CN113393084A - Operation ticket flow management system - Google Patents
Operation ticket flow management system Download PDFInfo
- Publication number
- CN113393084A CN113393084A CN202110525140.0A CN202110525140A CN113393084A CN 113393084 A CN113393084 A CN 113393084A CN 202110525140 A CN202110525140 A CN 202110525140A CN 113393084 A CN113393084 A CN 113393084A
- Authority
- CN
- China
- Prior art keywords
- job
- electronic
- ticket
- algorithm
- knowledge
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000010801 machine learning Methods 0.000 claims abstract description 22
- 238000012544 monitoring process Methods 0.000 claims abstract description 20
- 238000000605 extraction Methods 0.000 claims abstract description 17
- 230000006870 function Effects 0.000 claims abstract description 14
- 239000000284 extract Substances 0.000 claims abstract description 5
- 238000000034 method Methods 0.000 claims description 73
- 230000008569 process Effects 0.000 claims description 58
- 238000013135 deep learning Methods 0.000 claims description 25
- 238000012549 training Methods 0.000 claims description 24
- 238000004364 calculation method Methods 0.000 claims description 5
- 238000010276 construction Methods 0.000 claims description 5
- 238000001514 detection method Methods 0.000 claims description 5
- 230000001419 dependent effect Effects 0.000 claims description 3
- 238000001914 filtration Methods 0.000 claims description 3
- 238000005457 optimization Methods 0.000 claims description 3
- 230000011218 segmentation Effects 0.000 claims description 2
- 238000007689 inspection Methods 0.000 description 4
- 238000004519 manufacturing process Methods 0.000 description 3
- 238000012423 maintenance Methods 0.000 description 2
- 239000000463 material Substances 0.000 description 2
- 238000004458 analytical method Methods 0.000 description 1
- 238000013459 approach Methods 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000013479 data entry Methods 0.000 description 1
- 238000009826 distribution Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000005286 illumination Methods 0.000 description 1
- 238000009776 industrial production Methods 0.000 description 1
- 230000010354 integration Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000002093 peripheral effect Effects 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 239000013589 supplement Substances 0.000 description 1
- 230000001502 supplementing effect Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
- G06Q10/0631—Resource planning, allocation, distributing or scheduling for enterprises or organisations
-
- 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
- G06F16/33—Querying
- G06F16/3331—Query processing
- G06F16/334—Query execution
- G06F16/3344—Query execution using natural language analysis
-
- 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
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
- G06Q10/0635—Risk analysis of enterprise or organisation activities
-
- 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
- G06Q10/10—Office automation; Time management
- G06Q10/103—Workflow collaboration or project management
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Systems or methods specially adapted for specific business sectors, e.g. utilities or tourism
- G06Q50/06—Electricity, gas or water supply
Abstract
The invention provides a job ticket flow management system, which comprises an automatic industry knowledge extraction module, a business ticket flow management module and a business ticket flow management module, wherein the automatic industry knowledge extraction module automatically extracts entities from electronic texts of industry files and relations among the entities, and automatically constructs a structured industry knowledge graph; the workflow and safety risk proposing module based on the knowledge map utilizes the knowledge map to provide an electronic job ticket template containing the job standardized workflow proposal and the safety risk prompt for the given text description; the electronic operation ticket management module completes the functions of issuing, examining and approving electronic operation tickets, counting reports and the like; the automatic flow specification and safety risk monitoring module automatically searches, arranges and deploys the existing machine learning algorithm in the system, and realizes automatic monitoring and alarming on the operation flow and the safety risk. The invention improves the efficiency and quality of operation supervision.
Description
Technical Field
The invention relates to a job ticket management system, in particular to a job ticket process management system.
Background
The operation ticket system is a basic system for ensuring industrial safety production, and is an effective safety measure for preventing misoperation and danger in the production operation management process. A generic job ticket typically contains the contents of the production job, the steps, the security risks, the personnel involved and the approval part. The traditional paper operation ticket has great problems in issuing, checking, implementing and filing, and is gradually replaced by an electronic operation ticket.
Compared with paper operation tickets, the electronic operation tickets naturally have the advantages of easy backup and filing. Besides, a database is established in advance for the specific content, steps and security risks of common temporary jobs. The electronic job ticket system can ensure the accuracy and reliability of the content when the job ticket is issued. The conventional electronic operation ticket system is generally integrated with a process management system, so that the automation of the process of examination and approval of each item in operation is realized to a certain degree, and the phenomena of signature replacement, signature change and signature supplement are avoided. Through integration with a reporting system, the existing electronic job ticket system can also make an analysis report on the historical electronic job ticket, thereby helping a manager to better master the implementation condition of the historical job. However, the existing electronic operation ticket system has a series of problems facing the final target of the industrial production safety automation process management.
When a conventional electronic job ticket system issues a job ticket for a specific job, information such as job contents, steps, security risks, and the like in the job ticket is retrieved from a database created manually. The contents of the database are retrieved, structured and imported from expert or specification files by developers of the management system. Since the existing database is fixed, it is necessary to re-establish data entries for new jobs whenever a job ticket needs to be generated for that job. In addition, when an industry specification is updated, the database contents also need to be manually updated. This increases the difficulty of use of the system and also causes the traditional electronic job ticket system to always lag behind the updating of the industry specifications and job categories.
After the electronic operation ticket and the process management system are introduced, although the problems of operation ticket signing substitution, changing and supplementing can be solved, the effects of implementation, supervision and risk identification still depend on whether the responsible person fully functions to verify various process requirements and specifications in the operation ticket. Therefore, the electronic job ticket system itself does not actually have a function of preventing or supervising security risks due to staff's job loss or operation violations against the contents of the job ticket, even if such risk information is already embodied in the job ticket.
The actual conditions in the actual operation are complex, and even if the operation is the same, the safety specifications and requirements are different under the actual conditions of different scenes, weather, peripheral equipment distribution and the like. Therefore, it is difficult for the conventional electronic job ticket to completely list the safety risk information of the job in the current scene, and the content of the job ticket cannot be automatically adjusted according to the specific real condition. Even if it is possible to list as many security risks as possible for some jobs, the number of items of security risks is enormous due to the need to take care of too many real-world situations, which undoubtedly increases the burden on the supervising personnel.
In order to solve a series of problems of difficult database update, high safety risk and heavy burden of supervision personnel in the conventional electronic operation ticket system, the invention needs to invent an intelligent operation ticket full-flow management system which comprises automatic acquisition of operation ticket content and automatic generation of operation ticket safety specifications and automatic inspection of the operation ticket safety specifications, thereby realizing full-flow safety management from issuing and approval of electronic operation tickets to automatic safety inspection.
Disclosure of Invention
The invention aims to solve a series of problems of difficult database updating, high safety risk and heavy burden of supervision personnel in the existing electronic operation ticket system, and provides an operation ticket flow management system which comprises operation flow safety specifications and can automatically acquire operation ticket contents and automatically generate automatic inspection of the operation flow safety specifications, so that the whole-flow safety management from issuing, examining and approving of electronic operation tickets to automatic safety inspection is realized.
The invention provides a job ticket flow management system for realizing the purpose, which comprises an automatic industry knowledge extraction module, a knowledge graph-based job flow and safety risk proposing module, an electronic job ticket management module and an automatic flow specification and safety risk monitoring module.
The automatic industry knowledge extraction module automatically extracts entities and the relation between the entities from electronic texts of industry files such as operation specifications, safety specifications and the like, so that a structured industry knowledge map containing industry knowledge is automatically constructed; the knowledge map consists of an implicit knowledge map and an explicit knowledge map; the pre-trained deep learning language model forms an implicit knowledge map, and an entity automatically extracted from the language model and a traditional map structure knowledge map constructed by relations form an explicit knowledge map; the automatic extraction of entities and relationships between entities is done using a transform-based deep learning language model, such as Bert and GPT.
The knowledge-based workflow and security risk proposal module utilizes implicit and explicit knowledge maps to present an electronic job ticket template containing the job-standardized workflow recommendations and security risk tips for a given textual description of an industrial job.
And the electronic operation ticket management module is used for completing functions of issuing, examining and approving, counting reports and the like of the electronic operation ticket.
And the automatic flow specification and safety risk monitoring module is used for automatically searching, arranging and deploying the existing machine learning algorithm in the system according to the flow specification and the natural language text description of the safety risk in the electronic job ticket, so that the automatic monitoring and alarming of the flow and the safety risk are realized.
The training process of the automatic industry knowledge extraction module comprises the following steps: firstly, pre-training a deep learning language model by using Chinese electronic text data of a large number of industry files through two pre-training tasks of 'missing text prediction' and 'performing following prediction by using the former text', so as to obtain a pre-training model containing industry knowledge; after the pre-training model is obtained, the implicit knowledge map is an encoder part of the deep learning language model; then, automatically constructing a structured explicit knowledge graph by using an attention module and a specific filtering rule in a pre-training model; after a new industry file is delivered, the automatic industry knowledge extraction module can automatically utilize the new industry file to finely adjust the existing model and reconstruct the explicit knowledge graph, so that the full-automatic construction and updating of the industry knowledge graph are realized.
The construction process of the knowledge graph-based operation process and the safety risk proposing module comprises the following steps: establishing an electronic job ticket database which is composed of a large number of triple text data of job description, job flow and potential safety risk; inquiring operation flow and safety risk information from a knowledge map which is formed by an explicit knowledge map and an implicit knowledge map; describing the operation flow and the safety risk information by using a natural language text; after the natural language text descriptions of the operation flow and the safety risk information inferred from the explicit knowledge map and the implicit knowledge map are obtained respectively, the similarity between the operation flow and the safety risk information is calculated, repeated information is removed, and therefore the final operation flow and safety risk suggestion are obtained.
The method for inquiring the explicit knowledge map operation flow and the safety risk information is a traditional NLP-based method. The method for inquiring the operation flow and the safety risk information of the explicit knowledge graph is a process from operation text description to knowledge graph inquiry sentences based on a pre-trained deep learning language model. The pre-training process uses a binary dataset consisting of job text descriptions and corresponding query statements. The workflow and security risk information of the explicit knowledge-graph are jointly generated by a traditional NLP-based method and a pre-trained deep learning language model-based method, and corresponding natural language text descriptions are generated by fixed text templates. The natural language text description of the workflow and security risk information of the implicit knowledgemap is generated directly from the job text description by a decoder. The decoder is trained by adopting triple text data of operation description, operation flow and potential safety risk, parameters of the implicit knowledge map are fixed and unchanged in the training process, and only part of the decoder participates in the training optimization process.
The knowledge graph-based job flow and security risk proposing module can automatically generate flow and security risk information natural language text description corresponding to the job after a user submits a new job and text description thereof, so as to be used for generating an electronic job ticket by a traditional electronic job ticket management module.
The electronic operation ticket management module utilizes the sensor to acquire sensing information and a machine learning algorithm to automatically supervise the operation and can give an alarm to a supervision result. The sensor comprises a camera, a thermometer and a gas detection device; the sensing information types include a photograph and a numerical value. The content of the electronic job ticket comes from a knowledge-graph-based workflow and security risk proposing module.
The operation ticket flow management system can automatically or semi-automatically complete operation supervision by an automatic flow specification and safety risk monitoring module according to the flow and safety risk information in the electronic operation ticket by using a machine learning algorithm.
The automatic flow specification and safety risk monitoring module automatically searches, arranges and deploys the existing machine learning algorithm in the system according to the flow specification and the natural language text description of the safety risk in the electronic job ticket, and automatically monitors and alarms the flow and the safety risk; the process is as follows: firstly, establishing a lookup table for a machine learning algorithm according to functions, use conditions and performance indexes; after a lookup table of the completed algorithm information is constructed, a natural language text description containing the operation flow specification and the safety risk, a corresponding query statement of the algorithm lookup table, the starting condition of the algorithm, the input requirement of the algorithm, the output operation of the algorithm and the ending condition of the algorithm are collected and labeled to form a six-element data set; then, the data are used for finely adjusting the deep learning language model, so that an automatic algorithm searching and arranging model based on the deep learning language model is obtained; and finally, establishing a machine learning algorithm execution engine, wherein the execution engine searches and arranges the output result of the model by using an automatic algorithm, arranges the sequence of algorithm execution, calls the corresponding algorithm according to the starting and ending conditions, applies the required logic and numerical operation to the result directly output by the algorithm, and stores the execution results of all the algorithms in a global cache for other algorithms or an electronic job ticket management module to read.
For a given electronic operation ticket, the operation flow specification and the safety risk in the electronic operation ticket are automatically analyzed item by item, all machine learning algorithms, the starting condition, the dependent algorithm, the output operation and the ending condition of each algorithm, which are required for realizing supervision of the electronic operation ticket, are searched, all algorithms are arranged and executed, the execution results of all algorithms are finally output, and whether the corresponding operation flow specification and the safety risk item are in accordance with the actual algorithm detection calculation results or not is obtained. The results can be called by the electronic job ticket management module and can be recorded, alarmed and the like according to user-defined settings.
The operation ticket flow management system has the advantages that the operation ticket flow management system can realize full-flow closed-loop operation management and control, reduce the labor and material cost of deployment and subsequent maintenance, has low system operation difficulty, and improves the efficiency and quality of operation supervision.
In order to make the aforementioned and other objects, features and advantages of the present invention comprehensible, preferred embodiments accompanied with figures are described in detail below.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, and it is obvious that the drawings in the following description are some embodiments of the present invention, and other drawings can be obtained by those skilled in the art without creative efforts.
Fig. 1 is a schematic view of the working process of the present invention.
Detailed Description
The technical solution in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention. It is to be understood that the described embodiments are merely exemplary of the invention, and not restrictive of the full scope of the invention. The components of embodiments of the present invention generally described and illustrated in the figures herein may be arranged and designed in a wide variety of different configurations. Thus, the following detailed description of the embodiments of the present invention, presented in the figures, is not intended to limit the scope of the invention, as claimed, but is merely representative of selected embodiments of the invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments of the present invention without making any creative effort, shall fall within the protection scope of the present invention.
The present invention is described in further detail below to enable those skilled in the art to practice the invention with reference to the description.
Referring to fig. 1, fig. 1 is a schematic view of a work flow of an embodiment of the present invention. As shown in fig. 1, the job ticket flow management system of this embodiment includes an automation industry knowledge extraction module, a knowledge graph-based job flow and security risk proposing module, an electronic job ticket management module, and an automation flow specification and security risk monitoring module.
In this embodiment, the automated industry knowledge extraction module automatically extracts entities and relationships between the entities from electronic texts of industry files such as job specifications, safety specifications, and the like, thereby automatically constructing a structured industry knowledge graph containing industry knowledge. The knowledge map based workflow and security risk proposal module uses the knowledge map to establish an electronic job ticket database for the text description of the given industrial job and provides an electronic job ticket template containing the job standardized workflow proposal and security risk prompt. The automatic flow specification and safety risk monitoring module automatically searches, arranges and deploys the existing machine learning algorithm in the system according to the flow specification and the natural language text description of the safety risk in the electronic job ticket, thereby realizing the automatic monitoring and alarming of the operation flow and the safety risk. The electronic operation ticket management module obtains sensing information from the sensor, completes functions of issuing, examining and approving electronic operation tickets, counting reports and the like according to self-recognition language text description and automatic flow specification of the electronic operation ticket template and a calculation result of the safety risk monitoring module, and performs operations of recording, alarming and the like according to user-defined setting.
In some embodiments, the automated industry knowledge extraction module automatically extracts entities and relationships between the entities from electronic texts of industry files such as job specifications, safety specifications and the like, so as to automatically construct a structured industry knowledge graph containing industry knowledge. The module utilizes a group of non-supervised and supervised learning transform-based deep learning language models (such as Bert and GPT) to complete automatic entity and relationship extraction, and the knowledge graph is formed by a pre-trained deep learning language model and a traditional graph structure knowledge graph constructed by automatically extracting entities and relationships from the language model, and is respectively referred to as an implicit knowledge graph and an explicit knowledge graph in the following. Taking the Bert as an example of a deep learning language model in the module, we pre-train the Bert by two pre-training tasks of ' missing text prediction ' and ' performing the following prediction by using the former text in Chinese electronic text data of a large number of industry files, so as to obtain the Bert pre-training model containing industry knowledge. After the pre-training model is obtained, the implicit knowledge map is the encoder part of the model. We then use the attention module in the pre-trained model and specific filtering rules to automatically build a structured explicit knowledge-graph. In the actual application process, after a new industry file is delivered, the module can automatically utilize the texts to finely adjust the existing model and reconstruct the explicit knowledge graph, so that the full-automatic construction and updating of the industry knowledge graph are realized.
The knowledgeable-based workflow and security risk proposal module of certain embodiments is an implicit and explicit knowledgeable graph constructed using an automated industry knowledge extraction module that, for a textual description of a given industrial job, gives an electronic job ticket template containing the job's standardized workflow suggestions and security risk tips. To build this module, first, we build an electronic job ticket database, consisting of a large number of triple text data of job descriptions, job flows and potential security risks. For explicit knowledge graphs, two methods are used simultaneously to query workflow and security risk information from them. The first method utilizes a traditional NLP-based method, which firstly carries out word segmentation and part-of-speech tagging on the text description of the operation and generates a query sentence according to predefined keywords and rule matching. The second approach implements the process from job text description to knowledge-graph query statement based on a pre-trained deep learning language model (e.g., Bert and GPT), which uses the binary dataset we collect consisting of job text descriptions and corresponding query statements. The query sentences generated by the two methods are commonly used for querying the flow and safety risk information of the operation from the real knowledge graph, and the information generates corresponding natural language text description through a fixed text template. For implicit knowledge maps, we connect a decoder behind it, which is responsible for generating natural language text descriptions of workflow and security risk information directly from the job text description. The decoder is trained by using the collected triple text data of the operation description, the operation flow and the potential safety risk, in the training process, parameters of an implicit knowledge map (namely an encoder part of a depth language model in the module I) are fixed and unchanged, and only the decoder part participates in the training optimization process. After the workflow inferred from the reality knowledge map and the safety risk information natural language text description are obtained respectively, similarity between every two of the information is further calculated, repeated information is removed, and therefore the final workflow and safety risk suggestion are obtained. In the practical application process, after a user submits a new job and a text description thereof in the electronic job ticket management module, the module can automatically generate a natural language text description of the flow and the safety risk information of the corresponding job so as to be used for generating the electronic job ticket by the traditional electronic job ticket management module.
The electronic job ticket management module of some embodiments has substantially the same function as the traditional electronic job ticket management system, and mainly completes the functions of issuing, examining and approving electronic job tickets, counting reports and the like, and the main difference between the functions is as follows:
(1) a sensor component and a supervision alarm function are newly added. Because the system mainly utilizes the sensing information and a machine learning algorithm to automatically supervise the operation, the system is additionally provided with the functions of accessing sensors (various cameras, temperature, gas and the like), viewing the sensing content (pictures and numerical values) and alarming the supervision result.
(2) The source of the job ticket content. In a conventional electronic job ticket management system, the content of the job ticket is from a predefined content template; in the system, the contents come from a workflow and security risk proposing module based on a knowledge graph.
(3) Supervisory means and logic. In a traditional electronic job ticket management system, the implementation and supervision processes of jobs are not in the management and control range of the system; in the system, the supervision of the operation can be automatically/semi-automatically completed by an automatic process specification and safety risk monitoring module according to the process and safety risk information in the electronic operation ticket by using a machine learning algorithm.
The automatic flow specification and safety risk monitoring module of some embodiments is used for automatically searching, arranging and deploying the existing machine learning algorithm in the system according to the natural language text description of the flow specification and safety risk in the electronic job ticket, so as to realize automatic monitoring and alarming of the flow and safety risk. To implement this module, we first build a look-up table for the machine learning algorithm according to function, usage conditions and performance indicators. In functional aspects, the lookup table content includes the target objects of the algorithm (e.g., people, equipment, safety equipment, people relationships, object-object relationships, etc.), functional definitions (e.g., location information, presence, category determination, numerical regression, prediction, etc.); in the aspect of using conditions, the contents of the lookup table comprise limiting conditions of using weather, scenes, illumination, camera types, camera poses, calculation force requirements and the like of the algorithm; in the aspect of performance indexes, the contents of the lookup table comprise performance parameters such as the expected accuracy, the false alarm rate and the operation speed of the algorithm. After a lookup table of algorithm information is constructed, a data set of a six-element group consisting of a natural language text description containing a work flow specification and a safety risk, a corresponding algorithm lookup table query statement, an algorithm starting condition, an algorithm input requirement, an algorithm output operation and an algorithm ending condition is collected and labeled. Then, we use these data to fine-tune the deep-learning language model (e.g. Bert and GPT, the chinese electronic text data already in large industry documents is pre-trained according to the pre-training method described in module one), so as to obtain an automatic algorithm search and arrangement model based on the deep-learning language model. Finally, a machine learning algorithm execution engine is established in the module, the execution engine searches and arranges the output results of the model by using an automatic algorithm, arranges the sequence of algorithm execution, calls corresponding algorithms according to starting and ending conditions, applies required logic and numerical operation to the results directly output by the algorithms, and stores the execution results of all the algorithms in a global cache for other algorithms or an electronic job ticket management module to read. In the using process, for a given electronic job ticket, the module can automatically analyze job flow specifications and security risks in the electronic job ticket item by item, search all machine learning algorithms, starting conditions, dependent algorithms, output operation and ending conditions of each algorithm required for realizing supervision of the electronic job ticket, arrange and execute all algorithms, finally output execution results of all algorithms, and obtain a calculation result whether corresponding job flow specifications and security risk items conform to actual algorithm detection. The results can be called by the electronic job ticket management module and can be recorded, alarmed and the like according to user-defined settings.
The operation ticket flow management system can realize the full-flow closed-loop operation management and control, reduce the labor and material cost of deployment and subsequent maintenance, has low system operation difficulty, and improves the efficiency and quality of operation supervision.
While embodiments of the invention have been disclosed above, it is not intended to be limited to the uses set forth in the specification and examples. It can be applied to all kinds of fields suitable for the present invention. Additional modifications will readily occur to those skilled in the art. Therefore, the invention is not to be limited to the specific details and illustrations shown and described herein, without departing from the general concept defined by the claims and their equivalents.
Claims (18)
1. The job ticket flow management system comprises
The automatic industry knowledge extraction module automatically extracts entities and the relation between the entities from electronic texts of industry files such as operation specifications, safety specifications and the like, so that a structured industry knowledge map containing industry knowledge is automatically constructed; the knowledge map consists of an implicit knowledge map and an explicit knowledge map; the pre-trained deep learning language model forms an implicit knowledge map, and the entity automatically extracted from the language model and the traditional map structure knowledge map of the relation construction form an explicit knowledge map. The automatic extraction of the entities and the relationship between the entities is completed by utilizing a deep learning language model.
A knowledge map-based workflow and security risk proposing module, which utilizes the implicit knowledge map and the explicit knowledge map to give an electronic job ticket template containing a standardized workflow proposal and a security risk prompt of the job for the text description of the given industrial job;
the electronic operation ticket management module is used for completing functions of issuing, examining and approving electronic operation tickets, counting reports and the like;
and the automatic flow specification and safety risk monitoring module is used for automatically searching, arranging and deploying the existing machine learning algorithm in the system according to the flow specification and the natural language text description of the safety risk in the electronic job ticket, so that the automatic monitoring and alarming of the flow and the safety risk are realized.
2. The job ticket flow management system of claim 1 wherein the deep learning language model of the automated industry knowledge extraction module is transform-based.
3. The job ticket process management system of claim 1 wherein the deep learning language model is a Bert model.
4. The job ticket process management system of claim 1 wherein the deep learning language model is a GPT model.
5. The job ticket process management system of claim 1, wherein the training process of the automated industry knowledge extraction module is: firstly, pre-training a deep learning language model by using electronic text data through two pre-training tasks of 'missing text prediction' and 'performing following prediction by using the former text', so as to obtain a pre-training model containing industry knowledge; after the pre-training model is obtained, the implicit knowledge map is an encoder part of the deep learning language model; then, automatically constructing a structured explicit knowledge graph by using an attention module and a specific filtering rule in a pre-training model; after a new industry file is delivered, the automatic industry knowledge extraction module can automatically utilize the new industry file to finely adjust the existing model and reconstruct the explicit knowledge graph, so that the full-automatic construction and updating of the industry knowledge graph are realized.
6. The work ticket process management system of claim 1 wherein the knowledge-graph based work process and security risk proposal module is constructed by: establishing an electronic job ticket database which is composed of a large number of triple text data of job description, job flow and potential safety risk; inquiring operation flow and safety risk information from a knowledge map which is formed by an explicit knowledge map and an implicit knowledge map; describing the operation flow and the safety risk information by using a natural language text; after the natural language text descriptions of the operation flow and the safety risk information inferred from the explicit knowledge map and the implicit knowledge map are obtained respectively, the similarity between the operation flow and the safety risk information is calculated, repeated information is removed, and therefore the final operation flow and safety risk suggestion are obtained.
7. The work ticket process management system of claim 6 wherein the explicit graph work process and security risk information query method is a traditional NLP-based method that first performs word segmentation and part-of-speech tagging on the text description of the job and generates a query statement according to predefined keyword and rule matching.
8. The work ticket process management system of claim 6 wherein the explicit knowledge-graph workflow and security risk information query method is based on a pre-trained deep learning language model to implement a process from job text description to knowledge-graph query statement; the pre-training process uses a binary dataset consisting of job text descriptions and corresponding query statements.
9. The job ticket process management system according to claims 7 and 8, wherein the explicit knowledge-graph workflow and security risk information are jointly generated by a conventional NLP-based method and a pre-trained deep learning language model-based method, and the corresponding natural language text description is generated by a fixed text template.
10. The job ticket process management system of claim 6 wherein the natural language textual description of the implicit knowledgebase workflow and security risk information is generated directly from the job textual description by a decoder.
11. The job ticket process management system of claim 10 wherein the decoder is trained using triple text data of job description, job process and potential security risk, and during the training process, parameters of the implicit knowledge map are fixed and only part of the decoder participates in the training optimization process.
12. The work ticket flow management system of claim 1 wherein the knowledge-graph based work flow and security risk suggestion module automatically generates a natural language text description of the flow and security risk information of the corresponding job for a legacy electronic work ticket management module to generate an electronic work ticket after a user submits a new job and its text description.
13. The job ticket process management system of claim 1, wherein the electronic job ticket management module utilizes sensors to obtain sensing information and machine learning algorithm to automatically supervise the job and can alarm the supervision result.
14. The job ticket process management system of claim 13 wherein the sensors include a camera, a thermometer, a gas detection device; the sensing information types include a photograph and a numerical value.
15. The job ticket process management system of claim 1 wherein the content of the electronic job ticket is derived from a knowledge-graph based job process and security risk proposal module.
16. The job ticket process management system of claim 1 wherein the job ticket process management system is capable of automatically or semi-automatically performing job monitoring by an automated process specification and security risk monitoring module based on process and security risk information in the electronic job ticket using a machine learning algorithm.
17. The job ticket process management system of claim 1, wherein the automated process specification and security risk monitoring module automatically searches, arranges and deploys the existing machine learning algorithm in the system according to the job process specification and the natural language text description of the security risk in the electronic job ticket, and automatically monitors and alarms the job process and the security risk;
firstly, establishing a lookup table for a machine learning algorithm according to functions, use conditions and performance indexes;
after a lookup table of the completed algorithm information is constructed, collecting and marking a natural language text description containing the operation flow specification and the safety risk, a corresponding query statement of the algorithm lookup table, the starting condition of the algorithm, the input requirement of the algorithm, the output operation of the algorithm and the ending condition of the algorithm to form a six-element data set;
then, the data are used for finely adjusting the deep learning language model, so that an automatic algorithm searching and arranging model based on the deep learning language model is obtained;
and finally, establishing a machine learning algorithm execution engine, wherein the execution engine searches and arranges the output result of the model by using an automatic algorithm, arranges the sequence of algorithm execution, calls the corresponding algorithm according to the starting and ending conditions, applies the required logic and numerical operation to the result directly output by the algorithm, and stores the execution results of all the algorithms in a global cache for other algorithms or an electronic job ticket management module to read.
18. The job ticket process management system of claim 1, wherein the automated process specification and security risk monitoring module automatically analyzes the job process specification and security risk of a given electronic job ticket item by item, searches all machine learning algorithms, the starting condition, the dependent algorithm, the output operation and the ending condition of each algorithm required for realizing supervision of the electronic job ticket item, arranges and executes all algorithms, finally outputs the execution results of all algorithms, and obtains a calculation result whether the corresponding job process specification and security risk item conform to actual algorithm detection; the results can be called by the electronic job ticket management module and can be recorded, alarmed and the like according to user-defined settings.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110525140.0A CN113393084A (en) | 2021-05-13 | 2021-05-13 | Operation ticket flow management system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110525140.0A CN113393084A (en) | 2021-05-13 | 2021-05-13 | Operation ticket flow management system |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN113393084A true CN113393084A (en) | 2021-09-14 |
Family
ID=77617888
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110525140.0A Pending CN113393084A (en) | 2021-05-13 | 2021-05-13 | Operation ticket flow management system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113393084A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114970564A (en) * | 2022-06-16 | 2022-08-30 | 北京汉端科技有限公司 | Cloud platform based micro-service civil aviation safety management SOP system |
| US11811626B1 (en) | 2022-06-06 | 2023-11-07 | International Business Machines Corporation | Ticket knowledge graph enhancement |
Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108460136A (en) * | 2018-03-08 | 2018-08-28 | 国网福建省电力有限公司 | Electric power O＆M information knowledge map construction method |
| CN108875051A (en) * | 2018-06-28 | 2018-11-23 | 中译语通科技股份有限公司 | Knowledge mapping method for auto constructing and system towards magnanimity non-structured text |
| CN108920587A (en) * | 2018-06-26 | 2018-11-30 | 清华大学 | Merge the open field vision answering method and device of external knowledge |
| CN110275960A (en) * | 2019-06-11 | 2019-09-24 | 中国电子科技集团公司电子科学研究院 | Representation method and system based on the knowledge mapping and text information for censuring sentence |
| US20190340518A1 (en) * | 2018-05-04 | 2019-11-07 | Zestfinance, Inc. | Systems and methods for enriching modeling tools and infrastructure with semantics |
| CN110825884A (en) * | 2019-11-05 | 2020-02-21 | 腾讯科技（深圳）有限公司 | Embedded representation processing method and device based on artificial intelligence and electronic equipment |
| CN110889556A (en) * | 2019-11-28 | 2020-03-17 | 福建亿榕信息技术有限公司 | Enterprise operation risk prediction method and system |
| CN111177393A (en) * | 2020-01-02 | 2020-05-19 | 广东博智林机器人有限公司 | Knowledge graph construction method and device, electronic equipment and storage medium |
| CN111353013A (en) * | 2018-12-05 | 2020-06-30 | 中兴通讯股份有限公司 | Method and system for realizing intelligent delivery and reception |
| CN112765486A (en) * | 2021-01-22 | 2021-05-07 | 重庆邮电大学 | Knowledge graph fused attention mechanism movie recommendation method |
-
2021
- 2021-05-13 CN CN202110525140.0A patent/CN113393084A/en active Pending
Patent Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108460136A (en) * | 2018-03-08 | 2018-08-28 | 国网福建省电力有限公司 | Electric power O＆M information knowledge map construction method |
| US20190340518A1 (en) * | 2018-05-04 | 2019-11-07 | Zestfinance, Inc. | Systems and methods for enriching modeling tools and infrastructure with semantics |
| CN108920587A (en) * | 2018-06-26 | 2018-11-30 | 清华大学 | Merge the open field vision answering method and device of external knowledge |
| CN108875051A (en) * | 2018-06-28 | 2018-11-23 | 中译语通科技股份有限公司 | Knowledge mapping method for auto constructing and system towards magnanimity non-structured text |
| CN111353013A (en) * | 2018-12-05 | 2020-06-30 | 中兴通讯股份有限公司 | Method and system for realizing intelligent delivery and reception |
| CN110275960A (en) * | 2019-06-11 | 2019-09-24 | 中国电子科技集团公司电子科学研究院 | Representation method and system based on the knowledge mapping and text information for censuring sentence |
| CN110825884A (en) * | 2019-11-05 | 2020-02-21 | 腾讯科技（深圳）有限公司 | Embedded representation processing method and device based on artificial intelligence and electronic equipment |
| CN110889556A (en) * | 2019-11-28 | 2020-03-17 | 福建亿榕信息技术有限公司 | Enterprise operation risk prediction method and system |
| CN111177393A (en) * | 2020-01-02 | 2020-05-19 | 广东博智林机器人有限公司 | Knowledge graph construction method and device, electronic equipment and storage medium |
| CN112765486A (en) * | 2021-01-22 | 2021-05-07 | 重庆邮电大学 | Knowledge graph fused attention mechanism movie recommendation method |
Non-Patent Citations (2)
| Title |
|---|
| 尚鸿志;刘玉东;: "国内外作业许可制度建立与实施的初步探讨", 中国安全生产科学技术, no. 2, 15 December 2012 (2012-12-15) * |
| 曾厉: "基于智能成票规则及算法的电气工作票系统", 《电工技术》, 25 December 2019 (2019-12-25) * |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US11811626B1 (en) | 2022-06-06 | 2023-11-07 | International Business Machines Corporation | Ticket knowledge graph enhancement |
| CN114970564A (en) * | 2022-06-16 | 2022-08-30 | 北京汉端科技有限公司 | Cloud platform based micro-service civil aviation safety management SOP system |
| CN114970564B (en) * | 2022-06-16 | 2023-02-03 | 北京汉端科技有限公司 | Cloud platform based micro-service civil aviation safety management SOP system |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Wu et al. | Literature review and prospect of the development and application of FMEA in manufacturing industry | |
| CN107147639B (en) | A kind of actual time safety method for early warning based on Complex event processing | |
| CN112612902A (en) | Knowledge graph construction method and device for power grid main device | |
| CN111460167A (en) | Method for positioning pollution discharge object based on knowledge graph and related equipment | |
| CN106557541A (en) | For the apparatus and method for automatically analyzing of the execution data of product bug detection | |
| CN106067094A (en) | A kind of dynamic assessment method and system | |
| Yim et al. | Using process topology in plant-wide control loop performance assessment | |
| CN113393084A (en) | Operation ticket flow management system | |
| Li et al. | A policy-based process mining framework: mining business policy texts for discovering process models | |
| CN113656805B (en) | Event map automatic construction method and system for multi-source vulnerability information | |
| CN114417015B (en) | High-speed train maintainability knowledge graph construction method | |
| CN111183421A (en) | Service providing system, business analysis support system, method, and program | |
| Almatared et al. | Digital twin in the architecture, engineering, and construction industry: A bibliometric review | |
| Zhang et al. | Decision support oriented ontological modeling of product knowledge | |
| CN115511119A (en) | Intelligent diagnosis method and system for heat supply system based on knowledge map and Bayes | |
| Rajbahadur et al. | Pitfalls analyzer: quality control for model-driven data science pipelines | |
| JP2006236299A (en) | Integrated knowledge based system | |
| JP2018147351A (en) | Knowledge model construction system and knowledge model construction method | |
| Samosir et al. | Identifying Requirements Association Based on Class Diagram Using Semantic Similarity | |
| Yurin et al. | Semi-automated formalization and representation of the engineering knowledge extracted from spreadsheet data | |
| Liu et al. | EXPERIENCE: Algorithms and case study for explaining repairs with uniform profiles over IoT data | |
| CN114862233A (en) | Intelligent decision method and intelligent decision system | |
| WO2020141394A1 (en) | Method and system to generate control logic for performing industrial processes | |
| CN112668836A (en) | Risk graph-oriented associated risk evidence efficient mining and monitoring method and device | |
| Zhang et al. | Research Review of Design Pattern Mining |
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 |