TWM586422U - Factory operation and maintenance expert system - Google Patents

Abstract

This creation is a plant operation and maintenance expert system, which includes at least an inference engine, a knowledge base, a fact database, a deep learning or mechanical learning model, and a user interface. The fact database can be linked to maintenance information systems or decentralized control systems. The rules of its knowledge base can be established by deep learning models or mechanical learning models for users to consult the operation and maintenance of at least one device in the factory. Therefore, this creation can effectively save and inherit the relevant information and experience of operation and maintenance, and effectively analyze and use a large amount of historical maintenance records and operation data, and then derive new ideas and concepts. In particular, users can search through the user interface Or the question and answer method to get the real cause analysis of the equipment failure and related processing methods or the best operating conditions.

Description

Plant Operation and Maintenance Expert System

This creative department is about a plant operation and maintenance expert system, in particular, it can effectively save and inherit the relevant information and experience of operation and maintenance, and at the same time effectively analyze and use a large number of historical maintenance records and operation data, and then draw new ideas and concepts. Users You can get the real cause analysis of the equipment failure and related processing methods or the best operating conditions by searching or answering questions. The decentralized control system can send real-time operation data and get feedback suggestions through communication protocols. Intelligent effect.

In the conventional technology, the general maintenance information management system itself is an independent system. After the plant personnel judges that the equipment is abnormal, they manually open a maintenance work order in the system to raise maintenance requirements, and then have experience The maintenance personnel find out the true cause of the fault, and then troubleshoot it, and report the case to the system.

However, with the huge historical maintenance record data of the system, information including the cause of failure is stored in the system database, but they can only be analyzed and applied manually, so it is clear that the traditional maintenance information management system is no longer capable. Moreover, the general distributed control system (DCS) can only achieve the automation effect of Industry 3.0.

In addition, to find out the true cause of equipment failure often depends on the experience and judgment of maintenance personnel, but this experience may not be properly passed down. For big data, the facets and capabilities of purely manual analysis have become quite limited, and it is not easy to find the correlation between them, and it is impossible to dig out valuable information. In addition, without the connection with big data analysis, the intelligent effect of Industry 4.0 cannot be achieved.

Therefore, there is a great need for an innovative plant operation and maintenance expert system that can effectively preserve and inherit relevant data and experience of operation and maintenance, and effectively analyze and use a large number of historical maintenance records and operation data, and then derive new ideas and concepts, especially, using The user can obtain the real cause analysis of the equipment failure and related processing methods or the best operating conditions by searching or answering questions. The decentralized control system can send real-time operation data and get feedback suggestions through the communication protocol, and specifically achieve Industry 4.0. The intelligent effect of the machine reduces the maintenance personnel's fault cause judgment and maintenance time, improves the equipment and the entire plant operation rate and operation efficiency, and has equipment abnormality warning function to solve all the problems of the conventional technology.

The main purpose of this creation is to provide a factory operation and maintenance expert system, which at least includes an inference engine, a knowledge base, a fact database, a deep learning or mechanical learning model, and a user interface for users to consult the operation of at least one device in the factory And maintenance.

The user can input the question through the user interface, and compare the rules in the fact database and the knowledge base through the reasoning engine to determine the possible answers to the question. You can also use the user interface to perform additions, edits, and saves on the fact database and knowledge base.

The data source of the above fact database can be linked with the operation and maintenance manual, maintenance information management system or decentralized management system, and the rule source of the knowledge base can be the rules established by the fact database data through deep learning or mechanical learning models.

In addition, the decentralized control system can also connect to the user interface through the communication protocol, transmit real-time operation data, and compare the rules in the fact database and the knowledge base through the inference engine to reason and get recommendations for operation and maintenance. The corresponding instructions are transmitted to the distributed control system to control at least one device.

Therefore, this creation can effectively save and inherit the relevant information and experience of operation and maintenance, and effectively analyze and use a large number of historical maintenance records and operation data, and then derive new ideas and concepts. In particular, users can use search or question and answer methods to Obtain the cause analysis of the equipment failure and the related processing methods or the best operating conditions. The distributed control system can send real-time operation data and get feedback suggestions through the communication protocol to achieve the intelligent effect of Industry 4.0.

The following is a more detailed description of the implementation of this creation with illustrations and component symbols, so that those skilled in the art can implement it after studying this manual.

Please refer to the first figure, the schematic diagram of the factory operation and maintenance expert system of this creative embodiment. As shown in the first figure, the plant operation and maintenance expert system of this creation includes a knowledge base 10, an inference engine 20, a user interface 30, a deep learning or mechanical learning model 40, and a fact database 50 for users U to consult the factory Operation and maintenance recommendations for at least one device (not shown).

Specifically, the user U can input the question through the user interface 30, compare the rules in the fact database 50 and the knowledge base 10 through the inference engine 20, and infer a possible answer to the question, and present it in the user interface 30 . Alternatively, the user U may perform tasks such as adding, editing, and storing the fact database 50 and the knowledge base 10 through the user interface 30.

The knowledge base 10 stores the operation and maintenance related rules of at least one device in the factory, and the fact database 50 stores the operation, maintenance related data, files and data of at least one device in the factory. The inference engine 20 is connected to the knowledge base 10 and the fact database 50, reads the fact database 50 and the rules in the knowledge base 10 to compare and judge, and infers operation or maintenance suggestions.

The data source of the above fact database 50 can be linked with the operation and maintenance manual, the maintenance information management system 70 or the decentralized management system 60. The rule source of the knowledge database 10 can be the data of the fact database 50 via deep learning or mechanical learning model 40 Established rules.

The deep learning or mechanical learning model 40 is connected to the fact database 50, and converts, through deep learning or mechanical learning, data, documents and data related to the operation and maintenance of at least one device in the factory stored in the real database 50. The relevant rules are established and stored in the knowledge base 10.

The user interface 30 can be used to input questions related to operation or maintenance, transmit them to the inference engine 20, and be used to present operation or maintenance suggestions derived from the inference engine 20.

In addition, this creation can further receive real-time or non-real-time operation data from the at least one device in the factory of the Distributed Control System (DCS) 60, and automatically transmit the input to the user interface 30 through a communication protocol. The inference engine 20 compares and judges the rules in the fact database 50 and the knowledge base 10 to infer operation and maintenance recommendations, and may even return corresponding instructions to the distributed control system 60 to control at least one device.

The sources of the operation and maintenance data, documents and data in the fact database 50 include an operation and maintenance manual, a maintenance management information system (Maintenace Management Information System (MMIS) 70 or a decentralized management system 60).

More specifically, the inference engine 20 may automatically receive the operation status of the at least one device, compare the rules in the fact database 50 and the knowledge database 10, and automatically determine and provide fault early warning, maintenance reference or best operation instructions, and present To user interface 30. In addition, the user U can also obtain the cause analysis of the failure of the at least one device in a search or Q & A manner through the user interface 30.

In summary, the characteristics of this creation are that it can effectively save and inherit relevant information and experience of operation and maintenance, and at the same time effectively analyze and use a large number of historical maintenance records and operation data, and then derive new ideas and concepts. In particular, you can search or It is a question-and-answer method to obtain the real cause analysis of the equipment failure and related processing methods. The distributed control system can also send real-time operation data through communication protocols and get feedback suggestions to achieve the intelligent effect of Industry 4.0.

In addition, this creation can also reduce the maintenance personnel's fault cause judgment and maintenance time, improve the equipment and the entire plant operating rate and operating efficiency, with equipment abnormal warning function.

The above are only for explaining the preferred embodiment of the creation, and are not intended to restrict the creation in any form. Therefore, any modification or change related to the creation under the same spirit of creation , Should still be included in the scope of protection of this creative intention.

10‧‧‧ Knowledge Base
20‧‧‧Inference Engine
30‧‧‧user interface
40‧‧‧ Deep Learning or Machine Learning Model
50‧‧‧ fact database
60‧‧‧ Distributed Control System (DCS)
70‧‧‧ Maintenance Information Management System (MMIS)
U‧‧‧User

The first figure shows a schematic diagram of a plant operation and maintenance expert system according to this creative embodiment.

Claims (3)

A plant operation and maintenance expert system is used to consult the operation and maintenance suggestions of at least one device in a plant, including:
A knowledge base for storing rules related to the operation and maintenance of the at least one device in the plant;
A fact database for storing data, documents and data related to the operation and maintenance of the at least one device in the plant;
An inference engine is connected to the knowledge base and fact database, reads the fact database, compares and judges with the rules in the knowledge base, and infers operation or maintenance recommendations;
A deep learning or mechanical learning model is linked to the fact database to use the deep learning or mechanical learning data, files and data related to the operation and maintenance of the at least one device in the factory stored in the fact database. Methods, conversions are established into relevant rules, and stored in the knowledge base; and a user interface is used to input operation or maintenance related issues, transmitted to the inference engine, and used to present operations or inferences derived from the inference engine. Maintenance recommendations. According to the plant operation and maintenance expert system described in item 1 of the scope of patent application, it can receive real-time or non-immediate operation data of the at least one device in the plant of the distributed control system (DCS), and communicate through It can automatically send input to the user interface through agreement, and infer operation or maintenance suggestions through the inference engine. It can also return corresponding instructions to the distributed control system to control at least one device in the plant. According to the plant operation and maintenance expert system described in item 2 of the scope of the patent application, the source of the operation and maintenance data, documents and data in the fact database includes an operation and maintenance manual, maintenance information management system (Maintenace Management Information System, MMIS) ) Or the decentralized management system.