KR20190135923A - Facility management method and apparatus performing the same - Google Patents

Abstract

The present invention relates to a facility management method and an apparatus executing the same. The facility management method, which is executed in the facility management apparatus, comprises the steps of: analyzing each failure report to generate facility status information; generating a knowledge list by listing corresponding facility situation information according to whether facility status information of failure reports match with each other; updating a previously generated knowledge map by using the knowledge list; and performing failure diagnosis and prediction on a facility by using the previously generated knowledge map when receiving an alarm signal of facility status from a sensor installed in the facility. Therefore, the present invention has advantages of detecting abnormalities of facilities in large industrial facilities such as steel mills and automatically providing knowledge for diagnosing facilities and predicting failure.

Description

FACILITY MANAGEMENT METHOD AND APPARATUS PERFORMING THE SAME}

The present invention relates to a facility management method and a device for implementing the same, and more particularly, a facility management method for automatically detecting a failure of a facility in a large industrial facility such as a steel mill and automatically providing knowledge for predicting the diagnosis and failure of the facility. It relates to a device for doing this.

Defects or damages to mechanical systems in large industrial installations can lead to shutdowns of facilities, which can lead to huge economic losses due to inoperability or industrial accidents, which can lead to human and social damage.

Therefore, in order to maintain normal operation and to maximize the economic profits, the company operates safety management processes to prevent damages caused by failures, and is making a lot of efforts for preventive maintenance and failure prediction.

Preventive maintenance is the most important process to prevent failures in order to maintain the best condition of equipment at all times. The limitation of the preventive maintenance process is that it cannot prevent the failures that occur continually even after performing periodic maintenance, because the facilities where periodic maintenance is effective are limited.

Unnecessary excessive maintenance that is performed with or without facility abnormalities causes a problem of inefficiency and economic loss of work. Trouble shooting is to minimize the damage by quickly recovering the plant's operation rate when appropriate.

However, troubleshooting is entirely dependent on manpower and fault diagnosis and action are taken by the experience of the facility specialist. The problem is that the number of skilled equipment specialists is limited, and the absence of equipment specialists increases the risk of failure diagnosis.

In addition, even in the case of experts, unsystematic fault diagnosis is carried out by experience, so there is always a possibility of human error, and there is always a possibility of occurrence of a diagnosis error for an inexperienced situation.

The alarm system is used to overcome the limitations of facility management by personnel by collecting the status of the facility in real time from the sensors installed in the facility and providing it to the operator of the control room.

However, changes in the initial alarm system design due to continuous equipment operation, replacement of equipment, deterioration, and the addition of sensors cause changes in alarm generation patterns and numerous alarms.

This phenomenon is called alarm flooding, and it provides work fatigue by providing information that exceeds the limits that facility monitoring personnel can handle, thus relying on the experience of the personnel rather than using alarms in actual work. Identify and process

In order to solve this problem, various studies for alarm processing have been conducted. Most of the research is interested in solving problems with fluding. The purpose is to reduce the number of alarms to a level that can be handled by monitoring personnel by classifying alarm types using mathematical statistical models, analyzing alarm patterns, and extracting specific alarms.

The limitation of this study is that even if an alarm is provided at a manageable level, expert judgment such as alarm analysis, cause diagnosis, and failure prediction is required, and this process requires a lot of time and effort.

Fault knowledge is difficult to generalize to expert experience, and knowledge must be found from vastly accumulated data. Because of these problems, even if a facility abnormality is recognized, a failure cannot be avoided because the measures for the problems that occur eventually become delayed.

It is an object of the present invention to provide a facility management method and apparatus for performing the same, which automatically discovers an abnormality of a facility in a large industrial facility such as a steel mill and automatically provides knowledge for diagnosing and predicting a failure of the facility.

In addition, an object of the present invention is to provide a facility management method and an apparatus for performing the same, by making it possible to reuse the accumulated failure experience by defining the failure experience knowledge of the expert, which is difficult to generalize.

In addition, the present invention analyzes and processes a failure report written in natural language to change the structured knowledge to build a knowledge map to diagnose the situation where the failure occurred and to predict problems that can occur after a specific phenomenon And an apparatus for executing the same.

In addition, the present invention, when an alarm indicating the status of the facility received from the sensor installed in the facility is generated by mapping with the knowledge map to find out the problem predictable to the current status of the facility or to diagnose the process caused the facility status It is an object of the present invention to provide a facility management method and a device for implementing the same.

The objects of the present invention are not limited to the above-mentioned objects, and other objects and advantages of the present invention, which are not mentioned above, can be understood by the following description, and more clearly by the embodiments of the present invention. Also, it will be readily appreciated that the objects and advantages of the present invention may be realized by the means and combinations thereof indicated in the claims.

It is an object of the present invention to provide a facility management method and apparatus for performing the same, which automatically discovers an abnormality of a facility in a large industrial facility such as a steel mill and automatically provides knowledge for diagnosing and predicting a failure of the facility.

In addition, an object of the present invention is to provide a facility management method and an apparatus for performing the same, by making it possible to reuse the accumulated failure experience by defining the failure experience knowledge of the expert, which is difficult to generalize.

In addition, the present invention analyzes and processes a failure report written in natural language to change the structured knowledge to build a knowledge map to diagnose the situation where the failure occurred and to predict problems that can occur after a specific phenomenon And an apparatus for executing the same.

In addition, the present invention, when an alarm indicating the status of the facility received from the sensor installed in the facility is generated by mapping with the knowledge map to find out the problem predictable to the current status of the facility or to diagnose the process caused the facility status It is an object of the present invention to provide a facility management method and a device for implementing the same.

The objects of the present invention are not limited to the above-mentioned objects, and other objects and advantages of the present invention, which are not mentioned above, can be understood by the following description, and more clearly by the embodiments of the present invention. Also, it will be readily appreciated that the objects and advantages of the present invention may be realized by the means and combinations thereof indicated in the claims.

According to the present invention as described above, there is an advantage that it can automatically provide knowledge for detecting abnormalities of facilities and predicting the diagnosis and failure of facilities in large industrial facilities such as steel mills.

In addition, according to the present invention, it is possible to reuse the accumulated failure experience by defining the failure experience knowledge of an expert, which is difficult to generalize, into a failure rule.

In addition, according to the present invention, by analyzing and processing the failure report written in natural language, it is changed into a structured knowledge to build a knowledge map to diagnose the situation where the failure occurs and to predict a problem that may occur after a specific phenomenon. .

In addition, according to the present invention, when an alarm indicating a state of a facility received from a sensor installed in the facility occurs, it can be mapped with a knowledge map to find out a problem that can be predicted to the state of the current facility or to diagnose a process in which the state of the facility is induced. There is an advantage.

1 is a block diagram illustrating a facility management apparatus according to an embodiment of the present invention.
2 is a flowchart illustrating an embodiment of a facility management method according to the present invention.
3 is a flowchart for explaining another embodiment of the facility management method according to the present invention.
4 is an exemplary view for explaining a process of generating a knowledge map according to an embodiment of the present invention.

The above objects, features, and advantages will be described in detail with reference to the accompanying drawings, whereby those skilled in the art may easily implement the technical idea of the present invention. In describing the present invention, when it is determined that the detailed description of the known technology related to the present invention may unnecessarily obscure the gist of the present invention, the detailed description will be omitted. Hereinafter, exemplary embodiments of the present invention will be described in detail with reference to the accompanying drawings. In the drawings, the same reference numerals are used to indicate the same or similar components.

The term “failure report” used in this specification is a document describing the cause analysis of the failure and the action taken on the failure in the event of a failure. It includes know-how.

1 is a block diagram illustrating a facility management apparatus according to an embodiment of the present invention.

Referring to FIG. 1, the facility management apparatus 100 includes a facility management execution unit 110, a failure rule database 120, a failure report analysis unit 130, a knowledge list generation unit 140, and a knowledge map management unit 150. It includes.

When the facility management execution unit 110 receives an alarm signal for a facility status from a sensor installed in the facility, the facility management execution unit 110 performs a failure prediction and diagnosis of a facility corresponding to the alarm signal based on the knowledge map generated by the knowledge map manager 150. Run

First, when the facility management execution unit 110 receives an alarm signal for a facility state from a sensor installed in the facility, the failure rule database 120 determines a facility situation condition corresponding to the alarm signal.

Thereafter, the facility management execution unit 110 extracts the facility status information corresponding to the facility status condition, and then the facility status information matching the facility status information among the nodes of the knowledge map generated by the knowledge map manager 150 is displayed. Check whether the saved node exists.

Then, the facility management execution unit 110 checks whether there is a node in which the facility status information matching the facility status information is stored among the nodes of the knowledge map generated by the knowledge map manager 150. .

If there is a node in which the facility status information matching the facility status information is stored among the nodes of the knowledge map, the facility management execution unit 110 uses the node and other nodes connected to the node to troubleshoot the facility. Predict and diagnose.

In one embodiment, the facility management execution unit 110 uses the facility status information stored in each node of the node of the knowledge map that matches the facility status information that matches the facility status information and the parent node of the node, the alarm The failure of the equipment corresponding to the signal can be diagnosed.

In another embodiment, the facility management execution unit 110 uses the facility status information stored in each node of the node of the knowledge map, the facility status information matching the facility status information, and each of the child nodes of the node. The failure of the equipment corresponding to the alarm signal can be predicted.

In this case, when the child node is used as the parent node, the facility management execution unit 110 may predict the failure of the facility corresponding to the alarm signal by using the facility status information stored in each of the child nodes connected to the parent node.

As described above, the facility management execution unit 110 detects a failure of the facility corresponding to the alarm signal by using the information stored in each node in which the facility status information matching the facility status information among the nodes of the knowledge map is stored. The reason for the prediction and diagnosis is that each node of the knowledge map is connected to each other based on the similarity of the facility status information stored in the node.

The failure rule database 120 stores a failure rule generated based on the expert's experience knowledge.

The failure rule stored in the failure rule database 120 includes facility status conditions and facility status information, and facility status information includes failure-facing facilities and failure status information.

Therefore, the facility management execution unit 110 determines the facility status condition corresponding to the alarm signal from the failure rule database 120, extracts the facility status information corresponding to the facility status condition, and then generates the information by the knowledge map manager 150. Among the nodes of the knowledge map, it is possible to check whether there is a node in which the facility status information matching the facility status information is stored.

Hereinafter, when the facility management execution unit 110 receives an alarm signal for a facility status from a sensor installed in the facility, a process of generating a knowledge map that is the basis for executing a failure prediction and diagnosis of a facility corresponding to the alarm signal. Will be described.

The failure report analysis unit 130 receives the failure report, analyzes the failure report, generates facility status information, and provides the knowledge list generation unit 140. In this case, the facility status information includes a failure occurrence facility and failure status information (that is, the status of the facility and the operation of the facility).

In one embodiment, the failure report analysis unit 130 analyzes the failure report in sentence units, extracts failure occurrence equipment and failure situation information from each sentence, and provides equipment situation information including failure occurrence equipment and failure situation information. Create

The above-mentioned failure report is a document describing the cause analysis of the failure and the action taken on the failure in case of failure, and includes the expert's know-how describing the causal relationship and the steps of the failure.

Since a failure report is conventionally documented, it is difficult to obtain desired information because there is no systematic management for reuse. However, by using the contents of the failure report having a causal relationship of the failure, it is possible to diagnose the cause of the failure and to predict a problem that may occur after a specific phenomenon.

As described above, the failure report analysis unit 130 analyzes the failure report to generate facility status information and then provides the failure status information to the knowledge list generation unit 140.

The knowledge list generating unit 140 extracts the failure occurrence facility and the failure situation information from the facility situation information received from the failure report analyzer 130, and generates a plurality of nodes including the failure occurrence facility and the failure situation information. Then, the knowledge list generating unit 140 generates a knowledge list by listing each node where the failure occurrence facilities and the failure situation information match for each of the plurality of nodes, and provides the knowledge list to the knowledge map manager 150.

When the knowledge map manager 150 receives the knowledge list from the knowledge list generator 140, the knowledge map manager 150 updates the previously generated knowledge map using the knowledge list.

In one embodiment, the knowledge map manager 150 compares the information stored in each of the nodes of the knowledge map and the nodes of the knowledge list generated in advance to calculate the similarity between the nodes. Then, the knowledge map manager 150 updates the previously generated knowledge map using the knowledge list based on the similarity.

In the above embodiment, the knowledge map management unit 150 is the facility status information stored in each node of the knowledge list and the facility status information stored in each node of the pre-generated knowledge map based on Equation 1 below. Text similarity and relationship similarity between are determined, and text similarity and relationship similarity are used to calculate similarity between nodes.

In this case, the equipment status information stored in each node of the knowledge list and the equipment status information stored in each node of the previously generated knowledge map are stored in the same way, so that the information between the nodes is compared and compared. Text similarity and relationship similarity can be determined.

[Equation 1]

n _dj : j th node of graph d,

n _pi : i-th node of graph p,

Sim (n _dj , n _pi ): A function that determines the similarity between the j th node of graph d and the i th node of graph p,

SimText (n _dj , n _pi ): A function that determines the text similarity between the j th node of graph d and the i th node of graph p,

SimRel (n _dj , n _pi ): A function that determines the similarity of relationships between the j th node of graph d and the i th node of graph p

First, a process of calculating a text similarity between facility status information stored in each node of the knowledge list and facility status information stored in each node of the previously generated knowledge map will be described.

The knowledge map management unit 150 compares the facility status information stored in each node of the knowledge list and the facility status information stored in each node of the pre-generated knowledge map in text units to check whether the same word exists to determine whether the same word exists. Calculate

In one embodiment, the knowledge map manager 150 compares the facility status information stored in each node of the knowledge list and the facility status information stored in each node of the pre-generated knowledge map in text units, and the same word exists. The similarity is calculated according to the number of words.

In the above embodiment, the knowledge map management unit 150, each of the corresponding nodes when the similarity between the facility status information stored in each node of the knowledge list and the facility status information stored in each node of the pre-generated knowledge map is greater than or equal to a specific value. The relationship similarity is determined according to the number of other nodes connected with the.

In addition, the knowledge map manager 150 may refer to Equation 2 below. In one embodiment, the knowledge map manager 150 may compare the information stored in each node of the previously generated knowledge map and each node of the knowledge list. Calculate the similarity between nodes. Then, the knowledge map manager 150 updates the previously generated knowledge map using the knowledge list based on the similarity.

[Equation 2]

n _dj : j th node of graph d,

n _pi : i-th node of graph p,

Sim (n _dj , n _pi ): A function that determines the similarity between the j th node of graph d and the i th node of graph p,

EditDistance (): A function that calculates the number of times required to change the equipment status information stored in each node of the knowledge list to the equipment status information stored in each node of the previously generated knowledge map.

The knowledge map manager 150 compares the facility status information stored in each node of the knowledge list and the facility status information stored in each node of the pre-generated knowledge map to generate the facility status information stored in each node of the knowledge list in advance. The number of times required to change the facility status information stored in each node of the acquired knowledge map is calculated, and the similarity is calculated according to the number of times.

In an embodiment, the knowledge map manager 150 may determine that the number of times required to change the facility status information stored in each node of the knowledge list to the facility status information stored in each node of the previously generated knowledge map is equal to or less than a specific number of times. It is determined that the facility status information stored in each node is similar to the facility status information stored in each node of the previously generated knowledge map.

2 is a flowchart illustrating an embodiment of a facility management method according to the present invention. The embodiment disclosed in FIG. 2 relates to an embodiment in which a knowledge map can be generated using a failure report.

Referring to FIG. 2, the facility management apparatus 100 analyzes a failure report and generates facility status information, respectively (step S210).

In an embodiment of step S210, the facility management apparatus 100 analyzes the failure report in sentence units, extracts the failure occurrence equipment and failure situation information from each sentence, and includes the facilities including the failure occurrence equipment and failure situation information. Generate status information.

The facility management apparatus 100 generates a knowledge list by listing the corresponding facility status information according to whether the facility status information is mutually matched (step S220).

In one embodiment for step S220, the facility management apparatus 100 generates a plurality of nodes including the failure occurrence facility and failure situation information, and each of the failure occurrence facility and failure situation information is matched for each of the plurality of nodes. Create a knowledge list by listing the nodes of.

The facility management apparatus 100 updates the previously generated knowledge map using the knowledge list (step S230).

In one embodiment of step S230, the facility management apparatus 100 compares the information stored in each of the nodes of the knowledge map and the nodes of the knowledge list generated in advance to calculate the similarity between the nodes, and based on the similarity The previously generated knowledge map is updated by using a list.

To this end, the facility management apparatus 100 compares the facility status information stored in each node of the knowledge list and the facility status information stored in each node of the pre-generated knowledge map in text units, and if the same word exists, The similarity may be calculated according to the number of words.

If the similarity between the specific node of the knowledge list and the specific node of the pre-generated knowledge map is greater than or equal to a certain value, the facility management apparatus 100 displays the knowledge list in the form of a graph based on the specific node of the pre-generated knowledge map. It is configured to update the pre-generated knowledge map.

3 is a flowchart for explaining another embodiment of the facility management method according to the present invention. The embodiment disclosed in FIG. 3 relates to an embodiment in which failure prediction and diagnosis of a facility may be performed using a knowledge map using a failure report.

Referring to FIG. 3, the facility management apparatus 100 receives an alarm signal for a facility state from a sensor installed in a facility (step S310).

The facility management apparatus 100 extracts facility status information corresponding to the alarm signal from the failure rule database (step S320).

The facility management apparatus 100 checks whether there is a node in which the facility status information matching the facility status information exists among the respective nodes of the previously generated knowledge map (step S330).

The facility management apparatus 100 predicts and diagnoses a failure of the facility by using another node connected with the node (step S340).

In one embodiment of step S340, the facility management execution unit 110 is the facility status information stored in each of the node and the parent node of the node is stored the facility status information matching the facility status information of each node of the knowledge map You can diagnose the failure of the equipment corresponding to the alarm signal by using.

In another embodiment of step S340, the facility management execution unit 110 is the facility status stored in each node and the child node of the node is stored the facility status information matching the facility status information of each node of the knowledge map The information can be used to predict the failure of the equipment corresponding to the alarm signal.

In this case, when the child node is used as the parent node, the facility management execution unit 110 may predict the failure of the facility corresponding to the alarm signal by using the facility status information stored in each of the child nodes connected to the parent node.

4 is an exemplary view for explaining a process of generating a knowledge map according to an embodiment of the present invention.

Referring to FIG. 4, the facility management apparatus 100 calculates a similarity between facility A and facility B. FIG. Facility A is represented by graph P and facility B is represented by graph d. The graph of facility A (G _p ) includes node (n _p ) and the edge (e _p ) connecting the nodes, and the graph of facility B (G _d ) includes node (n _d ) and the edge (e) connecting nodes. _d ).

The similarity between the i th node (n _pi ) of the graph (G _p ) of the facility A and the j th node (n _dj ) of the graph (G _d ) of the facility B is expressed by Equations 1 and 2 above. Similarly, the text similarity of the i th node (n _pi ) of facility A's graph (G _p ) and the j th node (n _dj ) of facility B's graph (G _d ) and the i'th node of facility A's graph (G _p ) (n _pi ) and the similarity of the j th node (n _dj ) of the graph G _d of Facility B.

The present invention as described above, although been described and specific examples, the invention is not limited to the embodiments described above, which those skilled in the art to which the invention pertains many modifications to the described and Modifications are possible. Accordingly, the invention idea should be understood only by the claims set forth below, and all equivalent or equivalent modifications thereof will fall within the scope of the invention idea.

100: equipment management device
110: facility management execution unit
120: failure rule database
130: failure report analysis unit
140: knowledge list generation unit
150: knowledge map management

Claims (14)

In the facility management method performed in the facility management device,
Analyzing the failure report and generating facility status information, respectively;
Generating a knowledge list by listing corresponding facility situation information according to whether each facility situation information is matched with each other;
Updating a previously generated knowledge map using the knowledge list; And
Receiving an alarm signal for the status of the facility from a sensor installed in the facility, comprising the step of performing a diagnosis and prediction of the failure of the facility using the pre-generated knowledge map
How to manage your equipment.
The method of claim 1,
Generating facility status information by analyzing the failure report
Analyzing the failure report in units of sentences and extracting information on failure occurrence facilities and failure conditions from each sentence; And
Generating facility status information including the failure-producing facility and failure status information.
How to manage your equipment.
The method of claim 1,
Generating a knowledge list by listing the facility status information
Generating a plurality of nodes including the failure occurrence facility and failure situation information; And
And generating a knowledge list by listing each node where the failure occurrence facility and the failure situation information match for each of the plurality of nodes.
How to manage your equipment.
The method of claim 1,
The updating of the previously generated knowledge map using the knowledge list
Calculating similarity between nodes by comparing information stored in each node of the previously generated knowledge map and each node of the knowledge list; And
And updating the pre-generated knowledge map using the knowledge list based on the similarity.
How to manage your equipment.
The method of claim 4, wherein
Comparing the information stored in each node of the pre-generated knowledge map and the nodes of the knowledge list to calculate the similarity between nodes
Comparing facility status information stored in each node of the knowledge list and facility status information stored in each node of the pre-generated knowledge map in text units, when the same word exists, the similarity is calculated according to the number of words. Characterized in that it comprises a step of
How to manage your equipment.
The method of claim 4, wherein
Updating the pre-generated knowledge map using the knowledge list based on the similarity level
When the similarity between the specific node of the knowledge list and the specific node of the pre-generated knowledge map is equal to or greater than a specific value, the pre-generated list of the knowledge list in the form of a graph based on the specific node of the pre-generated knowledge map. And updating the knowledge map.
How to manage your equipment.
The method of claim 1,
Receiving an alarm signal for a facility status from a sensor installed in the facility;
Extracting facility status information corresponding to the alarm signal from a failure rule database;
Checking whether there is a node in which facility status information matching the facility status information exists among each node of the knowledge map; And
Predicting and diagnosing a failure of a facility by using the node and another node connected to the node according to the checking result;
How to manage your equipment.
A failure report analysis unit for analyzing the failure report and generating facility status information, respectively;
A knowledge list generation unit for generating a knowledge list by listing corresponding facility situation information according to mutual matching between the facility situation information;
A knowledge map manager for updating a knowledge map previously generated using the knowledge list; And
Receiving an alarm signal for the status of the facility from the sensor installed in the facility, characterized in that it comprises a facility management execution unit for performing a failure diagnosis and prediction of the facility using the pre-generated knowledge map
Facility management device.
The method of claim 8,
The failure report analysis unit
After analyzing the failure report in sentence units, it is possible to extract the failure occurrence equipment and failure situation information from each sentence, and generate facility situation information including the failure occurrence facility and failure situation information.
Facility management device.
The method of claim 8,
The knowledge list generation unit
Generate a plurality of nodes including the failure occurrence facility and failure situation information, and generate a knowledge list by listing each node where the failure occurrence facility and failure situation information is matched for each of the plurality of nodes; doing
Facility management device.
The method of claim 8,
The knowledge map management unit
Comparing the information stored in each node of the pre-generated knowledge map and the nodes of the knowledge list to calculate the similarity between the nodes, and updating the pre-generated knowledge map using the knowledge list based on the similarity. By
Facility management device.
The method of claim 11,
The knowledge map management unit
Comparing facility status information stored in each node of the knowledge list and facility status information stored in each node of the pre-generated knowledge map in text units, when the same word exists, the similarity is calculated according to the number of words. Characterized by
Facility management device.
The method of claim 11,
The knowledge map management unit
When the similarity between the specific node of the knowledge list and the specific node of the pre-generated knowledge map is equal to or greater than a specific value, the pre-generated list of the knowledge list in the form of a graph based on the specific node of the pre-generated knowledge map. Characterized by updating the knowledge map
Facility management device.
The method of claim 8,
The facility management execution unit
Extracts facility status information corresponding to the alarm signal from a failure rule database, checks whether a node in which facility status information matching the facility status information is stored among each node of the knowledge map exists, and checks According to the result, the failure of the facility is predicted and diagnosed using the node and other nodes connected to the node.
Facility management device.

Images