CN111695780B - Process flow quality multi-fault autonomous detection method and system - Google Patents
Process flow quality multi-fault autonomous detection method and system Download PDFInfo
- Publication number
- CN111695780B CN111695780B CN202010419509.5A CN202010419509A CN111695780B CN 111695780 B CN111695780 B CN 111695780B CN 202010419509 A CN202010419509 A CN 202010419509A CN 111695780 B CN111695780 B CN 111695780B
- Authority
- CN
- China
- Prior art keywords
- quality
- information
- subsystem
- fault
- flow
- 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
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
- G06Q10/0633—Workflow analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N20/00—Machine learning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/067—Enterprise or organisation modelling
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Systems or methods specially adapted for specific business sectors, e.g. utilities or tourism
- G06Q50/04—Manufacturing
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
- Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
- Y02P90/30—Computing systems specially adapted for manufacturing
Abstract
The invention provides a process flow quality multi-fault autonomous detection method and a system, wherein the method comprises the following steps: decomposing the process flow under the constraint of mass flow, and dividing the whole flow into a plurality of subsystems; constructing a quality index set and a quality index inverse mapping model which are sensitive to the quality fluctuation of the whole process in each subsystem; establishing a subsystem quality multi-fault detection model fusing hierarchy information to realize subsystem quality multi-fault autonomous detection; and fusing the quality multi-fault autonomous detection information of different subsystems of the process flow to realize the full-flow quality multi-fault autonomous detection fusing the local correlation information. The invention can effectively monitor and judge the quality faults of the production process in time and accurately; the method aims to solve the key challenge problem in the modern process industrial quality multi-fault detection, explores a practical and effective solution strategy, and has important practical application and popularization values.
Description
Technical Field
The invention relates to the technical field of control and monitoring of a production process, in particular to a method and a system for autonomously detecting multiple faults of process flow quality.
Background
The hot continuous rolling process of the strip steel is long in flow, multivariable coupling in the process, mass heredity in the process, multiple system levels and multiple quality indexes, and the production process relates to gas-liquid-solid multiphase coexistence of complex physical and chemical reactions.
The hot continuous rolling production process of the strip steel mainly comprises a plurality of production processes such as heating, rough rolling, flying shear, finish rolling, laminar flow, coiling and the like, and a long product processing flow taking a series structure as a main body is formed from raw materials to a final product; meanwhile, the corresponding integrated automation system has obvious hierarchy and mainly comprises an equipment layer, a real-time control layer, a process control layer, a manufacturing execution layer and the like. The processes of all the layers are definite and are mutually associated in a cooperative manner, and in addition, the raw material components, the equipment state, the process parameters, the product quality and the like of the layers cannot be comprehensively perceived in real time, so that the safety and stability analysis is complex and changeable, fault propagation and even evolution can be caused by any one or more link abnormalities, the production halt and maintenance of enterprises due to the return of goods of quality objection users are caused, and the economic benefit of the enterprises is influenced.
The continuous uninterrupted operation of the process industry described above makes it possible for a failure of any unit or subsystem to propagate and evolve at different levels through mass flow, energy flow, and information flow, resulting in the simultaneous or sequential occurrence of two or more failures, which may seriously affect the stable operation of the production process and the product quality. The characteristics of propagation, coupling, multiple concurrency and the like of the quality multiple faults and the complex mapping relation of one-to-many, many-to-one and many-to-many between the faults and symptoms make the quality multiple fault diagnosis of the process industrial process become a comprehensive and complex problem. Therefore, in the face of increasingly intense market competition, ensuring high-quality and high-efficiency operation of the process industry through a reasonable quality multi-fault diagnosis technology becomes an important content in sustainable development of the manufacturing industry.
Disclosure of Invention
The technical problem to be solved by the invention is to provide a process flow quality multi-fault autonomous detection method and system, so as to solve the quality multi-fault detection problem of the process industry and realize the effect of timely and accurately carrying out effective monitoring and judgment on the quality multi-fault of the production process.
In order to solve the technical problems, the invention provides the following technical scheme:
a process flow quality multi-fault autonomous detection method comprises the following steps:
decomposing the process flow to be detected under the constraint of mass flow based on the process knowledge and historical data of the process flow to be detected, and dividing the whole process into a plurality of subsystems;
constructing a quality index set sensitive to the quality fluctuation of the whole process in each subsystem, and preprocessing quality index data in the quality index set to construct a quality index inverse mapping model;
establishing a subsystem quality multi-fault detection model fusing level information based on a system level information and quality index inverse mapping model of a process flow to be detected, and realizing subsystem quality multi-fault autonomous detection;
and fusing the quality multi-fault autonomous detection information of different subsystems of the process flow to be detected, and realizing the full-process quality multi-fault autonomous detection of the process flow fusing the local correlation information.
Further, the decomposing the process flow to be detected under the constraint of mass flow based on the process knowledge and the historical data of the process flow to be detected, and dividing the whole process into a plurality of subsystems comprises:
analyzing the association and propagation relation of quality information among process variables by using process knowledge and historical data of a process to be detected and taking control association, reaction association, position association, type association, structure association and function association as a reference sequence, and determining a variable candidate set of the whole-process quality flow information;
determining a subsystem dominant variable based on the candidate set of variables;
measuring the correlation between the residual variables and the dominant variables in the variable candidate set by utilizing a statistical correlation analysis method, and determining auxiliary variables; and with the main variable as the center, gathering the auxiliary variables around the main variable with the correlation strength larger than a preset threshold value, and realizing the primary decomposition of the whole process;
and evaluating the rationality of the preliminary decomposition result based on the hierarchical information of the process flow, adjusting the preset threshold value, realizing the fine decomposition of the whole flow under the constraint of mass flow, and dividing the whole flow into a plurality of subsystems.
Further, the process knowledge includes hierarchical information, process mechanism and expert knowledge of the process flow to be detected.
Further, the determining a subsystem dominant variable based on the candidate set of variables includes:
aiming at a subsystem with a clear process mechanism, determining a dominant variable of the subsystem by using priori knowledge; and aiming at subsystems with unclear process mechanisms, determining dominant variables of the subsystems by using a partial least square or causal relationship analysis method.
Further, a quality index set sensitive to the quality fluctuation of the whole process in each subsystem is constructed, and quality index data in the quality index set is preprocessed to construct a quality index inverse mapping model, which comprises the following steps:
analyzing the sensitivity of each subsystem process variable to the typical quality fluctuation of the whole process;
based on the sensitivity analysis result, constructing a quality index set which is sensitive to the quality fluctuation of the whole process in each subsystem, and combining a process model of the process to be detected to realize the optimization of the quality index set;
preprocessing the quality index data in the optimized quality index set;
based on the preprocessed quality index data, determining the qualitative relation between the hierarchical characteristics of the quality information and the quality information among the hierarchies by using process knowledge, mining the quantitative quality information common among the hierarchies and specific in the hierarchies by adopting a dynamic time window and multivariate data modeling method, and realizing the analysis of the process and the quality data;
a method combining semi-supervised learning and multivariate regression is utilized to construct a high-dimensionality and multi-quality index inverse mapping model taking information depth perception as a characteristic.
Further, analyzing the sensitivity of each subsystem process variable to the typical quality fluctuation of the whole process flow, comprising:
and analyzing the sensitivity of the process variable of each subsystem to the typical quality fluctuation of the whole process by using variance decomposition, sensitivity analysis based on partial derivatives and sensitivity analysis based on connection weights.
Further, the pre-processing of the quality index data in the preferred quality index set includes:
analyzing the change rule of accumulated error and accuracy under the condition of cross-scale by adopting a cross-scale push-up and push-down method of multi-scale data mining, and realizing the synchronization of the process and the mass variable measurement scale;
the semi-supervised instant learning method is adopted to preprocess the process data with missing values and random noise, so that the redundancy of the data is reduced; on the basis of defining the standard running track of each production line, the starting time of running data of each batch is synchronized, and running data of equal-length batches is obtained by adopting an accelerated dynamic time warping method.
Furthermore, a subsystem quality multi-fault detection model fusing level information is established based on a system level information and quality index inverse mapping model of the process flow to be detected, so that subsystem quality multi-fault autonomous detection is realized; the method comprises the following steps:
based on system level information and a quality index inverse mapping model of a process flow to be detected, synchronizing multi-level information scales by using a scale synchronization method, and estimating a non-real-time measurement quality index by using a soft sensing model;
aiming at multi-level quality information, constructing a high-dimensional data tensor by the quality information in a system layered structure, and extracting the common characteristics between levels and the individual characteristic information in the levels in the process flow by adopting a tensor decomposition technology;
establishing a subsystem quality multi-fault detection model fusing hierarchy information by comprehensively considering time-varying, propagation, coupling and multiple concurrency characteristics of quality multi-fault according to the hierarchy level commonality characteristics and the hierarchy level individual characteristic information and adopting a variational reasoning and deep neural network method;
and based on the subsystem quality multi-fault detection model, a residual error interval is generated on line by using a centrosymmetric multi-cell technology and is used as a dynamic control limit of multi-fault detection, so that the subsystem quality multi-fault autonomous detection is realized.
Further, the fusion of the quality multi-fault autonomous detection information of different subsystems of the process flow to be detected to realize the full-process quality multi-fault autonomous detection of the process flow fusing the local correlation information includes:
and excavating static association and dynamic cooperation relation among the subsystems, fusing the quality multi-fault autonomous detection information of different subsystems through a preset information fusion and machine learning method, and realizing full-process quality multi-fault autonomous detection of the process flow to be detected by fusing local association information.
Accordingly, in order to solve the above technical problems, the present invention further provides the following technical solutions:
an autonomous detection system for multiple faults in process flow quality, the system comprising:
the flow decomposition module is used for decomposing the process flow to be detected under the mass flow constraint based on the process knowledge and the historical data of the process flow to be detected, and dividing the whole flow into a plurality of subsystems;
the quality index inverse mapping model building module is used for building a quality index set sensitive to full-process quality fluctuation in each subsystem and preprocessing quality index data to build a quality index inverse mapping model;
the subsystem quality multi-fault autonomous detection module fusing the hierarchical information is used for establishing a subsystem quality multi-fault detection model fusing the hierarchical information based on the system hierarchical information and the quality index inverse mapping model of the process flow to be detected, so as to realize subsystem quality multi-fault autonomous detection;
the process flow full-flow quality multi-fault autonomous detection module fusing the local correlation information is used for fusing the quality multi-fault autonomous detection information of different subsystems of the process flow to be detected, and realizing the process flow full-flow quality multi-fault autonomous detection fusing the local correlation information.
The technical scheme of the invention has the following beneficial effects:
the invention comprehensively considers the hierarchical information of the process industrial automation system, and realizes the full-process refined decomposition, the optimization of the quality index set and the construction of the inverse mapping model under the constraint of the quality flow, the autonomous detection of the quality multiple faults of the subsystem fusing the hierarchical information and the autonomous detection of the quality multiple faults of the full-process fusing the local associated information. On the basis of methods such as inverse mapping construction, multivariate statistics, variational Bayesian inference, ensemble learning and the like, an autonomous detection method suitable for the process industry is provided, the method aims at solving key challenge problems in the quality multi-fault detection of the modern process industry, explores practical and effective solution strategies, and has important practical application and popularization values.
Drawings
Fig. 1 is a flow chart of a process flow quality multi-fault autonomous detection method according to a first embodiment of the present invention;
FIG. 2 is a schematic diagram of the process layout of the hot continuous rolling production process of the strip steel;
fig. 3 is a schematic diagram of an implementation route of a multi-fault autonomous detection method for hot continuous rolling quality of strip steel according to a second embodiment of the present invention.
Detailed Description
In order to make the technical problems, technical solutions and advantages of the present invention more apparent, the following detailed description is given with reference to the accompanying drawings and specific embodiments.
First embodiment
Referring to fig. 1, the present embodiment provides a method for autonomously detecting multiple process quality faults, including:
s101, decomposing the process flow to be detected under the mass flow constraint based on the process knowledge and the historical data of the process flow to be detected, and dividing the whole process into a plurality of subsystems;
specifically, in this embodiment, the implementation process of the above steps is specifically as follows:
firstly, analyzing the association and propagation relation of quality information among process flow variables by using process knowledge (hierarchical information, process mechanism and expert knowledge) and historical data of a process flow to be detected and taking control association, reaction association, position association, type association, structure association and function association as a reference sequence, and determining a variable candidate set of the whole-flow quality flow information; aiming at a subsystem with a clear process mechanism, determining a dominant variable of the subsystem by using prior knowledge; and aiming at subsystems with unclear process mechanisms, determining dominant variables of the subsystems by using a partial least square or causal relationship analysis method. Measuring the correlation between the residual variables and the dominant variables in the variable candidate set by using a statistical correlation analysis method, and determining auxiliary variables; and with the main variable as the center, gathering the auxiliary variables around the main variable with the correlation strength larger than a preset threshold value, and realizing the primary decomposition of the whole process; and then, evaluating the rationality of the preliminary decomposition result based on the hierarchical information of the process flow, adjusting a preset threshold value, realizing the refined decomposition of the whole flow under the constraint of mass flow, and dividing the whole flow into a plurality of subsystems.
S102, constructing a quality index set sensitive to the quality fluctuation of the whole process in each subsystem, and preprocessing quality index data in the quality index set to construct a quality index inverse mapping model;
specifically, in this embodiment, the implementation process of the above steps is specifically as follows:
analyzing the sensitivity of each subsystem process variable to the typical quality fluctuation of the whole process by using a variance decomposition method, a sensitivity analysis method based on partial derivatives and a sensitivity analysis method based on connection weights;
based on the sensitivity analysis result, constructing a quality index set which is sensitive to the quality fluctuation of the whole process in each subsystem, and combining a process model of the process to be detected to realize the optimization of the quality index set;
preprocessing the quality index data in the optimized quality index set;
based on the preprocessed quality index data, determining the qualitative relation between the hierarchical characteristics of quality information and the quality information among the hierarchies by using process knowledge, and mining the quantitative quality information common among the hierarchies and specific in the hierarchies by adopting a dynamic time window and multivariate data modeling method to realize the analysis of the process and the quality data;
and constructing a high-dimensionality and multi-quality index inverse mapping model by using a method of combining semi-supervised learning and multivariate regression and taking information depth perception as characteristics.
The method for preprocessing the quality index data in the optimized quality index set comprises the following steps:
analyzing the change rule of accumulated error and accuracy under the condition of cross-scale by adopting a cross-scale push-up and push-down method of multi-scale data mining, and realizing the synchronization of the process and the mass variable measurement scale;
the semi-supervised instant learning method is adopted to preprocess the process data with missing values and random noise, so that the redundancy of the data is reduced; on the basis of defining the standard running track of each production line, the starting time of running data of each batch is synchronized, and running data of equal-length batches is obtained by adopting an accelerated dynamic time warping method.
S103, establishing a subsystem quality multi-fault detection model fusing level information based on a system level information and quality index inverse mapping model of the process flow to be detected, and realizing subsystem quality multi-fault autonomous detection;
specifically, in this embodiment, the implementation process of the above steps is specifically as follows:
based on system level information and a quality index inverse mapping model of a process flow to be detected, synchronizing multi-level information scales by using a scale synchronization method, and estimating a non-real-time measurement quality index by using a soft sensing model;
aiming at multi-level quality information, constructing a high-dimensional data tensor by the quality information in a system layered structure, and extracting the common characteristics between levels and the individual characteristic information in the levels in the process flow by adopting a tensor decomposition technology;
establishing a subsystem quality multi-fault detection model fusing hierarchy information by comprehensively considering time-varying, propagation, coupling and multiple concurrency characteristics of quality multi-fault according to the hierarchy level commonality characteristics and the hierarchy level individual characteristic information and adopting a variational reasoning and deep neural network method;
based on a subsystem quality multi-fault detection model, a central symmetry multi-cell technology is utilized to generate a residual error interval on line to serve as a dynamic control limit of multi-fault detection, and the subsystem quality multi-fault autonomous detection is realized.
And S104, fusing the quality multi-fault autonomous detection information of different subsystems of the process flow to be detected, and realizing the full-process quality multi-fault autonomous detection of the process flow fusing the local correlation information.
Specifically, in this embodiment, the implementation process of the above steps is specifically as follows:
and excavating static association and dynamic cooperation relation among the subsystems, fusing the quality multi-fault autonomous detection information of different subsystems through a preset information fusion and machine learning method, and realizing full-process quality multi-fault autonomous detection of the process flow to be detected by fusing local association information.
The method comprehensively considers the hierarchical information of the process industrial automation system, and realizes full-process refined decomposition, quality index set optimization and inverse mapping model construction, subsystem quality multi-fault autonomous detection of fusion hierarchical information, and full-process quality multi-fault autonomous detection of fusion local correlation information under the constraint of mass flow. On the basis of methods such as inverse mapping construction, multivariate statistics, variational Bayesian inference, ensemble learning and the like, an autonomous detection method suitable for the process industry is provided, the method aims at solving key challenge problems in process industry quality multi-fault detection, explores practical and effective solution strategies, and has important practical application and popularization values.
Second embodiment
Referring to fig. 2 and fig. 3, the embodiment further illustrates an embodiment of the method for automatically detecting multiple faults in process flow quality according to the present invention, by taking a data-driven hot continuous rolling production process of strip steel as an example; of course, it can be understood that the process flow quality multi-fault autonomous detection method is not limited to the strip steel hot continuous rolling process, and is also applicable to other data combined driven production processes, such as an automobile part production process.
As shown in figure 2, the production process of the hot continuous rolling of the strip steel has long process, multivariable coupling in the process, mass heredity in the process, multiple system levels and multiple quality indexes, and the production process relates to gas-liquid-solid multiphase coexistence of complex physical and chemical reactions. The production line comprises a plurality of production processes such as heating, rough rolling, flying shear, finish rolling, laminar flow, coiling and the like. The continuous uninterrupted operation of the complex production process enables any unit or subsystem to possibly propagate and evolve on different levels through mass flow, energy flow and information flow, so that the condition that two or more faults occur simultaneously or successively is generated, and serious influence is caused on the stable operation of the production process and the product quality.
As shown in fig. 3, on the basis of the practical engineering application driving of the hot continuous rolling of strip steel shown in fig. 1, the method for autonomously detecting multiple faults in process flow quality of the embodiment includes the following steps:
s1, carrying out fine decomposition on a full flow under the constraint of mass flow;
specifically, in this embodiment, the implementation process of the above steps is specifically as follows:
s101, on the basis of fully understanding the knowledge of the whole-flow process of the hot continuous rolling process of the strip steel, analyzing the structural characteristics of the whole-flow process of the hot continuous rolling process of the strip steel through a knowledge and data fusion technology, and researching a dominant-auxiliary variable selection method of each process under the mass flow constraint to divide the whole flow into a plurality of subsystems to realize the primary decomposition of the whole flow.
Firstly, analyzing the association and propagation relation of quality information among strip steel hot continuous rolling process variables by using process knowledge (hierarchical information, process mechanism, expert knowledge and the like) and taking control association, reaction association, position association, type association, structure association and function association as a reference sequence to determine a variable candidate set of the whole-process mass flow information;
then, aiming at a subsystem with a clear mechanism, determining a dominant variable by using prior knowledge; determining dominant variables by utilizing Partial Least Squares (PLS) and a causal relationship analysis method aiming at subsystems with uncertain mechanisms;
and finally, constructing main and auxiliary variable indexes by adopting statistical correlation analysis methods such as Mutual Information (MI), maximum Correlation Coefficient (MCC), maximum Information Coefficient (MIC) and the like, measuring the correlation between the residual variable and the dominant variable, and dividing the correlation into a far domain and a near domain, wherein the correlation between the variables in the near domain is strong, and the correlation between the variables in the far domain is weak. And (4) with the dominant variable as the center, gathering the auxiliary variables of the near-field variable around the dominant variable with strong correlation with the dominant variable, and realizing the primary decomposition of the whole process.
S102, comprehensively considering different levels of information such as production scheduling, tracking areas, process model setting and the like, researching the influence of the information on the full-process decomposition, and taking the influence as constraint guidance to realize the refined decomposition of the full-process of the hot continuous rolling of the strip steel under the constraint of mass flow.
The implementation process of the steps is as follows: the method mainly analyzes variables repeatedly covered by a plurality of local spaces, comprehensively considers different levels of information such as production scheduling, tracking areas, process model setting and the like and variable information not contained by any local space, evaluates the rationality of the preliminary decomposition result of the whole process of the hot continuous rolling of the strip steel by taking the variable information as a constraint condition, further adjusts a threshold value, and realizes the refined decomposition of the whole process under the constraint of mass flow.
And S2, constructing a quality index set optimization and inverse mapping model.
Specifically, in this embodiment, the implementation process of the above steps is specifically as follows:
s201, on the basis of the fine decomposition of the whole hot continuous rolling process of the strip steel, constructing a quality index set sensitive to the typical quality fluctuation of the whole process in each subsystem, and combining a process model, process knowledge and a data driving method to realize the optimization of the quality index set.
And analyzing the sensitivity of the process variables of each subsystem to the typical quality fluctuation of the whole process by methods of variance decomposition, sensitivity analysis based on partial derivatives, sensitivity analysis based on connection weights and the like, and realizing the optimization of a quality index set by combining a process model.
S202, different levels of information such as production scheduling, tracking areas and process model setting are comprehensively considered, the influence of the information on the whole process decomposition is researched, and the influence is taken as constraint guidance to realize the refined decomposition of the whole process of the hot continuous rolling of the strip steel under the constraint of mass flow.
Firstly, aiming at the multi-scale characteristics presented by the process and the quality variable of different levels and different procedures in the hot continuous rolling process of the strip steel, adopting a cross-scale push-up and push-down method of multi-scale data mining, analyzing the change rule of accumulated error and accuracy under the condition of cross-scale, and researching a process and quality variable measurement scale synchronization method; aiming at the problems of redundancy, uncertainty, missing values and the like easily occurring in the data of the hot continuous rolling process of the strip steel, the semi-supervised instant learning method is adopted to preprocess the process data with the missing values and random noise, so that the redundancy of the data is reduced; because the running tracks of different batches of the same strip steel hot continuous rolling production process are not completely consistent, on the basis of defining the running standard track of each production line, the starting time of the running data of each batch is synchronized, and the running data of equal-length batches is obtained by adopting an accelerated Dynamic Time Warping (DTW) method;
then, determining the hierarchical characteristics of the quality information and the qualitative relation of the quality information among the hierarchies by using process knowledge, mining the quantitative quality information common among the hierarchies and peculiar in the hierarchy by adopting a dynamic time window and multivariate data modeling method, and realizing the full-dimensional and multi-level intelligent analysis of the strip steel hot continuous rolling full-flow process and the quality data;
and finally, constructing a high-dimensionality and multi-quality index inverse mapping model by using a method combining semi-supervised learning and multivariate regression and taking information depth perception as characteristics, and laying a foundation for researches on aspects of quality multi-fault detection and the like of the strip steel hot continuous rolling process with level information.
And S3, fusing the multi-quality and multi-fault autonomous detection of the subsystem with hierarchical information.
Specifically, in this embodiment, the implementation process of the above steps is specifically as follows:
s301, on the basis of fully considering process knowledge, scheduling information, process data characteristics and different-level quality information interaction relations of subsystems in different levels, a unified representation method of multi-level quality information is researched, characteristics of propagation, coupling, multiple concurrency and the like of multiple quality faults are comprehensively considered, and a subsystem quality multiple fault detection model construction method fusing level information is researched.
Firstly, on the basis of full-flow refined decomposition under the constraint of mass flow, quality index set optimization and inverse mapping model construction, different levels of information such as production scheduling information, production process or control loop and the like in the hot continuous rolling process of strip steel are fully considered, a multi-level information scale is synchronized by using a scale synchronization method, and a key quality index which is difficult to measure or not measured in real time is estimated by using a soft sensing model;
then, constructing a high-dimensional data tensor by the quality information in a system layered structure aiming at multi-level quality information, and extracting the inter-level common characteristic and the intra-level individual characteristic information of the hot continuous rolling process of the strip steel by adopting tensor decomposition technologies such as CP decomposition, tucker decomposition and the like;
and finally, comprehensively considering the characteristics of time variation, propagation, coupling, multiple concurrency and the like of the quality multiple faults according to the inter-hierarchy commonality characteristics and the intra-hierarchy personality characteristic information, and establishing a subsystem quality multiple fault detection model fusing the hierarchy information by adopting a variational reasoning combined with a deep neural network method.
S302, researching self-adaptive detection statistics and a dynamic control limit design method based on the centrosymmetric multi-cell body technology, and realizing the quality multi-fault autonomous detection of the sub-system in the hot continuous rolling process of the strip steel.
On the basis of the constructed subsystem quality multi-fault detection model, a self-adaptive detection statistic design method is researched, the sensitivity and robustness of the designed detection model to the quality multi-fault are considered, the advantages of the centrosymmetric multi-cell technology in the aspects of uncertainty description, calculation and the like are fully utilized, a residual error interval is generated on line by utilizing the advantages, and the residual error interval is used as a dynamic control limit of multi-fault detection, so that the autonomy and the accuracy of the subsystem quality multi-fault real-time detection are improved.
And S4, fusing the full-process quality multi-fault autonomous detection of the local correlation information.
Specifically, in this embodiment, the implementation process of the above steps is specifically as follows:
and excavating static association and dynamic cooperation relation among subsystems, fusing the quality multi-fault autonomous detection information of different subsystems through information fusion and machine learning methods such as variational Bayesian inference, ensemble learning and the like, and realizing the full-process quality multi-fault autonomous detection of the full-process hot continuous rolling of the strip steel fusing local association information.
Firstly, aiming at the correlation and coupling characteristics among subsystems in the hot continuous rolling process of strip steel, analyzing the static correlation relationship among the subsystems by adopting methods such as typical correlation analysis (CCA), correlation projection analysis and the like, and analyzing the dynamic cooperation relationship among the subsystems by utilizing a method of combining dynamic CCA (DCCA) and Transfer Entropy (TE);
secondly, aiming at the problem of mutual information collaboration among subsystems in the process of hot continuous rolling of the strip steel, a distributed computing frame is utilized, a multivariable dimension reduction method is adopted, and relevant information of a correlation subsystem is compressed and transmitted to a target subsystem;
then, on the basis of the autonomous detection of the multiple faults of the quality of the subsystems fusing the hierarchical information, fusing the auxiliary information of the associated system and the monitoring information of the target subsystem to realize the cooperative detection of the multiple faults of the quality among the subsystems;
finally, as the model setting information, the scheduling information and the data characteristics of different subsystems are different, the designed detection statistics and dynamic control limits are different, and on the basis of the quality multi-fault cooperative detection among the subsystems, the self-adaptive detection statistics, the dynamic control limits and other information of the different subsystems are fused by methods such as integrated learning and variational Bayesian reasoning, so that the full-process quality multi-fault autonomous detection of the hot continuous rolling process of the strip steel with the fused local correlation information is realized.
The method for autonomously detecting the multiple faults of the hot strip rolling quality is provided by the embodiment, aims at the problems of quality safety and stability of products in the process industrial process such as the hot strip rolling, completely conforms to the multi-level and large-scale industrial characteristics and development direction of the process industry, has important scientific significance for preventing the quality reduction of the products and furthest exerting the process operation potential, becomes a research hotspot in the field of current process industrial process control, and has important theoretical value and wide engineering application prospect.
Third embodiment
The embodiment provides a many faults of process flow quality autonomous detection system, it includes:
the flow decomposition module is used for decomposing the process flow to be detected under the mass flow constraint based on the process knowledge and the historical data of the process flow to be detected, and dividing the whole flow into a plurality of subsystems;
the quality index inverse mapping model building module is used for building a quality index set sensitive to full-process quality fluctuation in each subsystem and preprocessing quality index data to build a quality index inverse mapping model;
the subsystem quality multi-fault autonomous detection module fusing the hierarchical information is used for establishing a subsystem quality multi-fault detection model fusing the hierarchical information based on the system hierarchical information and the quality index inverse mapping model of the process flow to be detected, so as to realize subsystem quality multi-fault autonomous detection;
the process flow full-flow quality multi-fault autonomous detection module fusing the local correlation information is used for fusing the quality multi-fault autonomous detection information of different subsystems of the process flow to be detected, and realizing the process flow full-flow quality multi-fault autonomous detection fusing the local correlation information.
The process flow quality multi-fault autonomous detection system of the embodiment corresponds to the process flow quality multi-fault autonomous detection method of the embodiment; the functions realized by the functional modules in the process flow quality multi-fault autonomous detection system of the embodiment correspond to the process steps in the process flow quality multi-fault autonomous detection method of the embodiment one by one; therefore, it is not described herein.
Furthermore, it should be appreciated by those skilled in the art that embodiments of the present invention may be provided as a method, apparatus, or computer program product. Accordingly, embodiments of the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, embodiments of the present invention may take the form of a computer program product embodied on one or more computer-usable storage media having computer-usable program code embodied in the medium.
Embodiments of the present invention are described with reference to flowchart illustrations and/or block diagrams of methods, terminal devices (systems), and computer program products according to embodiments of the invention. It will be understood that each flow and/or block of the flowchart illustrations and/or block diagrams, and combinations of flows and/or blocks in the flowchart illustrations and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, embedded processor, or other programmable data processing terminal to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing terminal, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing terminal to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks. These computer program instructions may also be loaded onto a computer or other programmable data processing terminal to cause a series of operational steps to be performed on the computer or other programmable terminal to produce a computer implemented process such that the instructions which execute on the computer or other programmable terminal provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
It should also be noted that, in this document, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or terminal that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or terminal. Without further limitation, an element defined by the phrases "comprising one of \ ...does not exclude the presence of additional like elements in a process, method, article, or terminal device that comprises the element.
Finally, it should be noted that while the above describes a preferred embodiment of the invention, it will be appreciated by those skilled in the art that, once the basic inventive concepts have been learned, numerous changes and modifications may be made without departing from the principles of the invention, which shall be deemed to be within the scope of the invention. Therefore, it is intended that the appended claims be interpreted as including the preferred embodiment and all changes and modifications that fall within the true scope of the embodiments of the present invention.
Claims (6)
1. A process flow quality multi-fault autonomous detection method is characterized by comprising the following steps:
decomposing the process flow to be detected under the constraint of mass flow based on the process knowledge and historical data of the process flow to be detected, and dividing the whole process into a plurality of subsystems;
constructing a quality index set sensitive to the quality fluctuation of the whole process in each subsystem, and preprocessing quality index data in the quality index set to construct a quality index inverse mapping model;
establishing a subsystem quality multi-fault detection model fusing level information based on a system level information and quality index inverse mapping model of a process flow to be detected, and realizing subsystem quality multi-fault autonomous detection;
fusing the quality multi-fault autonomous detection information of different subsystems of the process flow to be detected, and realizing the full-process quality multi-fault autonomous detection of the process flow fusing local correlation information;
the method for constructing the quality index set sensitive to the quality fluctuation of the whole process in each subsystem and preprocessing the quality index data in the quality index set to construct the quality index inverse mapping model comprises the following steps:
analyzing the sensitivity of each subsystem process variable to the typical quality fluctuation of the whole process;
based on the sensitivity analysis result, constructing a quality index set which is sensitive to the quality fluctuation of the whole process in each subsystem, and combining a process model of the process to be detected to realize the optimization of the quality index set;
preprocessing the quality index data in the optimized quality index set;
based on the preprocessed quality index data, determining the qualitative relation between the hierarchical characteristics of the quality information and the quality information among the hierarchies by using process knowledge, mining the quantitative quality information common among the hierarchies and specific in the hierarchies by adopting a dynamic time window and multivariate data modeling method, and realizing the analysis of the process and the quality data;
constructing a high-dimensionality and multi-quality index inverse mapping model by using a method of combining semi-supervised learning and multivariate regression, wherein the characteristic is information depth perception;
the analysis of the sensitivity of the process variables of each subsystem to the typical quality fluctuation of the whole process flow comprises the following steps:
analyzing the sensitivity of each subsystem process variable to the whole-process typical quality fluctuation by using a variance decomposition method, a sensitivity analysis based on partial derivatives and a sensitivity analysis method based on connection weights;
the preprocessing of the quality index data in the optimized quality index set includes:
analyzing the change rule of accumulated error and accuracy under the condition of cross-scale by adopting a cross-scale push-up and push-down method of multi-scale data mining, and realizing the synchronization of the process and the mass variable measurement scale;
the semi-supervised instant learning method is adopted to preprocess the process data with missing values and random noise, so that the redundancy of the data is reduced; on the basis of defining the standard running track of each production line, synchronizing the starting time of running data of each batch, and obtaining running data of equal-length batches by adopting an accelerated dynamic time warping method;
the subsystem quality multi-fault detection model fusing the hierarchy information is established based on the system hierarchy information and the quality index inverse mapping model of the process flow to be detected, and the subsystem quality multi-fault autonomous detection is realized; the method comprises the following steps:
based on system level information and a quality index inverse mapping model of a process flow to be detected, synchronizing multi-level information scales by using a scale synchronization method, and estimating a non-real-time measurement quality index by using a soft sensing model;
aiming at multi-level quality information, constructing a high-dimensional data tensor by the quality information in a system layered structure, and extracting the common characteristics between levels and the individual characteristic information in the levels in the process flow by adopting a tensor decomposition technology;
establishing a subsystem quality multi-fault detection model fusing hierarchy information by comprehensively considering time-varying, propagation, coupling and multiple concurrency characteristics of quality multi-fault according to the hierarchy level commonality characteristics and the hierarchy level individual characteristic information and adopting a variational reasoning and deep neural network method;
and based on the subsystem quality multi-fault detection model, a residual error interval is generated on line by using a centrosymmetric multi-cell technology and is used as a dynamic control limit of multi-fault detection, so that the subsystem quality multi-fault autonomous detection is realized.
2. The method for autonomously detecting multiple faults of process flow quality as claimed in claim 1, wherein the method for decomposing the process flow to be detected under the constraint of mass flow based on the process knowledge and historical data of the process flow to be detected, and dividing the whole process flow into a plurality of subsystems comprises:
analyzing the association and propagation relation of quality information among process variables by using process knowledge and historical data of a process to be detected and taking control association, reaction association, position association, type association, structure association and function association as a reference sequence, and determining a variable candidate set of the whole-process quality flow information;
determining a subsystem dominant variable based on the candidate set of variables;
measuring the correlation between the residual variables and the dominant variables in the variable candidate set by using a statistical correlation analysis method, and determining auxiliary variables; and with the main variable as the center, gathering the auxiliary variables around the main variable with the correlation strength larger than a preset threshold value, and realizing the primary decomposition of the whole process;
and evaluating the rationality of the preliminary decomposition result based on the hierarchical information of the process flow, adjusting the preset threshold value, realizing the refined decomposition of the whole flow under the constraint of mass flow, and dividing the whole flow into a plurality of subsystems.
3. The process flow quality multi-fault autonomous detection method of claim 2 wherein the process knowledge comprises hierarchical information, process mechanisms and expert knowledge of the process flow to be detected.
4. The process flow quality multi-fault autonomous detection method of claim 3 wherein said determining subsystem dominant variables based on said candidate set of variables comprises:
aiming at a subsystem with a clear process mechanism, determining a dominant variable of the subsystem by using priori knowledge; and aiming at subsystems with unclear process mechanisms, determining the dominant variables of the subsystems by using a partial least square or causal relationship analysis method.
5. The process flow quality multi-fault autonomous detection method of claim 1, wherein fusing the quality multi-fault autonomous detection information of different subsystems of the process flow to be detected to achieve process flow full-flow quality multi-fault autonomous detection fusing local correlation information, comprises:
and excavating static association and dynamic cooperation relation among the subsystems, fusing the quality multi-fault autonomous detection information of different subsystems through a preset information fusion and machine learning method, and realizing full-process quality multi-fault autonomous detection of the process flow to be detected by fusing local association information.
6. A process flow quality multi-fault autonomous detection system is characterized by comprising:
the flow decomposition module is used for decomposing the process flow to be detected under the mass flow constraint based on the process knowledge and the historical data of the process flow to be detected, and dividing the whole flow into a plurality of subsystems;
the quality index inverse mapping model building module is used for building a quality index set sensitive to full-process quality fluctuation in each subsystem and preprocessing quality index data to build a quality index inverse mapping model;
the subsystem quality multi-fault autonomous detection module fusing the hierarchical information is used for establishing a subsystem quality multi-fault detection model fusing the hierarchical information based on the system hierarchical information and the quality index inverse mapping model of the process flow to be detected, so as to realize subsystem quality multi-fault autonomous detection;
the process flow full-flow quality multi-fault autonomous detection module fusing the local correlation information is used for fusing the quality multi-fault autonomous detection information of different subsystems of the process flow to be detected, so as to realize the process flow full-flow quality multi-fault autonomous detection fusing the local correlation information;
constructing a quality index set sensitive to the quality fluctuation of the whole process in each subsystem and preprocessing quality index data to construct a quality index inverse mapping model, wherein the method comprises the following steps:
analyzing the sensitivity of each subsystem process variable to the typical quality fluctuation of the whole process;
based on the sensitivity analysis result, constructing a quality index set sensitive to the quality fluctuation of the whole process in each subsystem, and combining a process model of the process to be detected to realize the optimization of the quality index set;
preprocessing the quality index data in the optimized quality index set;
based on the preprocessed quality index data, determining the qualitative relation between the hierarchical characteristics of the quality information and the quality information among the hierarchies by using process knowledge, mining the quantitative quality information common among the hierarchies and specific in the hierarchies by adopting a dynamic time window and multivariate data modeling method, and realizing the analysis of the process and the quality data;
constructing a high-dimensionality and multi-quality index inverse mapping model by using a method of combining semi-supervised learning and multivariate regression, wherein the characteristic of the information depth perception is used as a characteristic;
the analysis of the sensitivity of the process variables of each subsystem to the typical quality fluctuation of the whole process flow comprises the following steps:
analyzing the sensitivity of each subsystem process variable to the whole-process typical quality fluctuation by using a variance decomposition method, a sensitivity analysis based on partial derivatives and a sensitivity analysis method based on connection weights;
the preprocessing of the quality index data in the optimized quality index set comprises the following steps:
analyzing the change rule of accumulated error and accuracy under the condition of cross-scale by adopting a cross-scale push-up and push-down method of multi-scale data mining, and realizing the synchronization of the process and the mass variable measurement scale;
the semi-supervised instant learning method is adopted to preprocess the process data with missing values and random noise, so that the redundancy of the data is reduced; on the basis of defining the standard running track of each production line, synchronizing the starting time of running data of each batch, and obtaining running data of equal-length batches by adopting an accelerated dynamic time warping method;
the subsystem quality multi-fault detection model fusing the hierarchy information is established based on the system hierarchy information and the quality index inverse mapping model of the process flow to be detected, and the subsystem quality multi-fault autonomous detection is realized; the method comprises the following steps:
based on system level information and a quality index inverse mapping model of a process flow to be detected, synchronizing multi-level information scales by using a scale synchronization method, and estimating a non-real-time measurement quality index by using a soft sensing model;
aiming at multi-level quality information, constructing a high-dimensional data tensor by the quality information in a system layered structure, and extracting the common characteristics between levels and the individual characteristic information in the levels in the process flow by adopting a tensor decomposition technology;
respectively aiming at the inter-level commonality characteristics and the intra-level personality characteristic information, comprehensively considering the time-varying, propagation, coupling and multiple concurrency characteristics of multiple faults of quality, and establishing a subsystem quality multiple fault detection model fusing level information by adopting a variational reasoning method and a deep neural network method;
and based on the subsystem quality multi-fault detection model, a residual error interval is generated on line by using a centrosymmetric multi-cell technology and is used as a dynamic control limit of multi-fault detection, so that the subsystem quality multi-fault autonomous detection is realized.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010419509.5A CN111695780B (en) | 2020-05-18 | 2020-05-18 | Process flow quality multi-fault autonomous detection method and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010419509.5A CN111695780B (en) | 2020-05-18 | 2020-05-18 | Process flow quality multi-fault autonomous detection method and system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111695780A CN111695780A (en) | 2020-09-22 |
| CN111695780B true CN111695780B (en) | 2023-04-18 |
Family
ID=72477887
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010419509.5A Active CN111695780B (en) | 2020-05-18 | 2020-05-18 | Process flow quality multi-fault autonomous detection method and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111695780B (en) |
Families Citing this family (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113835411B (en) * | 2021-09-07 | 2023-10-27 | 北京科技大学顺德研究生院 | Comprehensive diagnosis method for quality abnormality of steel rolling process flow |
| CN114235411B (en) * | 2021-12-28 | 2023-06-30 | 频率探索智能科技江苏有限公司 | Bearing outer ring defect positioning method |
| CN115780530B (en) * | 2023-01-19 | 2023-05-09 | 北京科技大学 | Steel rolling fault tracing method and device based on expert knowledge and data combined driving |
| CN116020879B (en) * | 2023-02-15 | 2023-06-16 | 北京科技大学 | Technological parameter-oriented strip steel hot continuous rolling space-time multi-scale process monitoring method and device |
| CN115879826B (en) * | 2023-02-20 | 2023-05-30 | 深圳普菲特信息科技股份有限公司 | Fine chemical process quality inspection method, system and medium based on big data |
| CN116627093B (en) * | 2023-04-19 | 2024-02-27 | 济南海马机械设计有限公司 | Nitrile glove processing control method, system, equipment and storage medium |
| CN117726240A (en) * | 2024-02-18 | 2024-03-19 | 中国标准化研究院 | Quality evaluation classification method and system based on convolutional neural network |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106355030A (en) * | 2016-09-20 | 2017-01-25 | 浙江大学 | Fault detection method based on analytic hierarchy process and weighted vote decision fusion |
| CN106354125A (en) * | 2016-11-09 | 2017-01-25 | 上海电机学院 | Method for utilizing block PCA (Principal Component Analysis) to detect fault of chemical process |
| CN106529079A (en) * | 2016-11-29 | 2017-03-22 | 上海电机学院 | Chemical process failure detection method based on failure-dependent principal component space |
| CN109656793A (en) * | 2018-11-22 | 2019-04-19 | 安徽继远软件有限公司 | A kind of information system performance stereoscopic monitoring method based on multi-source heterogeneous data fusion |
| CN109991951A (en) * | 2019-04-28 | 2019-07-09 | 齐鲁工业大学 | Multi-source fault detection and diagnosis method and apparatus |
| CN110333709A (en) * | 2019-06-19 | 2019-10-15 | 北京科技大学 | A kind of rolling process fault diagnosis method and system |
Family Cites Families (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20080275587A1 (en) * | 2002-09-25 | 2008-11-06 | Advanced Micro Devices, Inc. | Fault detection on a multivariate sub-model |
| US11348018B2 (en) * | 2017-12-19 | 2022-05-31 | Aspen Technology, Inc. | Computer system and method for building and deploying models predicting plant asset failure |
-
2020
- 2020-05-18 CN CN202010419509.5A patent/CN111695780B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106355030A (en) * | 2016-09-20 | 2017-01-25 | 浙江大学 | Fault detection method based on analytic hierarchy process and weighted vote decision fusion |
| CN106354125A (en) * | 2016-11-09 | 2017-01-25 | 上海电机学院 | Method for utilizing block PCA (Principal Component Analysis) to detect fault of chemical process |
| CN106529079A (en) * | 2016-11-29 | 2017-03-22 | 上海电机学院 | Chemical process failure detection method based on failure-dependent principal component space |
| CN109656793A (en) * | 2018-11-22 | 2019-04-19 | 安徽继远软件有限公司 | A kind of information system performance stereoscopic monitoring method based on multi-source heterogeneous data fusion |
| CN109991951A (en) * | 2019-04-28 | 2019-07-09 | 齐鲁工业大学 | Multi-source fault detection and diagnosis method and apparatus |
| CN110333709A (en) * | 2019-06-19 | 2019-10-15 | 北京科技大学 | A kind of rolling process fault diagnosis method and system |
Non-Patent Citations (1)
| Title |
|---|
| 刘白林 ; 任健强 ; 马颖 ; .结合知识推理的分层多信号流图在多故障诊断中的应用.西安工业大学学报.2016,(第03期),全文. * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111695780A (en) | 2020-09-22 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111695780B (en) | Process flow quality multi-fault autonomous detection method and system | |
| CN111724126A (en) | Accurate tracing method and system for process flow quality abnormity | |
| Zhang et al. | Digital twin enhanced dynamic job-shop scheduling | |
| Wang et al. | A digital twin-based big data virtual and real fusion learning reference framework supported by industrial internet towards smart manufacturing | |
| Menezes et al. | Predictive, prescriptive and detective analytics for smart manufacturing in the information age | |
| CN101923340B (en) | The method and apparatus of the procedure quality in forecasting process control system | |
| MacGregor et al. | Monitoring, fault diagnosis, fault-tolerant control and optimization: Data driven methods | |
| US10949740B2 (en) | Machine learning device, numerical controller, machine tool system, manufacturing system, and machine learning method for learning display of operation menu | |
| CN103116277A (en) | Displaying correlated factors and predicted faults | |
| KR101825881B1 (en) | Method of managing a manufacturing process and system using the same | |
| CN101408769A (en) | On-line energy forecasting system and method based on product ARIMA model | |
| Lai et al. | Data-driven dynamic bottleneck detection in complex manufacturing systems | |
| WO2020252666A1 (en) | Edge computing device and method for industrial internet of things, and computer-readable storage medium | |
| Choo et al. | Adaptive multi-scale prognostics and health management for smart manufacturing systems | |
| Qian et al. | Digital twin driven production progress prediction for discrete manufacturing workshop | |
| US20210325838A1 (en) | Prediction control development device, prediction control development method, and computer readable storage medium | |
| CN111340269B (en) | Real-time optimization method for process industrial process | |
| KR102474332B1 (en) | Prediction method of machine health stability in smart factory system, and recording medium thereof | |
| CN117393076B (en) | Intelligent monitoring method and system for heat-resistant epoxy resin production process | |
| Caiazzo et al. | An IoT-based and cloud-assisted AI-driven monitoring platform for smart manufacturing: design architecture and experimental validation | |
| CN116020879B (en) | Technological parameter-oriented strip steel hot continuous rolling space-time multi-scale process monitoring method and device | |
| Sun et al. | Prognostics and health management via long short-term digital twins | |
| Chen et al. | Exploring equipment electrocardiogram mechanism for performance degradation monitoring in smart manufacturing | |
| Karagiorgou et al. | Unveiling trends and predictions in digital factories | |
| Dong et al. | Hierarchical causal graph-based fault root cause diagnosis and propagation path identification for complex industrial process monitoring |
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 |