CN110929149B - Industrial equipment fault maintenance recommendation method and system - Google Patents
Industrial equipment fault maintenance recommendation method and system Download PDFInfo
- Publication number
- CN110929149B CN110929149B CN201911102558.XA CN201911102558A CN110929149B CN 110929149 B CN110929149 B CN 110929149B CN 201911102558 A CN201911102558 A CN 201911102558A CN 110929149 B CN110929149 B CN 110929149B
- Authority
- CN
- China
- Prior art keywords
- fault
- maintenance
- vector
- equipment
- entity
- 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
- 238000012423 maintenance Methods 0.000 title claims abstract description 84
- 238000000034 method Methods 0.000 title claims abstract description 25
- 238000012549 training Methods 0.000 claims abstract description 7
- 238000012163 sequencing technique Methods 0.000 claims abstract description 5
- 239000013598 vector Substances 0.000 claims description 54
- 230000000007 visual effect Effects 0.000 claims description 12
- 230000007246 mechanism Effects 0.000 claims description 6
- 230000011218 segmentation Effects 0.000 claims description 5
- 238000004458 analytical method Methods 0.000 claims description 4
- 238000013507 mapping Methods 0.000 claims description 4
- 210000002569 neuron Anatomy 0.000 claims description 4
- 230000008439 repair process Effects 0.000 claims description 4
- 238000013024 troubleshooting Methods 0.000 claims description 4
- 238000012800 visualization Methods 0.000 claims description 4
- 238000004364 calculation method Methods 0.000 claims description 3
- 239000000284 extract Substances 0.000 claims description 3
- 238000002372 labelling Methods 0.000 claims 1
- 238000003058 natural language processing Methods 0.000 abstract description 4
- 238000013145 classification model Methods 0.000 abstract description 2
- 230000008569 process Effects 0.000 description 6
- 238000007726 management method Methods 0.000 description 4
- 238000010586 diagram Methods 0.000 description 3
- 238000004519 manufacturing process Methods 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 230000007257 malfunction Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/90—Details of database functions independent of the retrieved data types
- G06F16/95—Retrieval from the web
- G06F16/953—Querying, e.g. by the use of web search engines
- G06F16/9535—Search customisation based on user profiles and personalisation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/35—Clustering; Classification
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/36—Creation of semantic tools, e.g. ontology or thesauri
- G06F16/367—Ontology
-
- 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
- G06Q10/20—Administration of product repair or maintenance
-
- 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
-
- 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
- Y02P90/80—Management or planning
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Databases & Information Systems (AREA)
- Business, Economics & Management (AREA)
- General Physics & Mathematics (AREA)
- Human Resources & Organizations (AREA)
- General Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Economics (AREA)
- Tourism & Hospitality (AREA)
- General Business, Economics & Management (AREA)
- Strategic Management (AREA)
- Marketing (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Entrepreneurship & Innovation (AREA)
- Manufacturing & Machinery (AREA)
- Health & Medical Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Primary Health Care (AREA)
- Life Sciences & Earth Sciences (AREA)
- Animal Behavior & Ethology (AREA)
- Computational Linguistics (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a method and a system for recommending fault maintenance of industrial equipment, wherein the method comprises the following steps of S1: training a fault entity and type classifier in a basic maintenance knowledge map to classify maintenance classes; s2: acquiring the problem of equipment failure; s3: extracting key words from the fault description input by a user to generate fault items; s4: and sorting out historical fault cases closest to the content of the fault items from the basic maintenance knowledge graph by utilizing machine sequencing, and recommending the maintenance decisions of the cases to the equipment point management personnel. The invention carries out natural language processing on the fault input by the maintenance personnel, inputs the fault into the classification model of iterative training for scoring, selects the historical maintenance records with high similarity and close scores in sequence, and recommends the maintenance scheme corresponding to the historical maintenance records to the maintenance personnel, so that the maintenance personnel work more quickly and conveniently, and quickly makes maintenance decision measures when encountering familiar faults.
Description
Technical Field
The invention relates to the field of fault maintenance of industrial equipment, in particular to a fault maintenance recommendation method and system for the industrial equipment.
Background
The industrial manufacturing industry internet is a key infrastructure for linking industrial whole systems, industrial whole chains and value whole chains and supporting industrial intelligent development. The global industry is advancing a brand new era of internet of things, and the demand for improving the automation of various operation processes is increasing.
The industrial equipment network is complicated, the equipment is thousands of, and once a fault occurs, the operation of the whole industrial equipment network is influenced if the fault is not processed in time.
In the equipment maintenance and guarantee process of the industrial manufacturing industry, a large amount of data such as equipment maintenance records, fault cases and the like can be generated, and certain maintenance specifications and maintenance standards are referred to when various equipment is operated and detected. Much value information is contained in these repair history data and specification standards. The information can effectively promote the maintenance work of the basic guarantee department, can provide information reference for decision-making personnel, and improves the efficiency and the accuracy of equipment maintenance guarantee.
Disclosure of Invention
The invention provides a method and a system for recommending fault maintenance of industrial equipment, which can accelerate the process of equipment maintenance, provide simple and convenient means for industrial equipment maintenance personnel and improve the problem of maintenance management in the industrial equipment.
In order to solve the technical problems, the technical scheme of the invention is as follows:
a fault repair recommendation method for industrial equipment comprises the following steps:
s1: training a fault entity and type classifier in a basic maintenance knowledge map to classify maintenance classes;
s2: acquiring the problem of equipment failure;
s3: extracting key words from the fault description input by the user to generate fault items;
s4: and sorting out historical fault cases closest to the content of the fault items from the basic maintenance knowledge graph by utilizing machine sequencing, and recommending the maintenance decisions of the cases to the equipment point management personnel.
Preferably, step S1 comprises the steps of:
s1.1: in the basic maintenance knowledge map, firstly, the part of speech is marked for the fault item in the maintenance items, then named entity identification is carried out, and a data set is automatically marked to generate a corpus;
s1.2: in the corpus, a word embedding vector w = { w = is obtained by inputting word2vector with a sentence as a large unit and a word as a small unit 1 ,w 2 ,…,w n Where n is the length of the sentence;
s1.3: inputting the word embedding vector w into a Bi-LSTM network, and outputting to obtain a vocabulary vector representation x, x = { x = } x 1 ,x 2 ,…,x n }；
S1.4: inputting the vocabulary vector representation output by the Bi-LSTM network into attentionIn the force network, the vocabulary vector output by the attention network is combined with the feature vectors of the rest words to form w' = { w = { (w) } 1 ′,w 2 ′,…,w n ′}；
S1.5: and inputting w' into a softmax layer of the output layer to obtain a classification result, and mapping the output of a plurality of neurons into a (0, 1) interval by the softmax layer to perform multi-classification.
Preferably, the fault entry is tagged with part-of-speech in step S1.1 using the chinese NLP named entity recognition sequence tagging tool YEDDA.
Preferably, YEDDA in step S1.1 makes use of CRF for named entity recognition.
Preferably, in step S1.1 a BIO or BMES annotation system is selected.
Preferably, in step S1.4, the calculation formula of the attention network is as follows:
z a ＝a⊙z
Preferably, the method for generating a fault search in step S3 adopts generation of a lexical dependency tree, and specifically includes the following steps:
s3.1: performing word segmentation and part-of-speech tagging on the equipment fault problem input by a user by using Stanford CoreNLP;
s3.2: and generating a minimum spanning tree by using a Prim algorithm, converting the minimum spanning tree into CoNLL format output, and displaying by using a visualization tool.
Preferably, step S4 specifically includes the steps of:
s4.1: inputting the natural language query sentence S of the equipment failure problem processed into the vocabulary language dependency tree into the failure entity and type classifier of S1 to obtain a category which the preliminary belongs to and a failure entity set which is corresponding to the category and possibly belongs to the same category, wherein the category which the preliminary belongs to and the failure entity set which is possibly belongs to the same category form a set e together;
s4.2: converting the s and the set e into a Vector form through a Term Vector layer, inputting the Vector form into a multilayer nonlinear perceptron network, and outputting a low-dimensional Vector containing semantic information;
s4.3: and calculating cosine similarity of the s and the set e, normalizing by softmax to obtain a unit in the set e which is most matched with the sentence s, and outputting historical maintenance operation data of the fault entity corresponding to the s to a user.
Preferably, the cosine similarity in step S4.3 is calculated as follows:
wherein A = A 1 ,A 2 ,A 3 ,…,A n Vector representing the problem of equipment failure, B = B 1 ,B 2 ,B 3 ,…,B n The vector corresponding to any category in the set e is represented, and cos (theta) represents cosine similarity.
An industrial equipment troubleshooting recommendation system comprising:
a basic maintenance knowledge graph module that classifies maintenance classes using a fault entity and type classifier;
the visual interface module provides a perfect visual interface, and equipment maintenance personnel input equipment fault problems through the visual interface module and acquire maintenance decision recommendation from the visual interface module;
the fault analysis and storage module extracts key words from fault description input by a user and generates a fault search query;
and the maintenance decision recommendation module selects the historical fault case closest to the content of the fault item from the basic maintenance knowledge map and recommends the maintenance decision of the case to the equipment point manager.
Compared with the prior art, the technical scheme of the invention has the beneficial effects that:
the invention carries out natural language processing on the fault input by the maintenance personnel, inputs the fault into the classification model of iterative training for scoring, and selects the historical maintenance records with high similarity and close score in sequence, recommends the maintenance scheme corresponding to the historical maintenance records to the maintenance personnel, combines the input fault with the existing historical maintenance records by adopting the learning sequencing iterative training process, leads the working efficiency of the maintenance personnel to be quicker and more convenient, and quickly makes maintenance decision measures when encountering familiar faults.
Drawings
FIG. 1 is a schematic flow chart of the method of the present invention.
Fig. 2 is a schematic diagram of the system module connection according to the present invention.
FIG. 3 is a diagram illustrating an exemplary lexical dependency tree process.
FIG. 4 is a process diagram of a multi-layered nonlinear perceptron.
Detailed Description
The drawings are for illustrative purposes only and are not to be construed as limiting the patent;
for the purpose of better illustrating the embodiments, certain features of the drawings may be omitted, enlarged or reduced, and do not represent the size of an actual product;
it will be understood by those skilled in the art that certain well-known structures in the drawings and descriptions thereof may be omitted.
The technical solution of the present invention is further described with reference to the drawings and the embodiments.
Example 1
The embodiment provides a method for recommending fault maintenance of industrial equipment, as shown in fig. 1, which includes the following steps:
s1: training a fault entity and type classifier in a basic maintenance knowledge map to classify maintenance classes;
s2: acquiring the problem of equipment failure;
s3: extracting key words from the fault description input by the user to generate fault items;
s4: and sorting the historical fault cases closest to the contents of the fault items from the basic maintenance knowledge graph by using machine sequencing, and recommending the maintenance decisions of the cases to the equipment point management personnel.
The step S1 includes the steps of:
s1.1: in the basic maintenance knowledge map, firstly, the part of speech of a fault item in the maintenance items is labeled, then named entity recognition is carried out, and a data set is automatically labeled to generate a corpus;
s1.2: in the corpus, a word embedding vector w = { w = is obtained by inputting word2vector with a sentence as a large unit and a word as a small unit 1 ,w 2 ,…,w n Where n is the length of the sentence;
s1.3: inputting the word embedding vector w into a Bi-LSTM network, and outputting to obtain a vocabulary vector representation x, x = { x = } x 1 ,x 2 ,…,x n }；
S1.4: the vocabulary vector representation output by the Bi-LSTM network is input into the attention network, and the vocabulary vector output by the attention network is combined with the feature vectors of the rest words to form w' = { w = 1 ′,w 2 ′,…,w n ′}；
S1.5: and inputting w' into a softmax layer of the output layer to obtain a classification result, and mapping the output of a plurality of neurons into a (0, 1) interval by the softmax layer to perform multi-classification.
In step S1.1, a part-of-speech is marked for the fault entry by using a Chinese NLP named entity recognition sequence marking tool YEDDA.
In step S1.1, YEDDA performs named entity recognition using CRF.
In step S1.1, a BIO or BMES annotation system is selected.
In step S1.4, the calculation formula of the attention network is as follows:
z a ＝a⊙z
S3, generating a lexical dependency tree by adopting a fault searching method, and specifically comprising the following steps of:
s3.1: performing word segmentation and part-of-speech tagging on an equipment fault problem input by a user by using Stanford CoreNLP; for example, "what is the reason why gate No. 2 of subway line No. 4 has a fault", the word segmentation and part of speech tagging become [ subway/n line No. four/nz, fault/v, cause/n ], and because there is a virtual root in the dependency syntax tree, a virtual node is added to the dependency syntax tree, so that there are four nodes in total: [ # # core # #/root, subway/n quarter line/nz, fault/v, cause/n ]; each node forms a directed edge with the other three, for a total of 4 x 3=12, there being only 9.
S3.2: generating a minimum spanning tree by using a Prim algorithm, converting the minimum spanning tree into CoNLL format output, and displaying the CoNLL format output by using a visualization tool, as shown in FIG. 3;
step S4 is shown in fig. 4, and specifically includes the following steps:
s4.1: inputting the natural language query sentence S of the equipment failure problem processed into the vocabulary language dependency tree into the failure entity and type classifier of S1 to obtain a category which the preliminary belongs to and a failure entity set which is corresponding to the category and possibly belongs to the same category, wherein the category which the preliminary belongs to and the failure entity set which is possibly belongs to the same category form a set e together;
s4.2: converting the s and the set e into a Vector form through a Term Vector layer, inputting the Vector form into a multilayer nonlinear perceptron network, and outputting a low-dimensional Vector containing semantic information;
s4.3: and calculating cosine similarity of the s and the set e, normalizing by using softmax to obtain a unit in the set e which is most matched with the sentence s, and outputting historical maintenance operation data of the fault entity corresponding to the s to a user.
The formula for calculating the cosine similarity in step S4.3 is as follows:
wherein A = A 1 ,A 2 ,A 3 ,…,A n Vector representing the problem of equipment failure, B = B 1 ,B 2 ,B 3 ,…,B n The vector corresponding to any category in the set e is represented, and cos (theta) represents cosine similarity.
Example 2
The embodiment provides an industrial equipment malfunction repair recommendation system, as shown in fig. 2, including:
a basic maintenance knowledge graph module that classifies maintenance classes using a fault entity and type classifier;
the visual interface module provides a perfect visual interface, equipment maintenance personnel input equipment fault problems through the visual interface module and obtain maintenance decision recommendation from the visual interface module;
the fault analysis and storage module extracts key words from fault description input by a user and generates a fault search query;
and the maintenance decision recommendation module selects the historical fault case closest to the fault item content from the basic maintenance knowledge graph and recommends the maintenance decision of the case to the equipment point management personnel.
The same or similar reference numerals correspond to the same or similar parts;
the terms describing positional relationships in the drawings are for illustrative purposes only and should not be construed as limiting the patent;
it should be understood that the above-described embodiments of the present invention are merely examples for clearly illustrating the present invention, and are not intended to limit the embodiments of the present invention. Other variations and modifications will be apparent to persons skilled in the art in light of the above description. And are neither required nor exhaustive of all embodiments. Any modification, equivalent replacement, and improvement made within the spirit and principle of the present invention should be included in the protection scope of the claims of the present invention.
Claims (6)
1. A method for recommending fault maintenance of industrial equipment is characterized by comprising the following steps:
s1: training a fault entity and type classifier in a basic maintenance knowledge map to classify maintenance classes;
s2: acquiring the problem of equipment failure;
s3: extracting key words from the fault description input by a user to generate fault items;
s4: sorting out historical fault cases closest to the contents of the fault items from a basic maintenance knowledge graph by utilizing machine sequencing, and recommending maintenance decisions of the cases to equipment point managers;
the step S1 includes the steps of:
s1.1: in the basic maintenance knowledge map, firstly, the part of speech is marked for the fault item in the maintenance items, then named entity identification is carried out, and a data set is automatically marked to generate a corpus;
s1.2: in the corpus, a word embedding vector w = { w = is obtained by inputting word2vector with a sentence as a large unit and a word as a small unit 1 ,w 2 ,…,w n H, where n is the length of the sentence;
s1.3: inputting the word embedding vector w into a Bi-LSTM network, and outputting to obtain a vocabulary vector representation x, x = { x = } x 1 ,x 2 ,…,x n }；
S1.4: the vocabulary vector representation output by the Bi-LSTM network is input into the attention network, and the vocabulary vector output by the attention network is combined with the feature vectors of other words to form w ′ ＝{w 1 ′ ,w 2 ′ ,…,w n ′ }；
S1.5: will w ′ Inputting a softmax layer of an output layer to obtain a classification result, and mapping the output of a plurality of neurons into a (0, 1) interval by the softmax layer to perform multi-classification;
s3, generating a lexical dependency tree by adopting a fault searching method, and specifically comprising the following steps of:
s3.1: performing word segmentation and part-of-speech tagging on the equipment fault problem input by a user by using Stanford CoreNLP;
s3.2: generating a minimum spanning tree by using a Prim algorithm, converting the minimum spanning tree into a CoNLL format for output, and displaying by using a visualization tool;
step S4 specifically includes the following steps:
s4.1: inputting the natural language query sentence S of the equipment failure problem processed into the vocabulary language dependency tree into the failure entity and type classifier of S1 to obtain a category which the preliminary belongs to and a failure entity set which is corresponding to the category and possibly belongs to the same category, wherein the category which the preliminary belongs to and the failure entity set which is possibly belongs to the same category form a set e together;
s4.2: converting the s and the set e into a Vector form through a Term Vector layer, inputting the Vector form into a multilayer nonlinear perceptron network, and outputting a low-dimensional Vector containing semantic information;
s4.3: calculating cosine similarity of the s and the set e, normalizing by using softmax to obtain a unit in the set e which is most matched with the sentence s, and outputting historical maintenance operation data of the fault entity corresponding to the s to a user;
the formula for calculating the cosine similarity in step S4.3 is as follows:
wherein A = A 1 ,A 2 ,A 3 ,…,A n Vector representing the problem of equipment failure, B = B 1 ,B 2 ,B 3 ,…,B n The vector corresponding to any category in the set e is represented, and cos (theta) represents cosine similarity.
2. The industrial equipment fault repair recommendation method according to claim 1, characterized in that part-of-speech is tagged to the fault entry in step S1.1 by using chinese NLP named entity recognition sequence tagging tool YEDDA.
3. The industrial equipment troubleshooting recommendation method of claim 1 wherein YEDDA in step S1.1 utilizes CRF for named entity identification.
4. The method for recommending fault maintenance of industrial equipment according to claim 1, characterized in that in step S1.1, a BIO or BMES labeling system is selected.
5. The method for recommending the trouble shooting of industrial equipment according to claim 1, wherein in step S1.4, the calculation formula of the attention network is as follows:
z a ＝a⊙z
6. An industrial equipment troubleshooting recommendation system, comprising:
a basic maintenance knowledge graph module that classifies maintenance classes using a fault entity and type classifier;
the visual interface module provides a perfect visual interface, and equipment maintenance personnel input equipment fault problems through the visual interface module and acquire maintenance decision recommendation from the visual interface module;
the fault analysis and storage module extracts key words from fault description input by a user and generates a fault search query;
the maintenance decision recommendation module selects a historical fault case closest to the content of the fault item from the basic maintenance knowledge map, and recommends the maintenance decision of the case to an equipment point manager;
the basic maintenance knowledge map module comprises the following steps:
in the basic maintenance knowledge map, firstly, the part of speech is marked for the fault item in the maintenance items, then named entity identification is carried out, and a data set is automatically marked to generate a corpus;
in the corpus, a word embedding vector w = { w = is obtained by inputting word2vector with a sentence as a large unit and a word as a small unit 1 ,w 2 ,…,w n Where n is the length of the sentence;
inputting the word embedding vector w into a Bi-LSTM network, and outputting to obtain a vocabulary vector representation x, x = { x = } x 1 ,x 2 ,…,x n }；
The vocabulary vector representation output by the Bi-LSTM network is input into the attention network, and the vocabulary vector output by the attention network and the rest vocabulary vectorsWord feature vector combination as w ′ ＝{w 1 ′ ,w 2 ′ ,…,w n ′ }；
Will w ′ Inputting a softmax layer of an output layer to obtain a classification result, and mapping the output of a plurality of neurons into a (0, 1) interval by the softmax layer to perform multi-classification;
the method for generating the fault search by the fault analysis and storage module adopts the vocabulary-based dependency tree generation method, and specifically comprises the following steps:
performing word segmentation and part-of-speech tagging on the equipment fault problem input by a user by using Stanford CoreNLP;
generating a minimum spanning tree by using a Prim algorithm, converting the minimum spanning tree into a CoNLL format for output, and displaying by using a visualization tool;
the maintenance decision recommendation module specifically comprises the following steps:
inputting the natural language query sentence S of the equipment failure problem processed into the vocabulary language dependency tree into the failure entity and type classifier of S1 to obtain a category which the preliminary belongs to and a failure entity set which is corresponding to the category and possibly belongs to the same category, wherein the category which the preliminary belongs to and the failure entity set which is possibly belongs to the same category form a set e together;
converting the s and the set e into a Vector form through a Term Vector layer, inputting the Vector form into a multilayer nonlinear perceptron network, and outputting a low-dimensional Vector containing semantic information;
calculating cosine similarity of the s and the set e, normalizing by using softmax to obtain a unit in the set e which is most matched with the sentence s, and outputting historical maintenance operation data of the fault entity corresponding to the s to a user;
the cosine similarity is calculated as follows:
wherein A = A 1 ，A 2 ，A 3 ，...，A n Vector representing the problem of equipment failure, B = B 1 ，B 2 ，B 3 ，...，B n The vector corresponding to any category in the set e is represented, and cos (alpha) represents cosine similarity.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911102558.XA CN110929149B (en) | 2019-11-12 | 2019-11-12 | Industrial equipment fault maintenance recommendation method and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911102558.XA CN110929149B (en) | 2019-11-12 | 2019-11-12 | Industrial equipment fault maintenance recommendation method and system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110929149A CN110929149A (en) | 2020-03-27 |
| CN110929149B true CN110929149B (en) | 2023-04-07 |
Family
ID=69852708
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201911102558.XA Active CN110929149B (en) | 2019-11-12 | 2019-11-12 | Industrial equipment fault maintenance recommendation method and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110929149B (en) |
Families Citing this family (14)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111985653B (en) * | 2020-06-24 | 2022-08-26 | 国网江苏省电力有限公司 | Power grid fault knowledge recommendation and knowledge management system and method based on knowledge graph |
| CN113868427A (en) * | 2020-06-30 | 2021-12-31 | 中国移动通信集团天津有限公司 | Data processing method and device and electronic equipment |
| CN111882084A (en) * | 2020-07-27 | 2020-11-03 | 西门子工厂自动化工程有限公司 | Method and apparatus for performing maintenance on field devices in a plant |
| CN112328875B (en) * | 2020-10-30 | 2024-07-12 | 北京精密机电控制设备研究所 | Equipment intelligent maintenance scheme recommendation method based on big data |
| CN112417175A (en) * | 2020-12-07 | 2021-02-26 | 北京明略软件系统有限公司 | Maintenance work order-oriented document retrieval method, system, computer and storage medium |
| CN112541600A (en) * | 2020-12-07 | 2021-03-23 | 上海电科智能系统股份有限公司 | Knowledge graph-based auxiliary maintenance decision method |
| CN113139659B (en) * | 2021-04-09 | 2024-07-05 | 智科云创(北京)科技有限公司 | Water conservancy monitoring method and system based on digital twin |
| CN113268604B (en) * | 2021-05-19 | 2024-06-07 | 国网辽宁省电力有限公司 | Knowledge base self-adaptive expansion method and system |
| CN113821365A (en) * | 2021-09-02 | 2021-12-21 | 上海羚数智能科技有限公司 | Method for recommending and updating knowledge for exception handling in production field |
| CN114218402B (en) * | 2021-12-17 | 2024-05-28 | 迈创企业管理服务股份有限公司 | Method for recommending computer hardware fault replacement parts |
| CN114707910B (en) * | 2022-05-24 | 2022-09-16 | 深圳市信润富联数字科技有限公司 | Circuit board intelligent maintenance system based on industrial internet |
| CN115146081B (en) * | 2022-08-31 | 2022-12-09 | 合肥中科迪宏自动化有限公司 | Construction method and diagnosis method of fault diagnosis knowledge graph of production equipment |
| CN116258484B (en) * | 2023-05-16 | 2023-08-18 | 青岛创新奇智科技集团股份有限公司 | Method, device, terminal equipment and medium for generating predictive operation and maintenance scheme of equipment |
| CN116664148B (en) * | 2023-08-01 | 2023-10-13 | 南京大全电气研究院有限公司 | After-sales defect elimination management method and device and electronic equipment |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106529684A (en) * | 2016-10-19 | 2017-03-22 | 华中科技大学 | Maintenance decision-making system for numerical control machine tool and method thereof |
| CN109471929A (en) * | 2018-11-06 | 2019-03-15 | 湖南云智迅联科技发展有限公司 | A method of it is matched based on map and carries out equipment maintenance record semantic search |
-
2019
- 2019-11-12 CN CN201911102558.XA patent/CN110929149B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106529684A (en) * | 2016-10-19 | 2017-03-22 | 华中科技大学 | Maintenance decision-making system for numerical control machine tool and method thereof |
| CN109471929A (en) * | 2018-11-06 | 2019-03-15 | 湖南云智迅联科技发展有限公司 | A method of it is matched based on map and carries out equipment maintenance record semantic search |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110929149A (en) | 2020-03-27 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110929149B (en) | Industrial equipment fault maintenance recommendation method and system | |
| CN108304468B (en) | Text classification method and text classification device | |
| CN107392143B (en) | Resume accurate analysis method based on SVM text classification | |
| CN110008472A (en) | A kind of method, apparatus, equipment and computer readable storage medium that entity extracts | |
| CN111353306B (en) | Entity relationship and dependency Tree-LSTM-based combined event extraction method | |
| CN112966525B (en) | Law field event extraction method based on pre-training model and convolutional neural network algorithm | |
| CN113268974A (en) | Method, device and equipment for marking pronunciations of polyphones and storage medium | |
| CN113343701B (en) | Extraction method and device for text named entities of power equipment fault defects | |
| CN113434688B (en) | Data processing method and device for public opinion classification model training | |
| CN113934909A (en) | Financial event extraction method based on pre-training language and deep learning model | |
| CN112036185B (en) | Method and device for constructing named entity recognition model based on industrial enterprise | |
| CN114911945A (en) | Knowledge graph-based multi-value chain data management auxiliary decision model construction method | |
| CN113157859A (en) | Event detection method based on upper concept information | |
| CN114970529A (en) | Weakly supervised and interpretable training of machine learning based Named Entity Recognition (NER) mechanisms | |
| CN114491024A (en) | Small sample-based specific field multi-label text classification method | |
| Anandika et al. | A study on machine learning approaches for named entity recognition | |
| CN115357700A (en) | Event map-based compressor fault diagnosis method | |
| CN113901224A (en) | Knowledge distillation-based secret-related text recognition model training method, system and device | |
| CN113065352B (en) | Method for identifying operation content of power grid dispatching work text | |
| CN115455196A (en) | Design map construction method based on RFPC (radio frequency personal computer) concept design framework | |
| CN116362247A (en) | Entity extraction method based on MRC framework | |
| CN115827871A (en) | Internet enterprise classification method, device and system | |
| CN115238093A (en) | Model training method and device, electronic equipment and storage medium | |
| CN112199114B (en) | Software defect report distribution method based on self-attention mechanism | |
| CN114372148A (en) | Data processing method based on knowledge graph technology and terminal equipment |
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 |