then it represents that AijAcquire numerical value reduction and A(i+1)lAcquire the event that numerical value reduces, the probability of event
It is expressed as P (- 1, -1)；
If M=1 and N=-1, then it represents that AijAcquire numerical value raising and A(i+1)lAcquire the event that numerical value reduces, the probability of event
It is expressed as P (1, -1)；
If M=-1 and N=1, then it represents that AijAcquire numerical value reduction and A(i+1)lAcquire the raised event of numerical value, the probability of event
It is expressed as P (- 1,1)；
If M=1 and N=0, then it represents that AijAcquire numerical value raising and A(i+1)lAcquire the constant event of numerical value, the probability tables of event
It is shown as P (1,0)；
If M=0 and N=1, then it represents that AijAcquire that numerical value is constant and A(i+1)lAcquire the raised event of numerical value, the probability tables of event
It is shown as P (0,1)；
If M=0 and N=-1, then it represents that AijAcquire that numerical value is constant and A(i+1)lAcquire the event that numerical value reduces, the probability of event
It is expressed as P (0, -1)；
If M=-1 and N=0, then it represents that AijAcquire numerical value reduction and A(i+1)lAcquire the constant event of numerical value, the probability of event
It is expressed as P (- 1,0)；
If two groups of data correlations, there are three types of situations, are to be positively correlated (same to increase same subtract), negatively correlated (shifting), one group respectively
Data do not change and another group of data also do not change；The relevant probability of happening of data is respectively: P (1,1) ∪ P (0,
0) ∪ P (- 1, -1) and P (1, -1) ∪ P (0,0) ∪ P (- 1,1)；
If two groups of data are unrelated, one group of data changes, and another group of data do not change, and the incoherent event of data is general
Rate is P (1,0) ∪ P (0,1) ∪ P (0, -1) ∪ P (- 1,0)；
Support s (M, N)=P (M ∪ N)；
Wherein, P (M ∪ N) indicates the percentage of the total event of state-event M and N concurrent Zhan；
Confidence level c (M, N)=P (N | M)；
Wherein, when P (N | M) indicates that state-event M occurs, probability that state-event N also occurs；
Finally, A is calculated using Apriori algorithm, the association situation of B:
Wherein minimum support is 10%, min confidence 75%, if support s (M, N) and confidence level c (M, N) is both greater than
The minimum support and min confidence pre-defined is then Qiang Guanlian, and there are sides by A, B.
8. the process industry network model construction method based on data as described in claim 1, characterized in that the step
(4) the step of are as follows:
It is inscribed when calculating each, sensors A(i+1)lAcquire numerical value dt(i+1)lWith sensors AijThe numerical value d of acquisitiontijRatio；
The ratio adduction inscribed when will be all, then average；Using the average value as the weight on side between back end.
9. the process industry network model based on data constructs system, characterized in that include: memory, processor and storage
The computer instruction run on a memory and on a processor, the computer instruction when being run by processor, execute with
Lower step:
Step (1): setting back end: flexible job shop includes several production equipments, and each production equipment is referred to as to flow
Cheng Jiedian disposes several sensors in each production equipment, and each sensor is referred to as back end；
Step (2): data prediction: the data that each sensor is acquired are stored into corresponding bivariate table, will be collected
Compare before and after numerical value, is+1,0 and -1 by subsequent data modification；Wherein ,+1 expression increase event, 0 expression invariant event, -1
Indicate diminution event；
Step (3): side is set for selected back end: increase that step (2) is obtained, it is constant and reduce event as
The input value of Apriori algorithm calculates separately incidence relation and front and back in the same production equipment between different sensors and holds in the mouth
Incidence relation in the different production equipments connect between sensor；If sensor is related, side is just set between back end,
Otherwise it is not provided with side；
Step (4): the weight on side between setting back end；
Step (5): according to the side and back end being arranged between back end, back end selected by step (1)-(4) it
Between side weight, network is attached, the process industry network model based on data is obtained.
10. a kind of computer readable storage medium, is stored thereon with computer instruction, characterized in that the computer instruction exists
When being run by processor, following steps are completed:
Step (1): setting back end: flexible job shop includes several production equipments, and each production equipment is referred to as to flow
Cheng Jiedian disposes several sensors in each production equipment, and each sensor is referred to as back end；
Step (2): data prediction: the data that each sensor is acquired are stored into corresponding bivariate table, will be collected
Compare before and after numerical value, is+1,0 and -1 by subsequent data modification；Wherein ,+1 expression increase event, 0 expression invariant event, -1
Indicate diminution event；
Step (3): side is set for selected back end: increase that step (2) is obtained, it is constant and reduce event as
The input value of Apriori algorithm calculates separately incidence relation and front and back in the same production equipment between different sensors and holds in the mouth
Incidence relation in the different production equipments connect between sensor；If sensor is related, side is just set between back end,
Otherwise it is not provided with side；
Step (4): the weight on side between setting back end；
Step (5): according to the side and back end being arranged between back end, back end selected by step (1)-(4) it
Between side weight, network is attached, the process industry network model based on data is obtained.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710941085.7A CN107544450B (en) | 2017-10-11 | 2017-10-11 | Process industry network model construction method and system based on data |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710941085.7A CN107544450B (en) | 2017-10-11 | 2017-10-11 | Process industry network model construction method and system based on data |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN107544450A CN107544450A (en) | 2018-01-05 |
| CN107544450B true CN107544450B (en) | 2019-06-21 |
Family
ID=60967808
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710941085.7A Active CN107544450B (en) | 2017-10-11 | 2017-10-11 | Process industry network model construction method and system based on data |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107544450B (en) |
Families Citing this family (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108681801B (en) * | 2018-02-22 | 2020-10-27 | 石化盈科信息技术有限责任公司 | Method and system for monitoring operation index in production process |
| CN109685386A (en) * | 2018-12-29 | 2019-04-26 | 齐鲁工业大学 | The complex network model construction method of multi-objective Job Shop based on data information |
| WO2021087896A1 (en) * | 2019-11-07 | 2021-05-14 | Alibaba Group Holding Limited | Data-driven graph of things for data center monitoring copyright notice |
| CN111930079A (en) * | 2020-06-30 | 2020-11-13 | 西门子工厂自动化工程有限公司 | Data acquisition method, device and system, electronic equipment and storage medium |
| CN114218197A (en) * | 2021-12-17 | 2022-03-22 | 上海繁易信息科技股份有限公司 | Object-oriented industrial Internet of things data modeling method and system |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102289738A (en) * | 2011-07-28 | 2011-12-21 | 冶金自动化研究设计院 | Method for establishing flow network model |
| CN102968523A (en) * | 2012-11-02 | 2013-03-13 | 浙江工业大学 | Modeling approach for shipbuilding enterprise subsection workshop spatial scheduling model based on cellular machine |
| CN103530710A (en) * | 2012-07-03 | 2014-01-22 | 鸿富锦精密工业（深圳）有限公司 | Production line model establishing system and method |
| EP2853972A2 (en) * | 2013-09-30 | 2015-04-01 | YPF Tecnologia S.A. | Device and method for detection and/or diagnosis of faults in a process, equipment and sensors |
| CN104914835A (en) * | 2015-05-22 | 2015-09-16 | 齐鲁工业大学 | Flexible job-shop scheduling multi-objective method |
| CN105094850A (en) * | 2015-08-28 | 2015-11-25 | 安庆师范学院 | Node layer relevance mapping based simple business process model establishing method and application |
| CN107103050A (en) * | 2017-03-31 | 2017-08-29 | 海通安恒（大连）大数据科技有限公司 | A kind of big data Modeling Platform and method |
-
2017
- 2017-10-11 CN CN201710941085.7A patent/CN107544450B/en active Active
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102289738A (en) * | 2011-07-28 | 2011-12-21 | 冶金自动化研究设计院 | Method for establishing flow network model |
| CN103530710A (en) * | 2012-07-03 | 2014-01-22 | 鸿富锦精密工业（深圳）有限公司 | Production line model establishing system and method |
| CN102968523A (en) * | 2012-11-02 | 2013-03-13 | 浙江工业大学 | Modeling approach for shipbuilding enterprise subsection workshop spatial scheduling model based on cellular machine |
| EP2853972A2 (en) * | 2013-09-30 | 2015-04-01 | YPF Tecnologia S.A. | Device and method for detection and/or diagnosis of faults in a process, equipment and sensors |
| CN104914835A (en) * | 2015-05-22 | 2015-09-16 | 齐鲁工业大学 | Flexible job-shop scheduling multi-objective method |
| CN105094850A (en) * | 2015-08-28 | 2015-11-25 | 安庆师范学院 | Node layer relevance mapping based simple business process model establishing method and application |
| CN107103050A (en) * | 2017-03-31 | 2017-08-29 | 海通安恒（大连）大数据科技有限公司 | A kind of big data Modeling Platform and method |
Also Published As
| Publication number | Publication date |
|---|---|
| CN107544450A (en) | 2018-01-05 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107544450B (en) | Process industry network model construction method and system based on data | |
| Xu et al. | Mid-term prediction of electrical energy consumption for crude oil pipelines using a hybrid algorithm of support vector machine and genetic algorithm | |
| Qiao et al. | An improved dolphin swarm algorithm based on Kernel Fuzzy C-means in the application of solving the optimal problems of large-scale function | |
| Dong et al. | Regional differences study of renewable energy performance: A case of wind power in China | |
| CN111353656B (en) | Steel enterprise oxygen load prediction method based on production plan | |
| CN106651089A (en) | Modeling and optimization method for production scheduling problem distribution set robust model | |
| CN109800898A (en) | A kind of intelligence short-term load forecasting method and system | |
| CN106650944A (en) | Metallurgical enterprise converter gas scheduling method based on knowledge | |
| CN110689162A (en) | Bus load prediction method, device and system based on user side classification | |
| Pei et al. | Twenty‐year retrospection on green manufacturing: A bibliometric perspective | |
| CN113344449A (en) | Method for predicting monthly-scale industrial water demand | |
| CN110221540A (en) | Continuous-stirring reactor system control method based on Hammerstein model | |
| Xia et al. | Efficient energy use in manufacturing systems—Modeling, assessment, and management strategy | |
| CN114240687A (en) | Energy hosting efficiency analysis method suitable for comprehensive energy system | |
| Zhang et al. | A regional water resource allocation model based on the human–water harmony theory in the Yellow River Basin | |
| Pang et al. | A photovoltaic power predicting model using the differential evolution algorithm and multi-task learning | |
| CN111859783B (en) | Water pressure prediction method, system, storage medium, equipment and urban water supply system | |
| Ma et al. | Comprehensive evaluation of renewable energy for power projects based on CA-DEA model | |
| Gong et al. | Dynamic Preventive Maintenance Optimization of Subway Vehicle Traction System Considering Stages | |
| CN108364136B (en) | Water resource shortage risk analysis method and system based on evidence reasoning | |
| Lu-ping et al. | Particle swarm optimization model of distributed network planning | |
| CN114004384A (en) | Project progress deviation early warning method and device | |
| Xie | Ecological network analysis of urban land resource use within China | |
| Li et al. | Identification of urban industry life cycle based on rough set neural network | |
| Jiang et al. | Model knowledge matching algorithm for steelmaking casting scheduling |
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 |