CN113610350B - Complex working condition fault diagnosis method, equipment, storage medium and device - Google Patents
Complex working condition fault diagnosis method, equipment, storage medium and device Download PDFInfo
- Publication number
- CN113610350B CN113610350B CN202110776972.XA CN202110776972A CN113610350B CN 113610350 B CN113610350 B CN 113610350B CN 202110776972 A CN202110776972 A CN 202110776972A CN 113610350 B CN113610350 B CN 113610350B
- Authority
- CN
- China
- Prior art keywords
- model
- procedure
- sub
- preset
- data
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 457
- 238000003745 diagnosis Methods 0.000 title claims abstract description 143
- 238000003860 storage Methods 0.000 title claims abstract description 14
- 238000004519 manufacturing process Methods 0.000 claims abstract description 144
- 238000013145 classification model Methods 0.000 claims abstract description 59
- 238000012549 training Methods 0.000 claims description 62
- 230000002159 abnormal effect Effects 0.000 claims description 39
- 238000012163 sequencing technique Methods 0.000 claims description 26
- 238000005070 sampling Methods 0.000 claims description 22
- 230000007717 exclusion Effects 0.000 claims description 10
- 238000002372 labelling Methods 0.000 claims description 9
- 238000012545 processing Methods 0.000 description 10
- 238000010586 diagram Methods 0.000 description 9
- 238000004891 communication Methods 0.000 description 5
- 238000010276 construction Methods 0.000 description 4
- 238000011156 evaluation Methods 0.000 description 4
- PCTMTFRHKVHKIS-BMFZQQSSSA-N (1s,3r,4e,6e,8e,10e,12e,14e,16e,18s,19r,20r,21s,25r,27r,30r,31r,33s,35r,37s,38r)-3-[(2r,3s,4s,5s,6r)-4-amino-3,5-dihydroxy-6-methyloxan-2-yl]oxy-19,25,27,30,31,33,35,37-octahydroxy-18,20,21-trimethyl-23-oxo-22,39-dioxabicyclo[33.3.1]nonatriaconta-4,6,8,10 Chemical compound C1C=C2C[C@@H](OS(O)(=O)=O)CC[C@]2(C)[C@@H]2[C@@H]1[C@@H]1CC[C@H]([C@H](C)CCCC(C)C)[C@@]1(C)CC2.O[C@H]1[C@@H](N)[C@H](O)[C@@H](C)O[C@H]1O[C@H]1/C=C/C=C/C=C/C=C/C=C/C=C/C=C/[C@H](C)[C@@H](O)[C@@H](C)[C@H](C)OC(=O)C[C@H](O)C[C@H](O)CC[C@@H](O)[C@H](O)C[C@H](O)C[C@](O)(C[C@H](O)[C@H]2C(O)=O)O[C@H]2C1 PCTMTFRHKVHKIS-BMFZQQSSSA-N 0.000 description 3
- 238000003066 decision tree Methods 0.000 description 3
- 230000003111 delayed effect Effects 0.000 description 3
- 230000006870 function Effects 0.000 description 3
- 238000004364 calculation method Methods 0.000 description 2
- 238000009826 distribution Methods 0.000 description 2
- 230000011218 segmentation Effects 0.000 description 2
- 238000003070 Statistical process control Methods 0.000 description 1
- 238000013075 data extraction Methods 0.000 description 1
- 230000001419 dependent effect Effects 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000005192 partition Methods 0.000 description 1
- 238000004886 process control Methods 0.000 description 1
- 238000000638 solvent extraction Methods 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/06—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
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/04—Manufacturing
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
- Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
- Y02P90/30—Computing systems specially adapted for manufacturing
Landscapes
- Engineering & Computer Science (AREA)
- Business, Economics & Management (AREA)
- Theoretical Computer Science (AREA)
- Human Resources & Organizations (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- General Physics & Mathematics (AREA)
- Strategic Management (AREA)
- Economics (AREA)
- General Business, Economics & Management (AREA)
- Tourism & Hospitality (AREA)
- Marketing (AREA)
- Bioinformatics & Computational Biology (AREA)
- Entrepreneurship & Innovation (AREA)
- Bioinformatics & Cheminformatics (AREA)
- General Engineering & Computer Science (AREA)
- Evolutionary Biology (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Computation (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Health & Medical Sciences (AREA)
- General Health & Medical Sciences (AREA)
- General Factory Administration (AREA)
- Development Economics (AREA)
- Educational Administration (AREA)
- Manufacturing & Machinery (AREA)
- Game Theory and Decision Science (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Primary Health Care (AREA)
Abstract
The invention discloses a complex working condition fault diagnosis method, equipment, a storage medium and a device, wherein the invention diagnoses manufacturing data of application equipment through a preset optimal performance sub-model and a preset classification model to obtain a diagnosis result; and determining a fault path corresponding to the problem procedure information contained in the manufacturing data according to a preset association algorithm and a diagnosis result. The invention diagnoses the manufacturing data through the preset optimal performance sub-model and the preset classification model, and determines the fault path corresponding to the problem procedure information contained in the manufacturing data according to the preset association algorithm and the diagnosis result.
Description
Technical Field
The present invention relates to the field of fault diagnosis, and in particular, to a method, an apparatus, a storage medium, and a device for diagnosing a fault under a complex working condition.
Background
At present, in the production and processing process of products, certain fluctuation of the specifications of the products, such as the dimensions, can occur for some reasons, the fluctuation has a great influence on the quality of the products, but the influence caused by the fluctuation can be completely avoided and eliminated by taking measures, and the measures are process control. In the traditional statistical process control method, quality diagnosis in the manufacturing process is mostly dependent on experience of operators, and certain requirements are met on the professional level and the service life of diagnosticians, so that the diagnosis accuracy varies from person to person, and the diagnosis result is inaccurate.
The foregoing is provided merely for the purpose of facilitating understanding of the technical solutions of the present invention and is not intended to represent an admission that the foregoing is prior art.
Disclosure of Invention
The invention mainly aims to provide a complex working condition fault diagnosis method, equipment, a storage medium and a device, and aims to solve the technical problem that in the prior art, diagnosis is carried out by relying on experience of an operator, so that a diagnosis result is inaccurate.
In order to achieve the above object, the present invention provides a method for diagnosing a fault in a complex working condition, the method for diagnosing a fault in a complex working condition comprising the steps of:
Acquiring manufacturing data of application equipment;
diagnosing the manufacturing data according to a preset optimal performance sub-model and a preset classification model to obtain a diagnosis result;
And determining problem procedure information contained in the manufacturing data according to a preset association algorithm and the diagnosis result, and determining a corresponding fault path according to the problem procedure information.
Preferably, the step of diagnosing the manufacturing data according to a preset optimal performance sub-model and a preset classification model to obtain a diagnosis result includes:
Dividing the original working procedure corresponding to the manufacturing data according to a preset optimal performance sub-model to obtain the duty ratio of the problem working procedure in the original working procedure;
determining a diagnostic priority based on the duty cycle;
and diagnosing the manufacturing data according to a preset classification model and the diagnosis priority to obtain a diagnosis result.
Preferably, the step of determining problem process information included in the manufacturing data according to a preset association algorithm and the diagnosis result, and determining a corresponding fault path according to the problem process information includes:
Determining a frequent item set corresponding to a problem procedure in the manufacturing data according to a preset association algorithm and the diagnosis result;
and determining a fault path corresponding to the problem procedure according to the frequent item set.
Preferably, before the step of acquiring manufacturing data of the application device, the method includes:
Acquiring historical process manufacturing data;
performing abnormal labeling on the process corresponding to the historical process manufacturing data to obtain labeled abnormal process and abnormal process data;
Constructing a problem procedure data set according to the marked abnormal procedure and the marked abnormal procedure data;
Training a model based on LightGBM algorithm according to the problem procedure data set to obtain an initial classification model;
Training the initial classification model according to the problem procedure data set to obtain a preset classification model;
Training a preset granularity-adjustable hierarchical model according to the historical procedure manufacturing data to obtain an initial optimal performance sub-model;
Training the initial optimal performance sub-model according to the historical procedure manufacturing data to obtain a preset optimal performance sub-model.
Preferably, the model based on LightGBM algorithm comprises a model based on single-side sampling algorithm and a model based on mutual exclusion feature algorithm;
The step of training a LightGBM algorithm-based model according to the problem procedure dataset to obtain an initial classification model includes:
Training the model based on the single-side sampling algorithm according to the problem procedure data set to obtain an initial learner;
Performing iterative training on the initial learner to obtain a preset learner;
Training the model based on the mutual exclusion feature algorithm according to the problem procedure data set to obtain a feature binding model;
And constructing an initial classification model according to the preset learner and the characteristic binding model.
Preferably, the step of training the model with adjustable preset granularity according to the historical procedure manufacturing data to obtain an initial optimal performance sub-model includes:
Training a hierarchical model with adjustable preset granularity according to the historical procedure manufacturing data to obtain different hierarchical sub-models;
evaluating each level sub-model to obtain an optimal sequencing sub-model;
training each optimal sequencing sub-model to obtain the occurrence frequency of each optimal sequencing sub-model;
and selecting an optimal performance sub-model from all the optimal sequencing sub-models according to the occurrence frequency, and taking the optimal performance sub-model as an initial optimal performance sub-model.
Preferably, the step of evaluating each hierarchical sub-model to obtain an optimal ranking sub-model includes:
Determining the number of process groups containing problem processes, the number of working procedures in a single process group containing problem processes and the number of problem processes in a single process group containing problem processes according to the historical process manufacturing data;
And evaluating each level submodel according to the number of the process groups containing the problem process, the number of the processes in the single process group containing the problem process and the number of the problem process in the single process group containing the problem process so as to obtain an optimal sequencing submodel.
In addition, in order to achieve the above object, the present invention also proposes a complex operating condition fault diagnosis apparatus including a memory, a processor, and a complex operating condition fault diagnosis program stored on the memory and operable on the processor, the complex operating condition fault diagnosis program being configured to implement the steps of the complex operating condition fault diagnosis method as described above.
In addition, in order to achieve the above object, the present invention also proposes a storage medium having stored thereon a complex operating condition fault diagnosis program which, when executed by a processor, implements the steps of the complex operating condition fault diagnosis method as described above.
In addition, in order to achieve the above object, the present invention also provides a complex operating condition fault diagnosis device, including:
the data acquisition module is used for acquiring manufacturing data of the application equipment;
the data diagnosis module is used for diagnosing the manufacturing data according to a preset optimal performance sub-model and a preset classification model so as to obtain a diagnosis result;
And the path determining module is used for determining problem procedure information contained in the manufacturing data according to a preset association algorithm and the diagnosis result, and determining a corresponding fault path according to the problem procedure information.
The invention obtains the manufacturing data of the application equipment; diagnosing the manufacturing data according to the preset optimal performance sub-model and the preset classification model to obtain a diagnosis result; and determining problem procedure information contained in the manufacturing data according to a preset association algorithm and a diagnosis result, and determining a corresponding fault path according to the problem procedure information. Because the invention diagnoses the manufacturing data through the preset optimal performance sub-model and the preset classification model, the problem procedure information contained in the manufacturing data is determined according to the preset association algorithm and the diagnosis result, and the corresponding fault path is determined according to the problem procedure information.
Drawings
FIG. 1 is a schematic diagram of a complex operating condition fault diagnosis device for a hardware operating environment according to an embodiment of the present invention;
FIG. 2 is a schematic flow chart of a first embodiment of a fault diagnosis method for complex working conditions according to the present invention;
FIG. 3 is a schematic flow chart of a second embodiment of the fault diagnosis method for complex working conditions according to the present invention;
FIG. 4 is a schematic flow chart of a third embodiment of a fault diagnosis method for complex working conditions according to the present invention;
FIG. 5 is a schematic diagram of a process topology division of a third embodiment of a fault diagnosis method for complex operating conditions according to the present invention;
FIG. 6 is a block diagram of a first embodiment of a complex operating condition fault diagnosis apparatus according to the present invention.
The achievement of the objects, functional features and advantages of the present invention will be further described with reference to the accompanying drawings, in conjunction with the embodiments.
Detailed Description
It should be understood that the specific embodiments described herein are for purposes of illustration only and are not intended to limit the scope of the invention.
Referring to fig. 1, fig. 1 is a schematic structural diagram of a complex operating condition fault diagnosis device of a hardware operating environment according to an embodiment of the present invention.
As shown in fig. 1, the complex operating condition fault diagnosis apparatus may include: a processor 1001, such as a central processing unit (Central Processing Unit, CPU), a communication bus 1002, a user interface 1003, a network interface 1004, a memory 1005. Wherein the communication bus 1002 is used to enable connected communication between these components. The user interface 1003 may include a Display (Display), and the optional user interface 1003 may also include a standard wired interface, a wireless interface, and the wired interface for the user interface 1003 may be a USB interface in the present invention. The network interface 1004 may optionally include a standard wired interface, a Wireless interface (e.g., a Wireless-Fidelity (Wi-Fi) interface). The Memory 1005 may be a high-speed random access Memory (Random Access Memory, RAM) or a stable Memory (NVM), such as a disk Memory. The memory 1005 may also optionally be a storage device separate from the processor 1001 described above.
Those skilled in the art will appreciate that the configuration shown in FIG. 1 is not limiting of the complex operating condition fault diagnosis apparatus and may include more or fewer components than shown, or certain components may be combined, or a different arrangement of components.
As shown in FIG. 1, memory 1005, which is considered a computer storage medium, may include an operating system, a network communication module, a user interface module, and a complex operating condition fault diagnosis program.
In the complex working condition fault diagnosis device shown in fig. 1, the network interface 1004 is mainly used for connecting a background server and performing data communication with the background server; the user interface 1003 is mainly used for connecting user equipment; the complex working condition fault diagnosis device invokes a complex working condition fault diagnosis program stored in the memory 1005 through the processor 1001, and executes the complex working condition fault diagnosis method provided by the embodiment of the invention.
Based on the hardware structure, the embodiment of the complex working condition fault diagnosis method is provided.
Referring to fig. 2, fig. 2 is a schematic flow chart of a first embodiment of the complex operating condition fault diagnosis method according to the present invention.
In this embodiment, the method for diagnosing a fault under a complex working condition includes the following steps:
step S10: manufacturing data of the application device is acquired.
It should be noted that, the execution body of the embodiment may be a complex working condition fault diagnosis system, or may be a device including a complex working condition fault diagnosis function. The device may be a computer, notebook, computer, cell phone, television, etc., and the embodiment is not limited thereto. The embodiment uses a complex working condition fault diagnosis system as an execution main body for explanation.
It should be understood that the application device may be a device that uses the fault diagnosis system for complex working conditions in the present solution during the manufacturing process, and the application device may be a device that processes the product.
It is understood that the manufacturing data may refer to data generated by the application device during the production and processing of the product, and the manufacturing data may be data collected by each sensor on the application device, or may be manufacturing data corresponding to the product. Manufacturing data corresponding to different product production processes are different.
In the specific implementation, the complex working condition fault diagnosis system can collect manufacturing data according to each sensor installed on the application equipment, and can also extract manufacturing data corresponding to product production and processing from a preset database.
Step S20: and diagnosing the manufacturing data according to the preset optimal performance sub-model and the preset classification model to obtain a diagnosis result.
It should be noted that the preset optimal performance sub-model may be a pre-trained optimal performance sub-model, and the optimal performance sub-model may be a model for optimally dividing production processes corresponding to the product. A process may refer to a particular step of making, producing something, or achieving a particular result, or may be a sequence of processing steps that make up the entire production process.
It is understood that the preset classification model may be a pre-trained classification model, and the classification model may be a model for diagnosing a product production process.
It should be understood that the diagnosis result may be data corresponding to a problem process and a problem process among the original processes in the manufacturing data.
In specific implementation, the complex working condition fault diagnosis system can classify manufacturing data according to a preset classification model to obtain a problem process set, diagnose problem processes in the problem process set one by one, judge that the process equipment has no fault if the sensor data of a single problem process is within a standard error range, judge that the process equipment has fault if the sensor data of the single problem process is not within the standard error range, reject the problem process with the fault of the process equipment judged in the problem process, and reserve the problem process with the fault judged as equipment fault as an updated problem process set, thereby rejecting the problem process diagnosis caused by the fault of the previous process, obtaining the problem process which really causes the product quality problem, and diagnose the updated problem process set according to the preset optimal performance submodel so as to obtain the problem process and the data corresponding to the problem process in the original process in the manufacturing data.
Step S30: and determining problem procedure information contained in the manufacturing data according to a preset association algorithm and the diagnosis result, and determining a corresponding fault path according to the problem procedure information.
It should be noted that, the preset association algorithm may be used to find potential links between the problem procedures, and the association relationship and the association rule between the problem procedures may be derived by applying the Apriori association algorithm.
It is understood that the problem process information may be process information including a process group in which the problem process is located, or may include data corresponding to the problem process.
In a specific implementation, the preset association algorithm may adopt an Apriori algorithm, and a diagnosis model is applied in a granularity division manner of an optimal performance submodel to diagnose a block, if the whole block has no problem, the next block is diagnosed, and if the block has a problem, then a single procedure in the block is diagnosed. And (3) completing the problem obtaining process by diagnosis, and obtaining a fault tracing chain by applying an association rule algorithm (Apriori).
The embodiment obtains manufacturing data of the application equipment; diagnosing the manufacturing data according to the preset optimal performance sub-model and the preset classification model to obtain a diagnosis result; and determining problem procedure information contained in the manufacturing data according to a preset association algorithm and a diagnosis result, and determining a corresponding fault path according to the problem procedure information. Because the embodiment diagnoses the manufacturing data through the preset optimal performance sub-model and the preset classification model, the problem procedure information contained in the manufacturing data is determined according to the preset association algorithm and the diagnosis result, and the corresponding fault path is determined according to the problem procedure information.
Referring to fig. 3, fig. 3 is a schematic flow chart of a second embodiment of the complex operating condition fault diagnosis method according to the present invention, and based on the first embodiment shown in fig. 2, the second embodiment of the complex operating condition fault diagnosis method according to the present invention is proposed.
In this embodiment, the step S20 includes:
step S201: and dividing the original working procedure corresponding to the manufacturing data according to a preset optimal performance sub-model so as to obtain the duty ratio of the problem working procedure in the original working procedure.
It should be noted that, the original process may refer to a complete process for producing a processed product, for example: the process for producing the product Q by the production process of L1-1 with reference to FIG. 5 may be 1: ABCDMOQ;2: EFGHMOQ, ABCDMOQ or EFGHMOQ, are two processing steps that are complete in producing the processed product Q. The process of shading is a problematic process. The process group in LI-1 is divided into { ABCD }, { M }, { EFGH }, and { OQ }.
It will be appreciated that the duty cycle of the problem process in the original process may be calculated from the number of problem processes in the original process and the number of original processes, for example: the number of problem processes in { ABCD } is 2, the number of original processes is 4, and the ratio of the problem processes in { ABCD } process group is 1/2; the problem process format in { EFGH } is 1, the number of original processes is 4, and the ratio of the problem process in { EFGH } process group is 1/4; the problem process format in { OQ } is 1, and the number of original processes is 2, that is, the ratio of the problem process to the { OQ } process group is 1/2.
In a specific implementation, the complex working condition fault diagnosis system can divide the complete processing procedure corresponding to the manufacturing data into procedure groups according to the preset optimal performance submodel so as to obtain the duty ratio of the problem procedure in the procedure groups.
Step S202: determining a diagnostic priority based on the duty cycle.
The diagnosis priority may determine the order of priority for diagnosing the process group according to the ratio of the problem process to the original process of the process group. Namely, the process group priority evaluation standard is based on the problem process ratio. The higher the ratio of the problem process to the dividing process group is, the higher the diagnosis priority is, and the diagnosis is prioritized in the diagnosis; the lower the ratio of the problem process in the divided process group is, the lower the diagnosis priority is, the diagnosis is delayed in the diagnosis, and the process equipment is reversely optimized for the high-priority process group so as to gradually reduce the ratio of the problem process.
In specific implementation, the complex working condition fault diagnosis system determines the diagnosis priority of each working procedure group according to the proportion of the problem working procedure in each working procedure group, and the higher the proportion of the problem working procedure in the dividing working procedure group is, the higher the diagnosis priority is, and the diagnosis is carried out in priority in diagnosis; the lower the ratio of the problem process in the divided process group is, the lower the diagnosis priority is, the diagnosis is delayed in the diagnosis, and the process equipment is reversely optimized for the high-priority process group so as to gradually reduce the ratio of the problem process.
Step S203: and diagnosing the manufacturing data according to a preset classification model and the diagnosis priority to obtain a diagnosis result.
The diagnosis result may be data corresponding to a problem step and a problem step in the original steps in the manufacturing data.
In a specific implementation, the complex working condition fault diagnosis system can determine diagnosis priority according to the duty ratio of the problem process in the division process group, and the preset classification model classifies the process of the manufacturing data according to the diagnosis priority, namely, the problem process and the data corresponding to the problem process in the original process in the manufacturing data are obtained.
In this embodiment, the step S30 includes: determining a frequent item set corresponding to a problem procedure in the manufacturing data according to a preset association algorithm and the diagnosis result; and determining a fault path corresponding to the problem procedure according to the frequent item set.
It should be noted that, the frequent item set may be an item set formed by one or more problem links found for the data corresponding to the problem procedure.
In the specific implementation, the complex working condition fault diagnosis system searches the problem procedure data firstly, and then removes the problem procedure data with the highest support degree, so as to determine the frequent item set, and finally connects the frequent item set. And obtaining a frequent item set corresponding to the problem procedure as a maximum fault path according to the association algorithm, and comparing the maximum fault path with the diagnosed problem procedure to obtain the current fault path.
The embodiment obtains manufacturing data of the application equipment; dividing the original working procedure corresponding to the manufacturing data according to a preset optimal performance sub-model to obtain the duty ratio of the problem working procedure in the original working procedure; determining a diagnostic priority based on the duty cycle; and diagnosing the manufacturing data according to the preset classification model and the diagnosis priority to obtain a diagnosis result. And determining problem procedure information contained in the manufacturing data according to a preset association algorithm and a diagnosis result, and determining a corresponding fault path according to the problem procedure information. According to the method, the original working procedures corresponding to the manufacturing data are divided according to the preset optimal performance submodel, so that the duty ratio of the problem working procedures in the original working procedures is obtained; determining a diagnostic priority based on the duty cycle; the method comprises the steps of diagnosing manufacturing data according to a preset classification model and a diagnosis priority to obtain a diagnosis result, determining problem procedure information contained in the manufacturing data according to a preset association algorithm and the diagnosis result, and determining a corresponding fault path according to the problem procedure information. The higher the ratio of the problem process to the dividing process group is, the higher the diagnosis priority is, and the diagnosis is prioritized in the diagnosis; the lower the proportion of the problem process in the divided process group is, the lower the diagnosis priority is, the diagnosis is delayed in the diagnosis, and the process equipment is reversely optimized for the high-priority process group, so that the proportion of the problem process is gradually reduced, the problems generated in the manufacturing process are more accurately positioned, the purpose of diagnosis is achieved, the diagnosis rate is improved, and the fault path of the problems generated in the manufacturing process is more accurately positioned.
Referring to fig. 4, fig. 4 is a schematic flow chart of a third embodiment of the complex operating condition fault diagnosis method according to the present invention, and based on the first embodiment shown in fig. 2, the third embodiment of the complex operating condition fault diagnosis method according to the present invention is proposed.
In this embodiment, before the step S10, the method further includes:
Step S01: historical process manufacturing data is obtained.
The historical process manufacturing data may refer to manufacturing data corresponding to normal and abnormal processes during the production process of the product, and the manufacturing data may include manufacturing data corresponding to normal production indexes corresponding to each process during the production process.
It is understood that the historical process manufacturing data may be manufacturing data corresponding to normal and abnormal processes for different products at the time of manufacturing processing.
In a specific implementation, the historical procedure manufacturing data can be stored in a preset database, the complex working condition fault diagnosis system can acquire the historical procedure manufacturing data from the preset database, the preset database can be classified and stored into data indexes and data according to the historical procedure manufacturing data, and different data are stored in different areas so as to facilitate later data extraction. For example: the data index area can be used for storing index data corresponding to each process of the application equipment in the production and manufacturing process under the normal process, the index data can be used for measuring whether data corresponding to each process in the manufacturing process are abnormal or not, the data update area can be used for storing data corresponding to each process of the application equipment and updating historical data, and the data update area can be used for storing data generated by the abnormal process and data generated by the normal process.
Step S02: and carrying out abnormal labeling on the process corresponding to the historical process manufacturing data so as to obtain an abnormal process and abnormal process data after labeling.
The abnormal labeling can be a process of labeling the abnormal process according to the data index corresponding to the normal process.
In a specific implementation, modeling is performed on the historical process manufacturing data index to determine whether the data index of each process in the historical process manufacturing data is abnormal, the abnormal process is classified, and the manufacturing data corresponding to the abnormal process is marked to obtain marked abnormal process data, namely, abnormal process data.
Step S03: and constructing a problem procedure data set according to the marked abnormal procedure and the marked abnormal procedure data.
The problem process data set may be a data set corresponding to the problem process, in which structured historical process manufacturing data is generated based on association rules between the marked abnormal data and the problem process.
It can be appreciated that the problem process data set may be divided into three parts, one part being training data, one part being tuning data, the other part being test data, and the data being normalized to facilitate later model training.
Step S04: training a model based on LightGBM algorithm according to the problem procedure data set to obtain an initial classification model.
It should be noted that LightGBM (LIGHT GRADIENT Boosting Machine) is a framework for implementing GBDT algorithm, supports efficient parallel training, not only has a faster training speed, but also has lower memory consumption and better accuracy, and supports the advantages of being capable of rapidly processing mass data in a distributed manner. It mainly comprises two algorithms: gradient-based One-sided sampling (GOSS, gradient-based One-SIDE SAMPLING) is bundled with mutually exclusive features (EFB, exclusive Feature Bundling). The LightGBM algorithm-based model may be a model built based on the LightGBM algorithm framework. GBDT (Gradient Boosting Decision Tree) is a model in machine learning, which is mainly trained by using a weak learner (such as a decision tree) to obtain an optimal model, and the model has the advantages of good training effect, difficult fitting and the like. Under the environment of large samples and high dimensionality, the traditional boosting algorithm is very time-consuming to scan all sample points for each feature to select the best segmentation point, and to solve the time-consuming problem, lightGBM uses two solutions as follows: firstly, sampling a sample based on a GOSS algorithm to calculate a gradient; and secondly, binding certain features together based on an EFB algorithm to reduce the dimension of the features, so that the consumption of searching the optimal segmentation point is reduced.
It is understood that the gradient-based single-side sampling algorithm is an algorithm that balances in terms of reducing the amount of data and ensuring accuracy. The algorithm reduces the calculated amount by distinguishing the examples of different gradients, reserving the examples of larger gradients and randomly sampling the smaller gradients, thereby achieving the purpose of improving the efficiency. The mutual exclusion feature binding algorithm is a mode of reducing feature dimensions (dimension reduction) in a feature binding mode, so that the calculation efficiency is improved. The features that are typically bundled are mutually exclusive (one feature value is zero and one feature value is non-zero, i.e., there is a non-zero association and no association is zero) so that the two features are bundled without losing information. If the two features are not completely mutually exclusive (in some cases both features are non-zero values), an index may be used to measure the degree of feature non-mutual exclusion, known as the conflict ratio, and when this value is small, we can choose to bundle the two features that are not completely mutually exclusive without affecting the final accuracy.
It should be understood that the initial classification model may be a model constructed based on LightGBM algorithm framework, i.e., the initial classification model may be a model constructed based on GOSS algorithm and EFB algorithm, and the initial classification model may be a model for classification of problem processes and diagnosis of problem processes.
In a specific implementation, the complex working condition fault diagnosis system can train a model based on LightGBM algorithm according to the problem procedure data set to obtain an initial classification model. The complex working condition fault diagnosis system can obtain an initial classification model by characterizing data in a problem working procedure set, sampling the data and then binding the features, the complex working condition fault diagnosis system can set the iteration step number d, the sampling rate a of large gradient data, the sampling rate b of small gradient data, the sampling rate b of the loss function and the type of weak learner (generally a decision tree) in the embodiment of the invention, and the types of the loss function and the weak learner can be replaced before training the data without specific limitation. Training may be performed by: (1) Sorting the sample points in descending order according to their absolute values; (2) Selecting a sample of 100% of the previous samples from the ordered results to generate a subset of large-gradient sample points; (3) Randomly selecting b (1-a) 100% sample points from 100% of the samples in the rest sample set (1-a), and generating a set of small gradient sample points; (4) Combining the large gradient sample and the sampled small gradient sample; (5) multiplying the small gradient sample by a weight coefficient; (6) Using the sampled samples, learn a new weak learner; (7) And (3) continuously repeating the steps (1) - (6) until the prescribed iteration times or convergence are reached, and finally outputting the trained strong learner. The method can greatly reduce the model learning rate while not losing the accuracy of the learner on the premise of not changing the data distribution. When a=0, the GOSS algorithm degenerates to a random sampling algorithm; when a=1, the GOSS algorithm becomes an algorithm taking the entire sample. In many cases, the GOSS algorithm trains out a model with a higher accuracy than the random sampling algorithm. On the other hand, sampling can increase the diversity of weak learners, thereby potentially improving the generalization ability of the trained model. In the model training process, high-dimensional data is usually sparse, in a sparse feature space, a plurality of features are mutually exclusive, namely, the features never take non-zero values at the same time, the mutually exclusive features can be bound into one feature, so that feature dimension is reduced, namely, the steps are as follows, the features can be firstly ordered according to the number of the non-zero values, then conflict ratios among different features are calculated, and finally each feature is traversed and the features are combined in an attempt to minimize the conflict ratios, so that the model training speed is improved.
Step S05: training the initial classification model according to the problem procedure data set to obtain a preset classification model.
It should be noted that, the preset classification model is a model obtained by training the initial classification model for multiple times according to the problem procedure data set.
Step S06: and training the hierarchical model with adjustable preset granularity according to the historical procedure manufacturing data to obtain an initial optimal performance sub-model.
It should be noted that, the preset granularity-adjustable hierarchical model may be a large calculation amount problem caused by performing model diagnosis based on LightGBM algorithm for simplifying and reducing each process data in the complex working condition operation of multiple processes, and the granularity-adjustable hierarchical model performs process topology map repartition for the process. The process of performing process topology map repartitioning by the granularity-adjustable hierarchical model can refer to the process topology partition schematic diagram of figure 5,
It can be understood that when training the granularity-adjustable hierarchical model according to historical process manufacturing data, three steps can be divided, firstly, a directed acyclic graph is used for topological sorting of the pilot relations among the processes, as shown by L4-1 in a process topological division diagram of fig. 5, each process is regarded as an independent unit, the granularity-adjustable hierarchical model is formed by dividing a process group with the degree of inclusion of 1 in the process topological sorting from fine to coarse by taking the process as an initial granularity-adjustable standard. The granularity-adjustable hierarchical model comprises hierarchical submodels corresponding to each process group.
It should be appreciated that the initial optimal performance sub-model may be a model selected from the hierarchical sub-models corresponding to the respective process groups, and the initial optimal performance sub-model may be used to optimize the model performance, thereby reducing the model computation. There may be multiple initial optimal performance sub-models.
In a specific implementation, the complex working condition fault diagnosis system can train the preset granularity-adjustable hierarchical model according to historical procedure manufacturing data so as to obtain an initial optimal performance sub-model.
Step S07: training the initial optimal performance sub-model according to the historical procedure manufacturing data to obtain a preset optimal performance sub-model.
It should be noted that the preset optimal performance sub-model is a model obtained by training the initial optimal performance sub-model for a plurality of times according to the historical process manufacturing data.
In the specific implementation, the complex working condition fault diagnosis system can perform multi-round training on the initial optimal performance sub-model according to the historical procedure manufacturing data, and select one optimal performance sub-model from a plurality of initial optimal performance sub-models as a preset optimal performance sub-model.
Further, the model based on LightGBM algorithm comprises a model based on a single-side sampling algorithm and a model based on a mutual exclusion feature algorithm; the step S04 includes: training the model based on the single-side sampling algorithm according to the problem procedure data set to obtain an initial learner; performing iterative training on the initial learner to obtain a preset learner; training the model of the mutual exclusion feature algorithm according to the problem procedure data set to obtain a feature binding model; and constructing an initial classification model according to the preset learner and the characteristic binding model.
It should be noted that, the model based on the single-side sampling algorithm may refer to a model based on a gradient single-side sampling algorithm, and the model may greatly reduce the learning rate of the model while not losing the accuracy of the learner without changing the data distribution.
It can be understood that the initial learner may be a new weak learner that learns through the weak learner by using a sampled sample, and the preset learner may be a strong learner that is trained by performing non-stop learning on the weak learner.
It should be appreciated that a model based on a mutually exclusive feature algorithm may refer to a model based on the EFB (Exclusive Feature Bundling, mutually exclusive feature binding) algorithm that may bind certain features together to reduce the dimensionality of the features, resulting in reduced consumption in finding the best cut point. The feature binding model may refer to a model that binds features corresponding to each process in the problem process data set. The feature binding model may bind features in the problem procedure dataset, thereby improving computational efficiency, for example: the feature binding model may rank the features in the problem process dataset by the number of non-zero values, then calculate the conflict ratio between different features, and finally traverse each feature and attempt to merge posts so that the conflict ratio is minimized. The feature binding model may be a model that binds data features corresponding to each process to obtain a feature binding set.
In the specific implementation, the complex working condition fault system can train a model based on a single-side sampling algorithm according to the problem procedure data set so as to obtain an initial learner; performing iterative training on the initial learner to obtain a preset learner; training a model of a mutual exclusion feature algorithm according to the problem procedure data set to obtain a feature binding model; and constructing an initial classification model according to the preset learner and the characteristic binding model.
Further, the step S06 includes: training a hierarchical model with adjustable preset granularity according to the historical procedure manufacturing data to obtain different hierarchical sub-models; evaluating each level sub-model to obtain an optimal sequencing sub-model; training each optimal sequencing sub-model to obtain the occurrence frequency of each optimal sequencing sub-model; and selecting an optimal performance sub-model from all the optimal sequencing sub-models according to the occurrence frequency, and taking the optimal performance sub-model as an initial optimal performance sub-model.
It should be noted that, the hierarchical sub-model may be a hierarchical sub-model corresponding to each process group, referring to a schematic process topology division diagram of fig. 5, for example: l1-1, L2-2, L2-3, L2-4, L3-1, L3-2, L3-3, L3-4, L4-1 correspond to one hierarchical submodel respectively, and the process division of each hierarchical submodel is different, namely, the processes contained in each process group are different.
It is understood that the optimal ranking sub-model may be a model selected from the various hierarchical sub-models, and the hierarchical sub-model with the optimal performance is selected as the optimal ranking sub-model according to the evaluation of the hierarchical sub-models, that is, the process division according to the selected hierarchical sub-model is used as the process division of the optimal ranking sub-model.
It should be understood that the occurrence frequency may be the number of occurrences of each optimal ranking sub-model, and the optimal ranking sub-model with a high occurrence frequency is selected as the initial optimal performance sub-model.
Further, the step of evaluating each hierarchical sub-model to obtain an optimal ranking sub-model includes: determining the number of process groups containing problem processes, the number of working procedures in a single process group containing problem processes and the number of problem processes in a single process group containing problem processes according to the historical process manufacturing data; and evaluating each level submodel according to the number of the process groups containing the problem process, the number of the processes in the single process group containing the problem process and the number of the problem process in the single process group containing the problem process so as to obtain an optimal sequencing submodel.
For further explanation, reference may be made to the process topology division schematic of fig. 5, for example: referring to the group of processes corresponding to the hierarchical submodel L2-2, each process is characterized by A, B, C, D, E, F, G, H, M, O, Q, where the specific detailed process paths can be divided into { A }, { B, C, D }, { M }, { O, Q }, and { E, F, G }, { H }, { M }, { O, Q }, when producing process Q, the problem process is marked by partial shading: B. d, G, Q, i.e., the process group containing the problem is { B, C, D }, { E, F, G }, { O, Q }; the individual process groups including the problem process include: A. b, C, D, E, F, G, H, M, A, O, Q; the problem processes in the single process group containing the problem processes are as follows: B. d, G, Q.
It is understood that the optimal ranking sub-model may refer to a partitioning sub-model that detects optimal processes. Referring to fig. 5, a process topology division diagram has coarse-fine granularity division, and a process division sequence corresponding to each hierarchical sub-model is evaluated to obtain an optimal sequencing sub-model.
In a specific implementation, the optimal ranking sub-model may be according to an evaluation formula:
k represents the number of process groups containing the problem process, m represents the number of processes in a single process group containing the problem process, n represents the number of problem processes in a single process group containing the problem process, and η represents a weight value. In a specific application, the weight value can be a preset value, and can be adjusted in the early stage, and the model performance is better as the evaluation result score is higher. And taking the submodel with the optimal performance as an optimal sequencing submodel.
Manufacturing data by acquiring a history process in the present embodiment; performing anomaly labeling on the process corresponding to the historical process manufacturing data to obtain labeled anomaly process and anomaly process data; constructing a problem procedure data set according to the marked abnormal procedure and the marked abnormal procedure data; training a model based on LightGBM algorithm according to the problem procedure data set to obtain an initial classification model; training the initial classification model according to the problem procedure data set to obtain a preset classification model; training a preset granularity-adjustable hierarchical model according to historical procedure manufacturing data to obtain an initial optimal performance sub-model; training the initial optimal performance sub-model according to the historical procedure manufacturing data to obtain a preset optimal performance sub-model. Acquiring manufacturing data of application equipment; diagnosing the manufacturing data according to the preset optimal performance sub-model and the preset classification model to obtain a diagnosis result; and determining problem procedure information contained in the manufacturing data according to a preset association algorithm and a diagnosis result, and determining a corresponding fault path according to the problem procedure information. Because the embodiment diagnoses the manufacturing data through the preset optimal performance sub-model and the preset classification model, the problem procedure information contained in the manufacturing data is determined according to the preset association algorithm and the diagnosis result, and the corresponding fault path is determined according to the problem procedure information.
In addition, the embodiment of the invention also provides a storage medium, wherein the storage medium is stored with a complex working condition fault diagnosis program, and the complex working condition fault diagnosis program realizes the steps of the complex working condition fault diagnosis method when being executed by a processor.
Referring to fig. 6, fig. 6 is a block diagram showing the structure of a first embodiment of the complex operating condition fault diagnosis apparatus of the present invention.
As shown in fig. 6, the device for diagnosing a fault under a complex working condition according to the embodiment of the present invention includes:
a data acquisition module 10 for acquiring manufacturing data of the application device;
A data diagnosis module 20, configured to diagnose the manufacturing data according to a preset optimal performance sub-model and a preset classification model, so as to obtain a diagnosis result;
The path determining module 30 is configured to determine problem procedure information included in the manufacturing data according to a preset association algorithm and the diagnosis result, and determine a corresponding fault path according to the problem procedure information.
The embodiment obtains manufacturing data of the application equipment; diagnosing the manufacturing data according to the preset optimal performance sub-model and the preset classification model to obtain a diagnosis result; and determining problem procedure information contained in the manufacturing data according to a preset association algorithm and a diagnosis result, and determining a corresponding fault path according to the problem procedure information. Because the embodiment diagnoses the manufacturing data through the preset optimal performance sub-model and the preset classification model, the problem procedure information contained in the manufacturing data is determined according to the preset association algorithm and the diagnosis result, and the corresponding fault path is determined according to the problem procedure information.
Further, the data diagnosis module 20 is further configured to divide the original process corresponding to the manufacturing data according to a preset optimal performance sub-model, so as to obtain the duty ratio of the problem process in the original process; determining a diagnostic priority based on the duty cycle; and diagnosing the manufacturing data according to a preset classification model and the diagnosis priority to obtain a diagnosis result.
Further, the path determining module 30 is further configured to determine a frequent item set corresponding to a problem procedure in the manufacturing data according to a preset association algorithm and the diagnosis result; and determining a fault path corresponding to the problem procedure according to the frequent item set.
Further, the complex working condition fault diagnosis device further comprises: the model construction module is used for acquiring historical procedure manufacturing data; performing abnormal labeling on the process corresponding to the historical process manufacturing data to obtain labeled abnormal process and abnormal process data; constructing a problem procedure data set according to the marked abnormal procedure and the marked abnormal procedure data; training a model based on LightGBM algorithm according to the problem procedure data set to obtain an initial classification model; training the initial classification model according to the problem procedure data set to obtain a preset classification model; training a preset granularity-adjustable hierarchical model according to the historical procedure manufacturing data to obtain an initial optimal performance sub-model; training the initial optimal performance sub-model according to the historical procedure manufacturing data to obtain a preset optimal performance sub-model.
Further, the model construction module is further used for training the model based on the single-side sampling algorithm according to the problem procedure data set so as to obtain an initial learner; performing iterative training on the initial learner to obtain a preset learner; training the model based on the mutual exclusion feature algorithm according to the problem procedure data set to obtain a feature binding model; and constructing an initial classification model according to the preset learner and the characteristic binding model.
Further, the model construction module is further used for training a preset granularity-adjustable hierarchical model according to the historical procedure manufacturing data so as to obtain different hierarchical sub-models; evaluating each level sub-model to obtain an optimal sequencing sub-model; training each optimal sequencing sub-model to obtain the occurrence frequency of each optimal sequencing sub-model; and selecting an optimal performance sub-model from all the optimal sequencing sub-models according to the occurrence frequency, and taking the optimal performance sub-model as an initial optimal performance sub-model.
Further, the model construction module is further used for determining the number of process groups containing problem processes, the number of working procedures in a single process group containing problem processes and the number of problem processes in a single process group containing problem processes according to the historical process manufacturing data; and evaluating each level submodel according to the number of the process groups containing the problem process, the number of the processes in the single process group containing the problem process and the number of the problem process in the single process group containing the problem process so as to obtain an optimal sequencing submodel.
Other embodiments or specific implementation manners of the complex working condition fault diagnosis device of the present invention may refer to the above method embodiments, and are not described herein again.
In addition, the embodiment of the invention also provides a storage medium, wherein the storage medium is stored with a complex working condition fault diagnosis program, and the complex working condition fault diagnosis program realizes the steps of the complex working condition fault diagnosis method when being executed by a processor.
It should be understood that the foregoing is illustrative only and is not limiting, and that in specific applications, those skilled in the art may set the invention as desired, and the invention is not limited thereto.
It should be noted that the above-described working procedure is merely illustrative, and does not limit the scope of the present invention, and in practical application, a person skilled in the art may select part or all of them according to actual needs to achieve the purpose of the embodiment, which is not limited herein.
In addition, technical details which are not described in detail in the embodiment can be referred to the complex working condition fault diagnosis method provided by any embodiment of the present invention, and are not described here again.
It should 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 system 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 system. Without further limitation, an element defined by the phrase "comprising one … …" does not exclude the presence of other like elements in a process, method, article, or system that comprises the element.
The foregoing embodiment numbers of the present invention are merely for the purpose of description, and do not represent the advantages or disadvantages of the embodiments. In the unit claims enumerating several means, several of these means may be embodied by one and the same item of hardware. The use of the terms first, second, third, etc. do not denote any order, but rather the terms first, second, third, etc. are used to interpret the terms as names.
From the above description of the embodiments, it will be clear to those skilled in the art that the above-described embodiment method may be implemented by means of software plus a necessary general hardware platform, but of course may also be implemented by means of hardware, but in many cases the former is a preferred embodiment. Based on such understanding, the technical solution of the present invention may be embodied essentially or in a part contributing to the prior art in the form of a software product stored in a storage medium (e.g. read only memory mirror (Read Only Memory image, ROM)/random access memory (Random Access Memory, RAM), magnetic disk, optical disk), comprising instructions for causing a terminal device (which may be a mobile phone, a computer, a server, or a network device, etc.) to perform the method according to the embodiments of the present invention.
The foregoing description is only of the preferred embodiments of the present invention, and is not intended to limit the scope of the invention, but rather is intended to cover any equivalents of the structures or equivalent processes disclosed herein or in the alternative, which may be employed directly or indirectly in other related arts.
Claims (6)
1. The complex working condition fault diagnosis method is characterized by comprising the following steps of:
Acquiring historical process manufacturing data;
performing abnormal labeling on the process corresponding to the historical process manufacturing data to obtain labeled abnormal process and abnormal process data;
Constructing a problem procedure data set according to the marked abnormal procedure and the marked abnormal procedure data;
Training a model based on LightGBM algorithm according to the problem procedure data set to obtain an initial classification model;
Training the initial classification model according to the problem procedure data set to obtain a preset classification model;
Training a hierarchical model with adjustable preset granularity according to the historical procedure manufacturing data to obtain different hierarchical sub-models;
evaluating each level sub-model to obtain an optimal sequencing sub-model;
training each optimal sequencing sub-model to obtain the occurrence frequency of each optimal sequencing sub-model;
selecting an optimal performance sub-model from all optimal sequencing sub-models according to the occurrence frequency, and taking the optimal performance sub-model as an initial optimal performance sub-model;
Training the initial optimal performance sub-model according to the historical procedure manufacturing data to obtain a preset optimal performance sub-model;
Acquiring manufacturing data of application equipment;
Dividing the original working procedure corresponding to the manufacturing data according to a preset optimal performance sub-model to obtain the duty ratio of the problem working procedure in the original working procedure;
determining a diagnostic priority based on the duty cycle;
diagnosing the manufacturing data according to a preset classification model and the diagnosis priority to obtain a diagnosis result;
Determining a frequent item set corresponding to a problem procedure in the manufacturing data according to a preset association algorithm and the diagnosis result;
and determining a fault path corresponding to the problem procedure according to the frequent item set.
2. The complex operating condition fault diagnosis method according to claim 1, wherein the model based on LightGBM algorithm comprises a model based on a single-side sampling algorithm and a model based on a mutual exclusion feature algorithm;
The step of training a LightGBM algorithm-based model according to the problem procedure dataset to obtain an initial classification model includes:
Training the model based on the single-side sampling algorithm according to the problem procedure data set to obtain an initial learner;
Performing iterative training on the initial learner to obtain a preset learner;
Training the model based on the mutual exclusion feature algorithm according to the problem procedure data set to obtain a feature binding model;
And constructing an initial classification model according to the preset learner and the characteristic binding model.
3. The complex operating condition fault diagnosis method according to claim 1, wherein the step of evaluating each level submodel to obtain an optimal ranking submodel comprises:
Determining the number of process groups containing problem processes, the number of working procedures in a single process group containing problem processes and the number of problem processes in a single process group containing problem processes according to the historical process manufacturing data;
And evaluating each level submodel according to the number of the process groups containing the problem process, the number of the processes in the single process group containing the problem process and the number of the problem process in the single process group containing the problem process so as to obtain an optimal sequencing submodel.
4. A complex operating condition fault diagnosis apparatus, characterized by comprising: a memory, a processor and a complex operating condition fault diagnosis program stored on the memory and operable on the processor, which complex operating condition fault diagnosis program when executed by the processor implements the steps of the complex operating condition fault diagnosis method according to any one of claims 1 to 3.
5. A storage medium having stored thereon a complex operating condition fault diagnosis program which when executed by a processor implements the steps of the complex operating condition fault diagnosis method according to any one of claims 1 to 3.
6. A complex operating condition fault diagnosis device, characterized in that the complex operating condition fault diagnosis device comprises:
The model building module is used for acquiring historical procedure manufacturing data; performing abnormal labeling on the process corresponding to the historical process manufacturing data to obtain labeled abnormal process and abnormal process data; constructing a problem procedure data set according to the marked abnormal procedure and the marked abnormal procedure data; training a model based on LightGBM algorithm according to the problem procedure data set to obtain an initial classification model; training the initial classification model according to the problem procedure data set to obtain a preset classification model; training a hierarchical model with adjustable preset granularity according to the historical procedure manufacturing data to obtain different hierarchical sub-models; evaluating each level sub-model to obtain an optimal sequencing sub-model; training each optimal sequencing sub-model to obtain the occurrence frequency of each optimal sequencing sub-model; selecting an optimal performance sub-model from all optimal sequencing sub-models according to the occurrence frequency, and taking the optimal performance sub-model as an initial optimal performance sub-model; training the initial optimal performance sub-model according to the historical procedure manufacturing data to obtain a preset optimal performance sub-model;
the data acquisition module is used for acquiring manufacturing data of the application equipment;
The data diagnosis module is used for dividing the original working procedure corresponding to the manufacturing data according to the preset optimal performance sub-model so as to obtain the duty ratio of the problem working procedure in the original working procedure; determining a diagnostic priority based on the duty cycle; diagnosing the manufacturing data according to a preset classification model and the diagnosis priority to obtain a diagnosis result;
The path determining module is used for determining a frequent item set corresponding to a problem procedure in the manufacturing data according to a preset association algorithm and the diagnosis result; and determining a fault path corresponding to the problem procedure according to the frequent item set.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110776972.XA CN113610350B (en) | 2021-07-08 | 2021-07-08 | Complex working condition fault diagnosis method, equipment, storage medium and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110776972.XA CN113610350B (en) | 2021-07-08 | 2021-07-08 | Complex working condition fault diagnosis method, equipment, storage medium and device |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113610350A CN113610350A (en) | 2021-11-05 |
| CN113610350B true CN113610350B (en) | 2024-04-23 |
Family
ID=78304315
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110776972.XA Active CN113610350B (en) | 2021-07-08 | 2021-07-08 | Complex working condition fault diagnosis method, equipment, storage medium and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113610350B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117171625B (en) * | 2023-10-23 | 2024-02-06 | 云和恩墨(北京)信息技术有限公司 | Intelligent classification method and device for working conditions, electronic equipment and storage medium |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109034220A (en) * | 2018-07-13 | 2018-12-18 | 福州大学 | A kind of intelligent photovoltaic array method for diagnosing faults based on optimal rotation forest |
| CN109657945A (en) * | 2018-12-06 | 2019-04-19 | 华中科技大学 | A kind of industrial process fault diagnosis method based on data-driven |
| CN111444169A (en) * | 2020-03-30 | 2020-07-24 | 国网江苏省电力有限公司连云港市赣榆区供电分公司 | Transformer substation electrical equipment state monitoring and diagnosis system and method |
| CN112346941A (en) * | 2019-08-08 | 2021-02-09 | 北京国双科技有限公司 | Fault diagnosis method and device |
-
2021
- 2021-07-08 CN CN202110776972.XA patent/CN113610350B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109034220A (en) * | 2018-07-13 | 2018-12-18 | 福州大学 | A kind of intelligent photovoltaic array method for diagnosing faults based on optimal rotation forest |
| CN109657945A (en) * | 2018-12-06 | 2019-04-19 | 华中科技大学 | A kind of industrial process fault diagnosis method based on data-driven |
| CN112346941A (en) * | 2019-08-08 | 2021-02-09 | 北京国双科技有限公司 | Fault diagnosis method and device |
| CN111444169A (en) * | 2020-03-30 | 2020-07-24 | 国网江苏省电力有限公司连云港市赣榆区供电分公司 | Transformer substation electrical equipment state monitoring and diagnosis system and method |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113610350A (en) | 2021-11-05 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108804641B (en) | Text similarity calculation method, device, equipment and storage medium | |
| CN112232413B (en) | High-dimensional data feature selection method based on graph neural network and spectral clustering | |
| CN109460793A (en) | A kind of method of node-classification, the method and device of model training | |
| CN108446616B (en) | Road extraction method based on full convolution neural network ensemble learning | |
| CN107103332A (en) | A kind of Method Using Relevance Vector Machine sorting technique towards large-scale dataset | |
| US9690969B2 (en) | Information processing apparatus, non-transitory computer readable medium, and information processing method | |
| CN113255892B (en) | Decoupled network structure searching method, device and readable storage medium | |
| CN113125960A (en) | Vehicle-mounted lithium ion battery charge state prediction method based on random forest model | |
| CN117294727B (en) | Cloud edge end collaborative management method based on cloud primordia and container technology | |
| CN113610350B (en) | Complex working condition fault diagnosis method, equipment, storage medium and device | |
| CN112308298A (en) | Multi-scenario performance index prediction method and system for semiconductor production line | |
| CN114511083A (en) | Model training method and device, storage medium and electronic device | |
| CN117036781A (en) | Image classification method based on tree comprehensive diversity depth forests | |
| CN115374019B (en) | Method and system for testing distributed UI test cases and computer storage medium | |
| CN108830302B (en) | Image classification method, training method, classification prediction method and related device | |
| CN116186384A (en) | Article recommendation method and system based on article implicit feature similarity | |
| CN112257977B (en) | Logistics project construction period optimization method and system with resource limitation under fuzzy man-hour | |
| CN114611668A (en) | Vector representation learning method and system based on heterogeneous information network random walk | |
| CN113296947A (en) | Resource demand prediction method based on improved XGboost model | |
| CN113393023A (en) | Mold quality evaluation method, apparatus, device and storage medium | |
| CN114529749A (en) | Pathological image classification device and method and use method of device | |
| CN115858629B (en) | KNN query method based on learning index | |
| CN115017125B (en) | Data processing method and device for improving KNN method | |
| CN111242235B (en) | Similar characteristic test data set generation method | |
| CN114357219A (en) | Mobile-end-oriented instance-level image retrieval method and device |
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 |