CN114398049A - Self-adaptive dynamic updating method for digital twin model of discrete manufacturing workshop - Google Patents

Self-adaptive dynamic updating method for digital twin model of discrete manufacturing workshop Download PDF

Info

Publication number
CN114398049A
CN114398049A CN202111438858.2A CN202111438858A CN114398049A CN 114398049 A CN114398049 A CN 114398049A CN 202111438858 A CN202111438858 A CN 202111438858A CN 114398049 A CN114398049 A CN 114398049A
Authority
CN
China
Prior art keywords
model
data
digital twin
virtual
production line
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
CN202111438858.2A
Other languages
Chinese (zh)
Inventor
钱伟伟
郭宇
张立童
张�浩
刘赛
崔凯
晏立超
陶亚宁
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Nanjing University of Aeronautics and Astronautics
Original Assignee
Nanjing University of Aeronautics and Astronautics
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Nanjing University of Aeronautics and Astronautics filed Critical Nanjing University of Aeronautics and Astronautics
Priority to CN202111438858.2A priority Critical patent/CN114398049A/en
Publication of CN114398049A publication Critical patent/CN114398049A/en
Pending legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F8/00—Arrangements for software engineering
    • G06F8/60—Software deployment
    • G06F8/65—Updates

Abstract

The invention discloses a self-adaptive dynamic updating method of a digital twin model of a discrete manufacturing workshop, which comprises the steps of firstly collecting data of a virtual workshop and an actual workshop, selecting a characteristic data set, respectively calculating a true value of a performance index and a theoretical value precision error predicted by a model, and then taking the true value and the theoretical value precision error as a triggering condition for dynamic correction of the twin model based on a Mann-Kendall precision error trend analysis method; then, DNN and LSTM are selected as base learners, an integrated learning method is adopted to conduct continuous iterative training, the weight updating mode of the base learners is improved from the two aspects of data importance and performance importance, an Adaboost-DNN-LSTM-based twin model dynamic correction algorithm is formed, and a difference base learner optimization method is adopted to complete the twin model dynamic correction. The digital twin model self-adaptive dynamic updating method provided by the invention provides a model updating method for the precise application of the digital twin of the discrete manufacturing system, and has important value for improving the intelligent level of workshop production management and control.

Description

Self-adaptive dynamic updating method for digital twin model of discrete manufacturing workshop
Technical Field
The invention relates to the technical field of workshop digitization, and mainly relates to a self-adaptive dynamic updating method for a digital twin model of a discrete manufacturing workshop.
Background
With the rise of industrial 4.0 and intelligent manufacturing, new technical means and tools such as internet of things, big data and artificial intelligence are continuously emerging, and the new techniques and tools assist the transformation and upgrade of the traditional manufacturing industry so as to promote the high-quality development of the manufacturing industry. The application of intelligent technology also puts new requirements on the transformation and the upgrade of the traditional discrete mechanical product production workshop. Digital twinning is an emerging technology for intelligent manufacturing in recent years and is also one of means for intelligent manufacturing floor application. As a new workshop operation mode, a digital twin manufacturing workshop has been raised in the manufacturing industry and gradually explored and applied, and accurate twin model correction is an important guarantee for completing digital twin construction of the workshop.
The current research on the digital twin of the plant is still in the preliminary exploration phase. The digital twin manufacturing workshop is used as the interaction and the fusion of a physical world and an information world and the execution basis of manufacturing activities, and the optimal production and management and control of the workshop are achieved under the drive of workshop data. However, in the digital twin application, the precision of the twin model is reduced along with the time, so that the model is inaccurate, and the efficiency and the precision of the workshop management and control are affected. In fact, how to perform dynamic adaptive update on the model of the digital twin manufacturing shop, that is, when the precision error of the twin model is significantly increased, the model is corrected, and it is challenging to correct the model deviation. The accuracy of the twin model is guaranteed to directly influence the digital twin construction precision, and further influence the effect of the digital twin construction of the workshop.
The digital twin is used as an effective means for realizing physical and information fusion, and is applied to a discrete manufacturing workshop, twin data is used as drive, a physical model and a virtual model are fused, and iterative optimization of virtual-real mapping and virtual-control real is completed. The digital twinning technology is used for describing and modeling the characteristics, behaviors, forming processes and performances of physical entity objects by utilizing a digital technology, and a digital mirror image which is completely the same as the physical entity objects in a physical space exists in a virtual space, so that a digital space model and a physical space model of a product and a production system are in real-time interaction, and the product and the physical space model can grasp the dynamic changes of each other in time and respond in real time. The digital twin model (hereinafter referred to as twin model) refers to a digital model which is the same as the physical entity model in geometric parameters and performance parameters; in the product operation stage, the twin model and the actual physical entity of the twin model can form real-time dynamic linkage. The testable twin model refers to the fact that repeated tests and experiments can be carried out on the twin model without driving a physical entity so as to verify the function and the performance of the physical entity. In order to achieve compliance with the function and performance of the physical entity during the operational phase, the design phase must be able to perform the testability of the twin model, i.e. the physical entity geometry, function and performance can be verified before production. However, in the construction process of a digital twin manufacturing workshop, the precision of the twin model is reduced along with the time, the management and control effect of the workshop is affected, if the twin model is not corrected in time, wrong scheme instructions may occur, and the operation efficiency of the workshop is obviously reduced. Through research on related patents and papers existing at present, the twin model has some progress in visualization, control logic consistency and data real-time interaction, but in the digital twin field of a discrete manufacturing workshop, the twin model often stays at the twin model construction stage, and the adaptability problem of the twin model in application is not considered, so that the problem of poor application effect of the twin model in the digital twin construction of a physical workshop is caused.
In summary, the model dynamic self-adaptive correction method for the digital twin manufacturing workshop is provided, and has important significance for ensuring the precision in the application of the twin model, improving the digital twin construction effect of the workshop, improving the manufacturing level and remolding the competitive advantage.
Disclosure of Invention
The purpose of the invention is as follows: aiming at the problems in the prior art, the invention provides a self-adaptive dynamic updating method for a digital twin model of a discrete manufacturing workshop, which comprises the steps of constructing a twin model precision error model, establishing twin model updating mechanism conditions, and performing dynamic self-adaptive correction on the twin model after reaching a twin model correction route. In the twin model precision error construction, firstly, the performance index of the twin model which is preferentially corrected needs to be determined, and the selection of the performance index is completed by taking a square loss function as a judgment basis for selecting the performance index. Secondly, performing significance inspection on the precision of the performance index based on a Mann-Kendall deviation trend analysis method, and judging whether the performance index needs to be corrected. And finally, slicing workshop production data, taking DNN and LSTM as a correction basis, carrying out model training, forming a strong learner by adopting an Adaboost integrated learning method, replacing a performance model in an inaccurate twin model by the strong learner, completing self-adaptive dynamic correction of the twin model, and solving the technical problem that the twin model cannot be dynamically corrected in the application of the existing discrete manufacturing workshop digital twin technology.
The technical scheme is as follows: in order to achieve the purpose, the invention adopts the technical scheme that:
a self-adaptive dynamic updating method for a digital twin model of a discrete manufacturing workshop comprises the following steps:
s1, collecting production data of a plurality of continuous production processes according to time sequence through an internet of things sensing device deployed in a digital twin manufacturing workshop, and selecting a characteristic data set; the characteristic data set comprises order task data, task data completed at the current moment, real-time production state data and predicted time; meanwhile, data acquisition is carried out on the virtual workshop to obtain a characteristic data set;
step S2, taking the square loss function as the response error of the performance target, and adopting a maximum and minimum normalization method to normalize the target response error and confirm the performance target needing to be corrected preferentially;
s3, performing significance test on the precision of the performance index by adopting a Mann-Kendall-based deviation trend analysis method; when the precision error exceeds a preset threshold value, the digital twin model is subjected to self-adaptive updating;
and S4, selecting DNN and LSTM as a base learner by adopting an integrated learning method based on the digital twin model updating mechanism of the step S3, carrying out difference comparison training on the base learner until a strong learner is formed, and finally updating the digital twin model.
Further, step S2 includes two processes of calculating a response error and selecting a performance target to be modified; in particular, the amount of the solvent to be used,
s2.1, constructing a performance target response error function; using a quadratic loss function
Figure BDA0003379144510000031
The target response error was constructed as follows:
Figure BDA0003379144510000032
wherein the content of the first and second substances,
Figure BDA0003379144510000033
actual value, y, representing the ith target response in the twin modeliF (x; i) represents a theoretical value of the ith target response in the twin model; based on the target response error function, completing the calculation of each performance target response error;
s2.2, selecting a performance target to be corrected; response error of target by adopting maximum and minimum normalization method
Figure BDA0003379144510000034
Figure BDA0003379144510000035
Normalization processing is carried out, and the influence of different dimensions is eliminated; the calculation method is as follows:
Figure BDA0003379144510000036
wherein the content of the first and second substances,
Figure BDA0003379144510000037
absolute value of error Δ ER for target response pairijThe calculation formula of (2) is as follows:
△ERij＝|ERi-ERj|
i and j respectively represent different target responses contained in the target response pair; further defining the proportion of each target response to the error difference, namely the fractional coefficient w of the target response pairijThe calculation formula is as follows:
Figure BDA0003379144510000038
n represents the total number of target responses; w is aijThe larger the value of (b), the larger the response error indicating the ith and jth targets, the higher the priority of correction required.
Further, the Mann-Kendall-based deviation trend analysis method in step S3 specifically includes:
s3.1, obtaining the testing precision change of the digital twin model; the calculation method comprises the following steps:
Figure BDA0003379144510000041
wherein the content of the first and second substances,
Figure BDA0003379144510000042
an actual value representing the i-th index at time t; f (x; i; t) represents the theoretical value of the ith index at the time t under the digital twin model; the test precision of the model at the t +1 th moment is TAi t+1And then the testing precision of the digital twin model is as follows:
Figure BDA0003379144510000043
a complete digital twin model test accuracy change of the ith index is expressed as:
Figure BDA0003379144510000044
s3.2, adopting a Mann-Kendall trend testing method to test the testing precision delta TA of the digital twin modeliThe trend change condition of (1); when the testing precision of the digital twin model shows a descending trend, triggering an updating mechanism of the digital twin model; specifically, the formula for the Mann-Kendall trend test is:
Figure BDA0003379144510000045
Figure BDA0003379144510000046
wherein T is the number of numerical values in the digital twin model test precision sequence, and S isi>When 0, the precision change is changed upwards; when S isi<When 0, the precision change is changed downwards; when T is>At 10, use of ZiStatistic amount, and calculate SiThe calculation formula is as follows:
Figure BDA0003379144510000047
Figure BDA0003379144510000048
wherein Var (S)i) Denotes SiG represents the number of groups of junctions, tj*Representing the number of values in the jth knot group; ziRepresenting a test statistic; when Z is<At 0, it represents Δ TAiThe twin model precision shows an ascending trend along with the descending change of time; when Z is>At 0 timeThen, a description will be given of Δ TAiThe accuracy of the twin model shows a descending trend along with the increasing change of time, and a digital twin model updating mechanism is triggered at the moment.
Further, the step S4 of adaptively updating the digital twin model specifically includes:
s4.1, updating an Adaboost weight function; updating the weight function by adopting a time weighting method, which comprises the following specific steps:
ωt＝exp(-ηt)(t＝1,2,…,T；η∈[0,0.99])
wherein t represents the data acquisition time, namely the data batch number acquired at the time t; when the data batches are sequentially increased, t is increased, and the more new the data is, the stronger the importance is;
s4.2, updating an iterative weighting mechanism of the sample; the sample weight updating mechanism in the iterative process of training the base learner by the Adaboost-DNN-LSTM algorithm is as follows:
Figure BDA0003379144510000051
Figure BDA0003379144510000052
Figure BDA0003379144510000053
where U represents the total number of iterations of Adaboost-DNN-LSTM, m*Representing the total number of samples;
Figure BDA0003379144510000054
denotes the ith*The weight of the u +1 th iteration of a sample,
Figure BDA0003379144510000055
is a parameter for controlling the sample weight adjustment direction; xi is a threshold value of the model error, when the error exceeds the threshold value, the weight of the class sample is increased, otherwise, the sample is reducedA weight;
step S4.3, determining the weight alpha of the base learneru(ii) a By alphauShows the voting weight of the u-th base learner, the accuracy error epsilon of the data set by the DNN and LSTM base learnersuThe following calculation formula is adopted:
Figure BDA0003379144510000061
Figure BDA0003379144510000062
wherein
Figure BDA0003379144510000063
Denotes the ith*The number of samples of the input is,
Figure BDA0003379144510000064
an actual value that represents the progress of the production,
Figure BDA0003379144510000065
represents a production progress prediction value of the u-th base learner,
Figure BDA0003379144510000066
representing the predicted value of the production schedule of the u-th base learner DNN,
Figure BDA0003379144510000067
representing the predicted value of the production progress of the u-th base learner LSTM;
s4.4, integrating a base learner; and integrating the trained base learners according to the voting weight to obtain a final strong learner Adaboost-DNN-LSTM, wherein the integration formula is as follows:
Figure BDA0003379144510000068
finally, the strong learner is used for updating the performance model in the original digital twin model to complete dynamic self-adaptive correction.
Further, in step S1, the digital twin manufacturing plant includes a virtual production line, a physical production line, a simulation model, a logic model, and a data model; the virtual production line is connected with the physical production line through a simulation model, a logic model and a data model, the simulation model is connected with the logic model, and the logic model is connected with the data model; the virtual production line is a digital mapping of the physical production line in a digital space; the simulation model is an artificial intelligence algorithm supporting the virtual production line to run in an actual production line state; the logic model is the operation logic of the actual production of the physical production line.
Further, the virtual production line comprises virtual production equipment, virtual conveying equipment, virtual monitoring equipment and other virtual equipment; the virtual production equipment, the virtual conveying equipment, the virtual monitoring equipment and other virtual equipment are respectively connected with the logic model and the data model; the physical production line comprises production equipment, conveying equipment, monitoring equipment and other equipment; the production equipment, the conveying equipment, the monitoring equipment and other equipment are respectively connected with the logic model and the data model; each device of the virtual production line corresponds to the same-name device on the physical production line respectively; the sensor is arranged on the virtual production line and is respectively connected with the production equipment, the conveying equipment, the monitoring equipment and other equipment.
Further, in the operation process of the virtual production line, the virtual sensor acquires data of the virtual production equipment in real time and transmits the data to the production line data model service platform, and the sensor reads the data from the production line data model service platform and acts on the production equipment; and the production equipment performs production action to obtain data of a physical production line, transmits the data of the physical production line to a production line data model service platform to be compared with the data of the virtual production line, and adjusts the path and the position of the virtual production equipment if the data of the production equipment is different from the data of the virtual production equipment so as to enable the operation results of the virtual production line and the physical production line to be consistent.
Has the advantages that:
the digital twin model self-adaptive dynamic updating method for the discrete manufacturing workshop provided by the invention fully utilizes knowledge information contained in mass manufacturing data, and effectively integrates Adaboost, DNN and LSTM, the theoretical method is simple and easy to realize, the calculation efficiency is greatly improved, and the self-adaptive updating requirement is met. Precise application of digital twinning for discrete manufacturing systems, comprising: the performance analysis, online decision and optimization provide a model updating method, and the method has important value for improving the intelligent management and control level of workshop production. The method can be used for guiding the construction process of the digital twin model, solving the problem of performance attenuation of the twin model along with time change, and effectively improving the construction precision of the twin model.
Drawings
FIG. 1 is a schematic diagram of a discrete manufacturing shop digital twin model adaptive dynamic update method provided by the present invention;
FIG. 2 is a flow chart of a digital twin model adaptive dynamic update method provided by the present invention;
FIG. 3 is a flowchart of the integrated Adaboost-DNN-LSTM algorithm provided in an embodiment of the present invention.
Detailed Description
The present invention will be further described with reference to the accompanying drawings. It is to be understood that the embodiments described are only a few embodiments of the present invention, and not all embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
The invention provides a self-adaptive dynamic updating method of a digital twin model of a discrete manufacturing workshop, which specifically comprises the following steps:
s1, collecting production data of a plurality of continuous production processes according to time sequence by an internet of things sensing device deployed in a digital twin manufacturing workshop, and selecting a characteristic data set; the characteristic data set comprises order task data, task data completed at the current moment, real-time production state data and predicted time; and similarly, data acquisition is carried out on the virtual workshop to obtain a characteristic data set. Wherein the two feature data sets are acquired independently of each other and have the same classification.
The data sources of the digital twin workshop are various, the data from different sources are different in terms of coding modes, data formats, application characteristics and the like, and the data of different manufacturing elements are also obviously different in form, semantics and identification. The digital twin manufacturing workshop comprises a virtual production line, a physical production line, a simulation model, a logic model and a data model; the virtual production line is connected with the physical production line through a simulation model, a logic model and a data model, the simulation model is connected with the logic model, and the logic model is connected with the data model; the virtual production line is a digital mapping of the physical production line in a digital space; the simulation model is an artificial intelligence algorithm supporting the virtual production line to run in an actual production line state; the logic model is the operation logic of the actual production of the physical production line.
The virtual production line comprises virtual production equipment, virtual conveying equipment, virtual monitoring equipment and other virtual equipment; the virtual production equipment, the virtual conveying equipment, the virtual monitoring equipment and other virtual equipment are respectively connected with the logic model and the data model; the physical production line comprises production equipment, conveying equipment, monitoring equipment and other equipment; the production equipment, the conveying equipment, the monitoring equipment and other equipment are respectively connected with the logic model and the data model; each device of the virtual production line corresponds to the same-name device on the physical production line respectively; the sensor is arranged on the virtual production line and is respectively connected with the production equipment, the conveying equipment, the monitoring equipment and other equipment.
In the running process of a virtual production line, a virtual sensor acquires data of virtual production equipment in real time and transmits the data to a production line data model service platform, and the sensor reads the data from the production line data model service platform and acts on the production equipment; and the production equipment performs production action to obtain data of a physical production line, transmits the data of the physical production line to a production line data model service platform to be compared with the data of the virtual production line, and adjusts the path and the position of the virtual production equipment if the data of the production equipment and the data of the virtual production equipment are different, so that the operation results of the virtual production line and the physical production line are consistent, and the data interaction of the testable digital twin body of the intelligent production line is realized.
Step S2, taking the square loss function as the response error of the performance target, and adopting a maximum and minimum normalization method to normalize the target response error and confirm the performance target needing to be corrected preferentially; in particular, the amount of the solvent to be used,
s2.1, constructing a performance target response error function; using a quadratic loss function
Figure BDA0003379144510000081
The target response error was constructed as follows:
Figure BDA0003379144510000082
wherein the content of the first and second substances,
Figure BDA0003379144510000083
actual value, y, representing the ith target response in the twin modeliF (x; i) represents a theoretical value of the ith target response in the twin model; based on the target response error function, completing the calculation of each performance target response error;
s2.2, selecting a performance target to be corrected; response error of target by adopting maximum and minimum normalization method
Figure BDA0003379144510000084
Figure BDA0003379144510000085
Normalization processing is carried out, and the influence of different dimensions is eliminated; the calculation method is as follows:
Figure BDA0003379144510000086
wherein the content of the first and second substances,
Figure BDA0003379144510000091
absolute value of error Δ ER for target response pairijThe calculation formula of (2) is as follows:
△ERij＝|ERi-ERj|
i and j respectively represent different target responses contained in the target response pair; further defining the proportion of each target response to the error difference, namely the fractional coefficient w of the target response pairijThe calculation formula is as follows:
Figure BDA0003379144510000092
n represents the total number of target responses; w is aijThe larger the value of (b), the larger the response error indicating the ith and jth targets, the higher the priority of correction required.
S3, performing significance test on the precision of the performance index by adopting a Mann-Kendall-based deviation trend analysis method; when the precision error exceeds a preset threshold value, the digital twin model is subjected to self-adaptive updating; in particular, the amount of the solvent to be used,
s3.1, obtaining the testing precision change of the digital twin model; the calculation method comprises the following steps:
Figure BDA0003379144510000093
wherein the content of the first and second substances,
Figure BDA0003379144510000094
an actual value representing the i-th index at time t; f (x; i; t) represents the theoretical value of the ith index at the time t under the digital twin model; the test precision of the model at the t +1 th moment is TAi t+1And then the testing precision of the digital twin model is as follows:
Figure BDA0003379144510000095
a complete digital twin model test accuracy change of the ith index is expressed as:
Figure BDA0003379144510000096
s3.2, adopting a Mann-Kendall trend testing method to test the testing precision delta TA of the digital twin modeliThe trend change condition of (1); when the testing precision of the digital twin model shows a descending trend, triggering an updating mechanism of the digital twin model; specifically, the formula for the Mann-Kendall trend test is:
Figure BDA0003379144510000101
Figure BDA0003379144510000102
wherein T is the number of numerical values in the digital twin model test precision sequence, and S isi>When 0, the precision change is changed upwards; when S isi<When 0, the precision change is changed downwards; when T is>At 10, use of ZiStatistic amount, and calculate SiThe calculation formula is as follows:
Figure BDA0003379144510000103
Figure BDA0003379144510000104
wherein Var (S)i) Denotes SiG represents the number of groups of junctions, tj*Representing the number of values in the jth knot group; ziRepresenting a test statistic; when Z is<At 0, it represents Δ TAiThe twin model precision shows an ascending trend along with the descending change of time; when Z is>When 0, it indicates Δ TAiThe accuracy of the twin model shows a descending trend along with the increasing change of time, and a digital twin model updating mechanism is triggered at the moment.
And S4, selecting DNN and LSTM as a base learner by adopting an integrated learning method based on the digital twin model updating mechanism of the step S3, carrying out difference comparison training on the base learner until a strong learner is formed, and finally updating the digital twin model. In particular, the amount of the solvent to be used,
s4.1, updating an Adaboost weight function; updating the weight function by adopting a time weighting method, which comprises the following specific steps:
ωt＝exp(-ηt)(t＝1,2,…,T；η∈[0,0.99])
wherein t represents the data acquisition time, namely the data batch number acquired at the time t; when the data batches are sequentially increased, t is increased, and the more new the data is, the stronger the importance is;
s4.2, updating an iterative weighting mechanism of the sample; the sample weight updating mechanism in the iterative process of training the base learner by the Adaboost-DNN-LSTM algorithm is as follows:
Figure BDA0003379144510000111
Figure BDA0003379144510000112
Figure BDA0003379144510000113
where U represents the total number of iterations of Adaboost-DNN-LSTM, m*Representing the total number of samples;
Figure BDA0003379144510000114
denotes the ith*The weight of the u +1 th iteration of a sample,
Figure BDA0003379144510000115
is a parameter for controlling the sample weight adjustment direction; xi is a threshold value of the model error, when the error exceeds the threshold value, the weight of the sample is increased, otherwise, the weight of the sample is reduced;
step S4.3, determining the weight alpha of the base learneru(ii) a By alphauIndicating the u-th basis learnerBy accuracy error ε of DNN, LSTM-based learner on data setuThe following calculation formula is adopted:
Figure BDA0003379144510000116
Figure BDA0003379144510000117
wherein
Figure BDA0003379144510000118
Denotes the ith*The number of samples of the input is,
Figure BDA0003379144510000119
an actual value that represents the progress of the production,
Figure BDA00033791445100001113
represents a production progress prediction value of the u-th base learner,
Figure BDA00033791445100001110
representing the predicted value of the production schedule of the u-th base learner DNN,
Figure BDA00033791445100001111
representing the predicted value of the production progress of the u-th base learner LSTM;
s4.4, integrating a base learner; and integrating the trained base learners according to the voting weight to obtain a final strong learner Adaboost-DNN-LSTM, wherein the integration formula is as follows:
Figure BDA00033791445100001112
finally, the strong learner is used for updating the performance model in the original digital twin model to complete dynamic self-adaptive correction.
The above description is only of the preferred embodiments of the present invention, and it should be noted that: it will be apparent to those skilled in the art that various modifications and adaptations can be made without departing from the principles of the invention and these are intended to be within the scope of the invention.

Claims (7)

1. A self-adaptive dynamic updating method for a digital twin model of a discrete manufacturing workshop is characterized by comprising the following steps:
s1, collecting production data of a plurality of continuous production processes according to time sequence through an internet of things sensing device deployed in a digital twin manufacturing workshop, and selecting a characteristic data set; the characteristic data set comprises order task data, task data completed at the current moment, real-time production state data and predicted time; meanwhile, data acquisition is carried out on the virtual workshop to obtain a characteristic data set;
step S2, taking the square loss function as the response error of the performance target, and adopting a maximum and minimum normalization method to normalize the target response error and confirm the performance target needing to be corrected preferentially;
s3, performing significance test on the precision of the performance index by adopting a Mann-Kendall-based deviation trend analysis method; when the precision error exceeds a preset threshold value, the digital twin model is subjected to self-adaptive updating;
and S4, selecting DNN and LSTM as a base learner by adopting an integrated learning method based on the digital twin model updating mechanism of the step S3, carrying out difference comparison training on the base learner until a strong learner is formed, and finally updating the digital twin model.
2. The adaptive dynamic updating method for the digital twin model in the discrete manufacturing plant as claimed in claim 1, wherein step S2 includes two processes of calculating response error and selecting the performance target to be modified; in particular, the amount of the solvent to be used,
s2.1, constructing a performance target response error function; using a quadratic loss function
Figure FDA0003379144500000011
The target response error was constructed as follows:
Figure FDA0003379144500000012
wherein the content of the first and second substances,
Figure FDA0003379144500000013
actual value, y, representing the ith target response in the twin modeliF (x; i) represents a theoretical value of the ith target response in the twin model; based on the target response error function, completing the calculation of each performance target response error;
s2.2, selecting a performance target to be corrected; adopting maximum and minimum normalization method to process target response error ERi
Figure FDA0003379144500000014
Normalization processing is carried out, and the influence of different dimensions is eliminated; the calculation method is as follows:
Figure FDA0003379144500000015
wherein the content of the first and second substances,
Figure FDA0003379144500000016
absolute value of error Δ ER for target response pairijThe calculation formula of (2) is as follows:
△ERij＝|ERi-ERj|
i and j respectively represent different target responses contained in the target response pair; further defining the proportion of each target response to the error difference, namely the fractional coefficient w of the target response pairijThe calculation formula is as follows:
Figure FDA0003379144500000021
n represents the total number of target responses; w is aijThe larger the value of (b), the larger the response error indicating the ith and jth targets, the higher the priority of correction required.
3. The discrete manufacturing shop digital twin model adaptive dynamic updating method according to claim 1, wherein the Mann-Kendall-based deviation trend analysis method in the step S3 specifically comprises:
s3.1, obtaining the testing precision change of the digital twin model; the calculation method comprises the following steps:
Figure FDA0003379144500000022
wherein the content of the first and second substances,
Figure FDA0003379144500000023
an actual value representing the i-th index at time t;
Figure FDA0003379144500000024
a theoretical value of an ith index representing the time t under a digital twin model; the test precision of the model at the t +1 th moment is TAi t+1And then the testing precision of the digital twin model is as follows:
Figure FDA0003379144500000025
a complete digital twin model test accuracy change of the ith index is expressed as:
Figure FDA0003379144500000026
s3.2, adopting a Mann-Kendall trend testing method to test the testing precision delta TA of the digital twin modeliThe trend change condition of (1); test accuracy of digital twin modelWhen the descending trend is presented, triggering a digital twin model updating mechanism; specifically, the formula for the Mann-Kendall trend test is:
Figure FDA0003379144500000027
Figure FDA0003379144500000028
wherein T is the number of numerical values in the digital twin model test precision sequence, and S isi>When 0, the precision change is changed upwards; when S isi<When 0, the precision change is changed downwards; when T is>At 10, use of ZiStatistic amount, and calculate SiThe calculation formula is as follows:
Figure FDA0003379144500000031
Figure FDA0003379144500000032
wherein Var (S)i) Denotes SiG represents the number of groups of junctions, tj*Representing the number of values in the jth knot group; ziRepresenting a test statistic; when Z is<At 0, it represents Δ TAiThe twin model precision shows an ascending trend along with the descending change of time; when Z is>When 0, it indicates Δ TAiThe accuracy of the twin model shows a descending trend along with the increasing change of time, and a digital twin model updating mechanism is triggered at the moment.
4. The adaptive dynamic updating method for the digital twin model in the discrete manufacturing plant as claimed in claim 1, wherein the step of adaptively updating the digital twin model in step S4 comprises the following specific steps:
s4.1, updating an Adaboost weight function; updating the weight function by adopting a time weighting method, which comprises the following specific steps:
ωt＝exp(-ηt)(t＝1,2,…,T；η∈[0,0.99])
wherein t represents the data acquisition time, namely the data batch number acquired at the time t; when the data batches are sequentially increased, t is increased, and the more new the data is, the stronger the importance is;
s4.2, updating an iterative weighting mechanism of the sample; the sample weight updating mechanism in the iterative process of training the base learner by the Adaboost-DNN-LSTM algorithm is as follows:
Figure FDA0003379144500000033
Figure FDA0003379144500000034
Figure FDA0003379144500000041
where U represents the total number of iterations of Adaboost-DNN-LSTM, m*Representing the total number of samples;
Figure FDA0003379144500000042
denotes the ith*The weight of the u +1 th iteration of a sample,
Figure FDA0003379144500000043
is a parameter for controlling the sample weight adjustment direction; xi is a threshold value of the model error, when the error exceeds the threshold value, the weight of the sample is increased, otherwise, the weight of the sample is reduced;
step S4.3, determining the weight alpha of the base learneru(ii) a By alphauShows the voting weight of the u-th base learner, the accuracy error epsilon of the data set by the DNN and LSTM base learnersuThe following calculation formula is adopted:
Figure FDA0003379144500000044
s.t.
Figure FDA0003379144500000045
wherein
Figure FDA0003379144500000046
Denotes the ith*The number of samples of the input is,
Figure FDA0003379144500000047
an actual value that represents the progress of the production,
Figure FDA0003379144500000048
represents a production progress prediction value of the u-th base learner,
Figure FDA0003379144500000049
representing the predicted value of the production schedule of the u-th base learner DNN,
Figure FDA00033791445000000410
representing the predicted value of the production progress of the u-th base learner LSTM;
s4.4, integrating a base learner; and integrating the trained base learners according to the voting weight to obtain a final strong learner Adaboost-DNN-LSTM, wherein the integration formula is as follows:
Figure FDA00033791445000000411
finally, the strong learner is used for updating the performance model in the original digital twin model to complete dynamic self-adaptive correction.
5. The adaptive dynamic discrete manufacturing shop digital twin model updating method according to claim 1, wherein the digital twin manufacturing shop in step S1 includes a virtual production line, a physical production line, a simulation model, a logic model and a data model; the virtual production line is connected with the physical production line through a simulation model, a logic model and a data model, the simulation model is connected with the logic model, and the logic model is connected with the data model; the virtual production line is a digital mapping of the physical production line in a digital space; the simulation model is an artificial intelligence algorithm supporting the virtual production line to run in an actual production line state; the logic model is the operation logic of the actual production of the physical production line.
6. The discrete manufacturing plant digital twin model adaptive dynamic updating method according to claim 5, wherein the virtual production line comprises virtual production equipment, virtual conveying equipment, virtual monitoring equipment and other virtual equipment; the virtual production equipment, the virtual conveying equipment, the virtual monitoring equipment and other virtual equipment are respectively connected with the logic model and the data model; the physical production line comprises production equipment, conveying equipment, monitoring equipment and other equipment; the production equipment, the conveying equipment, the monitoring equipment and other equipment are respectively connected with the logic model and the data model; each device of the virtual production line corresponds to the same-name device on the physical production line respectively; the sensor is arranged on the virtual production line and is respectively connected with the production equipment, the conveying equipment, the monitoring equipment and other equipment.
7. The self-adaptive dynamic updating method of the digital twin model of the discrete manufacturing workshop as claimed in claim 6, wherein during the operation of the virtual production line, the virtual sensor collects data of the virtual production equipment in real time and transmits the data to the production line data model service platform, and the sensor reads the data from the production line data model service platform and acts on the production equipment; and the production equipment performs production action to obtain data of a physical production line, transmits the data of the physical production line to a production line data model service platform to be compared with the data of the virtual production line, and adjusts the path and the position of the virtual production equipment if the data of the production equipment is different from the data of the virtual production equipment so as to enable the operation results of the virtual production line and the physical production line to be consistent.
CN202111438858.2A 2021-11-29 2021-11-29 Self-adaptive dynamic updating method for digital twin model of discrete manufacturing workshop Pending CN114398049A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202111438858.2A CN114398049A (en) 2021-11-29 2021-11-29 Self-adaptive dynamic updating method for digital twin model of discrete manufacturing workshop

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202111438858.2A CN114398049A (en) 2021-11-29 2021-11-29 Self-adaptive dynamic updating method for digital twin model of discrete manufacturing workshop

Publications (1)

Publication Number Publication Date
CN114398049A true CN114398049A (en) 2022-04-26

Family

ID=81225887

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202111438858.2A Pending CN114398049A (en) 2021-11-29 2021-11-29 Self-adaptive dynamic updating method for digital twin model of discrete manufacturing workshop

Country Status (1)

Country Link
CN (1) CN114398049A (en)

Cited By (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114596919A (en) * 2022-05-10 2022-06-07 安徽瑞邦数科科技服务有限公司 Index prediction method and system and application thereof in phosphoric acid production
CN114880930A (en) * 2022-05-11 2022-08-09 北京半人科技有限公司 Entity equipment digital twin modeling simulation prediction method based on online learning
CN115544819A (en) * 2022-12-06 2022-12-30 网思科技股份有限公司 Digital twin modeling method, system and readable storage medium for maintenance station
CN116700168A (en) * 2023-06-02 2023-09-05 中国五洲工程设计集团有限公司 Virtual-real synchronization method and system for production line
CN117608241A (en) * 2024-01-24 2024-02-27 山东建筑大学 Method, system, device and medium for updating digital twin model of numerical control machine tool
CN117788721A (en) * 2024-02-26 2024-03-29 彩讯科技股份有限公司 Digital twin content display method, system and medium based on artificial intelligence
CN117788721B (en) * 2024-02-26 2024-05-14 彩讯科技股份有限公司 Digital twin content display method, system and medium based on artificial intelligence

Cited By (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114596919A (en) * 2022-05-10 2022-06-07 安徽瑞邦数科科技服务有限公司 Index prediction method and system and application thereof in phosphoric acid production
CN114880930A (en) * 2022-05-11 2022-08-09 北京半人科技有限公司 Entity equipment digital twin modeling simulation prediction method based on online learning
CN114880930B (en) * 2022-05-11 2023-08-15 北京半人科技有限公司 Online learning-based entity equipment digital twin modeling simulation prediction method
CN115544819A (en) * 2022-12-06 2022-12-30 网思科技股份有限公司 Digital twin modeling method, system and readable storage medium for maintenance station
CN116700168A (en) * 2023-06-02 2023-09-05 中国五洲工程设计集团有限公司 Virtual-real synchronization method and system for production line
CN117608241A (en) * 2024-01-24 2024-02-27 山东建筑大学 Method, system, device and medium for updating digital twin model of numerical control machine tool
CN117608241B (en) * 2024-01-24 2024-04-05 山东建筑大学 Method, system, device and medium for updating digital twin model of numerical control machine tool
CN117788721A (en) * 2024-02-26 2024-03-29 彩讯科技股份有限公司 Digital twin content display method, system and medium based on artificial intelligence
CN117788721B (en) * 2024-02-26 2024-05-14 彩讯科技股份有限公司 Digital twin content display method, system and medium based on artificial intelligence

Similar Documents

Publication Publication Date Title
CN114398049A (en) Self-adaptive dynamic updating method for digital twin model of discrete manufacturing workshop
CN111590581B (en) Positioning compensation method and device for robot
CN109472057B (en) Product processing quality prediction device and method based on cross-process implicit parameter memory
CN106933105B (en) Track under confined condition updates integrated forecasting Iterative Learning Control Algorithm
CN110795780B (en) XGboost algorithm-based cable-stayed bridge finite element correction method
CN110571792A (en) Analysis and evaluation method and system for operation state of power grid regulation and control system
CN111638034B (en) Strain balance temperature gradient error compensation method and system based on deep learning
CN110263429B (en) Method for adjusting shape and surface precision of cable net reflecting surface antenna based on machine learning
US20230140696A1 (en) Method and system for optimizing parameter intervals of manufacturing processes based on prediction intervals
WO2021059787A1 (en) Control device, learning device, control method, learning method, and program
CN114800500B (en) Flexible constant force control method and system for polishing robot
CN108846200B (en) Quasi-static bridge influence line identification method based on iteration method
CN117033990A (en) Photovoltaic array fault feature extraction and modeling method
CN116595758A (en) Parameterized design system and parameterized design method for robot research and development application
CN114964313A (en) RVM-based fiber optic gyroscope temperature compensation method
CN101285816A (en) Copper matte air refining procedure parameter soft sensing instrument and its soft sensing method
CN112541296A (en) SO2 prediction method based on PSO-LSSVM
CN109187898B (en) Soft measurement method and device for ammonia nitrogen content of water in aquaculture environment
CN115206455B (en) Deep neural network-based rare earth element component content prediction method and system
JPH06332506A (en) Nonlinear controller
CN115526276A (en) Wind tunnel balance calibration load prediction method with robustness
CN114971375A (en) Examination data processing method, device, equipment and medium based on artificial intelligence
CN113807005A (en) Bearing residual life prediction method based on improved FPA-DBN
EP3889705A1 (en) Reduction of friction within a machine tool
CN114814707A (en) Intelligent ammeter stress error analysis method, equipment, terminal and readable medium

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