Fault classification device, fault classification procedure and fault classification program

Abstract

Fault classification device (1) for detecting a location that is the cause of a fault in a mechanism unit of a machine with multiple components, comprising: a fault unit detection unit (12) that captures a data set in which a fault unit of a machine is correlated with one or more components, wherein the fault unit is a group of one or more components whose causal fault locations cannot be separated from each other when designing the machine; a fault history recording unit (11) that records a fault history containing an event of faults that have occurred in the past and countermeasures for eliminating the faults; and a correlation unit (13) which stores information about one or more events in correlation with the fault unit in a fault history database (21) by comparing the component with the countermeasure component; a recovery receiving unit (15) that receives a recovery request containing a search keyword; and a recovery execution unit (16) that recovers information about the event based on the search keyword and outputs the fault unit correlated with the information about the event in order to specify the location in the mechanism unit of the machine that caused the fault.

Description

BACKGROUND OF THE INVENTION Area of ​​the invention

[0001] The present invention relates to a fault classification device, a fault classification method and a fault classification program for specifying locations of faults in a machine. Related area

[0002] When a fault occurs in a machine such as a machine tool, robot, or injection molding machine, a controller (e.g., a computer numerical controller (CNC)) that controls the machine conventionally detects faults within the controller itself or problems resulting from an overload or similar issue with a motor in a mechanical unit of the machine, and issues an alarm. This alarm information is useful for troubleshooting and plays a role in grouping the many causes of faults that occur.

[0003] To correct a currently occurring error, information from a previously stored error report can be retrieved using keyword-based retrieval. For example, patent document 1 discloses as a keyword-based retrieval method a method for displaying candidates for refined keywords in descending order of cumulative numbers of keywords present in a group of documents, as a guide to direct the retrieval process in a desired direction.

[0004] Patent document 1 (JP H05 - 314 182 A): Japanese unexamined patent application

[0005] US 2008 / 0034258A1 concerns a fault management device that enables the calculation of evaluation values ​​in relation to a fault occurring at an administrative objective as objective values ​​without the involvement of human judgment. SUMMARY OF THE INVENTION

[0006] However, a controller may not detect some machine faults and therefore may not generate an alarm. In such cases, there are multiple potential sources of the fault, and locating the source can be a significant undertaking. Even when an event is recovered from a previous fault report, pinpointing the exact location of a fault across multiple sections of a machine's mechanical unit remains challenging.

[0007] An object of the present invention is to provide a fault classification device, a fault classification method, and a fault classification program that can specify the causes of faults even when a controller does not generate an alarm. This object is achieved by a fault classification device having the features of claim 1, a fault classification method having the features of claim 4, and a fault classification program having the features of claim 5.

[0008] (1) A fault classification device (e.g., a fault classification device 1 to be described later) according to the present invention comprises: a fault unit acquisition unit (e.g., a fault unit acquisition unit 12 to be described later) that acquires a data record in which a fault unit of a machine is correlated with one or more components; a fault history acquisition unit (e.g., a fault history acquisition unit 11 to be described later) that acquires a fault history containing events of faults that have occurred in the past and countermeasure components; and a correlation unit (e.g., a correlation unit 13 to be described later) that stores information about one or more events correlated with the fault unit by matching the component with the countermeasure component.

[0009] (2) Furthermore, the fault classification device according to (1) may include a feature extraction unit (e.g. a feature extraction unit 14 to be described later) which extracts a fault keyword contained in the information about the event as feature data of each of the fault units correlated by the correlation unit, and the correlation unit may store one or more parts of the feature data correlated with the fault unit.

[0010] (3) Furthermore, the fault classification device according to (1) or (2) may include: a recovery receiving unit (e.g. a recovery receiving unit 15 to be described later) which receives a recovery request containing a search keyword; and a recovery execution unit (e.g. a recovery execution unit 16 to be described later) which recovers information about the event based on the search keyword and outputs the fault unit correlated with the information about the event.

[0011] (4) In the fault classification device according to (3) the recovery receiving unit can extract feature data contained in the recovery request as the search keyword.

[0012] (5) A fault classification method according to the present invention causes a computer (e.g., a fault classification device 1 to be described later) to perform: a fault unit acquisition step to acquire a data record in which a fault unit of a machine is correlated with one or more components; a fault history acquisition step of acquiring a fault history that includes events of faults that have occurred in the past and countermeasure components; and a correlation step of storing information about one or more events in correlation with the fault unit by matching the component with the countermeasure component.

[0013] (6) A fault classification program according to the present invention causes a computer (e.g., a fault classification device 1 to be described later) to perform: a fault unit acquisition step to acquire a data record in which a fault unit of a machine is correlated with one or more components; a fault history acquisition step to acquire a fault history containing events of faults that have occurred in the past and countermeasure components; and a correlation step to store information about one or more events in correlation with the fault unit by matching the component with the countermeasure component.

[0014] According to the present invention, it is possible to specify the causes of errors even when a controller does not generate an alarm. BRIEF DESCRIPTION OF THE DRAWINGS Fig. Figure 1 is a block diagram representing a functional configuration of a fault classification device according to one embodiment. Fig. Figure 2 is a representation that depicts a configuration example of a machine that is the subject of a fault classification procedure according to an embodiment. Fig. Figure 3 is a representation that shows an example of error history data according to one embodiment. Fig. Figure 4 is a representation that provides a definition example of a defect unit according to one embodiment. Fig. Figure 5 is a representation that shows an example of a data set that defines a unit of error according to one embodiment. Fig. Figure 6 is a representation that shows an example of an error history database to which an event code is appended according to an embodiment. Fig. Figure 7 presents an example in which an error history database is sorted according to an event code, according to one embodiment. Fig. Figure 8 is a representation that presents an example of an error history database based on feature data according to one embodiment. Fig. Figure 9 is a flowchart that presents a recovery process example of a fault classification device according to one embodiment. DETAILED DESCRIPTION OF THE INVENTION

[0015] The following describes an example of an embodiment of the present invention. Fig. Figure 1 is a block diagram representing a functional configuration of a fault classification device 1 according to the present embodiment.

[0016] The error classification device 1 is an information processing device (a computer) such as a server device or a PC and contains a control unit 10 and a storage unit 20. Furthermore, the error classification device 1 contains various data input / output devices, a communication device, and the like, and searches a database of the storage unit 20 based on the input of search keywords by a user in order to output a recovery result.

[0017] The control unit 10 is a component that controls the entire fault classification device 1 and implements various functions of the present embodiment by appropriately reading and executing different programs stored in the memory unit 20. The control unit 10 can be a CPU.

[0018] The memory unit 20 is a memory area for various programs to cause a hardware group to function as the fault classification device, and for various data, and can be a ROM, a RAM, a flash memory, or a hard disk drive (HDD). More specifically, in addition to a fault classification program to cause the control unit 10 to perform various functions of the present embodiment, the memory unit 20 stores a fault history database 21 and the like, which is designed as a recovery target. The fault history database 21 is located outside the fault classification device 1 and can read and write data by communicating with the fault classification device 1.

[0019] The control unit 10 contains a fault history acquisition unit 11, a fault unit acquisition unit 12, a correlation unit 13, a feature extraction unit 14, a recovery receiving unit 15, and a recovery execution unit 16. Using these functional units, the control unit 10 classifies faults based on event information from a machine.

[0020] Fig. Figure 2 is a diagram illustrating a configuration example of a machine that is the subject of a fault classification procedure according to the present embodiment. The machine is classified into a control unit and a mechanism unit. The control unit displays an alarm number and the content of an alarm detected by a controller (CNC) on a display device to inform an operator of a situation. Furthermore, a controller monitors the status of a motor in the mechanism unit, which is controlled by a server amplifier, for faults occurring in the mechanism unit and issues an overcurrent alarm, an overload alarm, or the like for the motor.

[0021] Although an alarm can be triggered for faults that cause serious problems in machine operations, it can often be difficult to detect faults occurring in the mechanical unit based solely on feedback from the motor, since dynamic power is indirectly transmitted from the motor to the respective sections of the mechanical unit. In this case, an operator identifies vibrations, abnormal noises, and the like as being generated in the mechanical unit and searches for the fault location by checking the respective machine components one by one.

[0022] The fault history recording unit 11 records a fault history that includes various fault events that have occurred in the machine in the past and countermeasures.

[0023] Fig. Figure 3 is a representation that presents an example of fault history data according to the present embodiment. In the fault history data, for example, a description such as "A grinding sound is generated from the vicinity of a table," which describes the event of a fault that occurred in the past, is correlated with the number of a countermeasure part for eliminating the fault. In this example, an alarm number is recorded when the controller (CNC) issues an alarm in relation to the fault that occurred.

[0024] The fault unit detection unit 12 captures a data record in which a fault unit of a machine is correlated with one or more components. A fault unit is a unit in which a fault occurring in the mechanism unit of a machine is checked and countermeasures are taken against it, whereby, when a machine is designed, several fault units are included, each consisting of one or more components whose causes of a fault cannot be separated.

[0025] Fig. Figure 4 is a representation that provides a definition example of a defect unit according to the present embodiment. For example, a mechanism unit of a machine tool is divided into the following defect units. - M001: Location contains a motor that drives a machine tool, a pulse encoder that provides feedback on the rotational position of the motor, and the like. - C001: Location contains a joint or the like, which is a connecting section for transmitting dynamic power from a motor to a ball screw or the like. - B001: The location contains, in addition to a ball screw, a ball nut and a bearing driven by a motor via a joint to align a table, lubrication points to which grease is supplied automatically or manually, or the like. - T001: The location contains, in addition to a table and a linear motion guide (LM guide), lubrication points to which grease is supplied automatically or manually, or the like.

[0026] Fig. Figure 5 is a representation that presents an example of a data record defining a defect unit according to the present embodiment. In a data record, a part number and a part name, which constitute a defect unit, are registered with respect to an event code, which is an identifier specifying the defect unit. In this example, for each event code, an event message to notify about the causes of a defect, a countermeasure procedure, and the like are registered.

[0027] The correlation unit 13 matches a component that is correlated with an event code with a countermeasure component of an error history and stores information about one or more events in correlation with the event code (of the error unit) in the error history database 21 of the storage unit 20.

[0028] Fig. Figure 6 is a representation that shows an example of an error history database 21 to which an event code is appended according to the present embodiment. In this example, each in Fig. The error history data shown in point 3 is appended with an event code. Using this error history database 21, an event code can be specified based on an event or an alarm indicating an error, and the error unit is specified.

[0029] The format of the error history database 21 is not limited to this, and changes can be made to achieve efficient data recovery. The correlation unit 13 can store one or more characteristic data points in the error history database 21, indicating an error event, in correlation with an event code.

[0030] Feature extraction unit 14 extracts an error keyword contained in event information as feature data for each event code correlated by correlation unit 13. The error keyword extraction method is not limited here, and, for example, a method proposed in Japanese patent application No. 2017-159990, which uses morpheme analysis, can be employed. In this case, error keywords can be determined by referencing various databases of unnecessary words excluded from extraction targets, as well as synonyms (from a thesaurus) classified as having the same meaning.

[0031] Fig. Figure 7 presents an example in which the error history database 21 is sorted according to an event code, as described in the present embodiment. In this way, several event details and countermeasure components are correlated with the same event code.

[0032] Fig. Figure 8 is a representation that illustrates an example of the fault history database 21 based on feature data according to the present embodiment. The feature extraction unit 14 extracts a fault keyword from a description of an event correlated with an event code, and the fault keyword is stored as feature data specifying the event in the fault history database 21. The number assigned to a fault keyword and a countermeasure in the drawing indicates the number of occurrences in a fault history with the same event code.

[0033] The recovery receiving unit 15 receives a recovery request for the fault history database 21, which contains a search keyword. The recovery request can be executed by entering one or more keywords, and a sentence describing an event observed by an operator can be entered. In this case, the recovery receiving unit 15 extracts feature data contained in a description as a search keyword. If the same procedure is used here for both the search keyword extraction and the fault keyword extraction, both keywords are aligned, and the recovery accuracy is improved.

[0034] The recovery execution unit 16 recovers the event information from the fault history database 21 based on the search keyword and outputs an event code (a fault unit) correlated with the event information. The recovery execution unit 16 can retrieve an event message corresponding to the event code from the database. Fig. Output the definition of the error unit shown in section 5. During recovery, a weighting can be applied to the keyword matching, corresponding to the number of occurrences of the error keyword.

[0035] Fig. Figure 9 is a flowchart illustrating a recovery process example of the fault classification device 1 according to the present embodiment. In step S1, the recovery receiving unit 15 receives the input of a sentence describing an event that occurred in a machine, as a recovery request.

[0036] In step S2, the recovery receiving unit 15 performs a morpheme analysis on the input sentence as the recovery request and extracts feature data as a search keyword.

[0037] In step S3, the recovery execution unit 16 compares the search keyword with the error keyword in the error history database 21 to determine whether an error history similar to the recovery request exists. If the result is "Yes," the process proceeds to step S4; if the result is "No," the process terminates.

[0038] In step S4, the recovery execution unit 16 extracts an event code that is assigned to the recovered similar fault history.

[0039] In step S5, the recovery execution unit 16 outputs the extracted event code and an event message corresponding to the event code.

[0040] According to the present embodiment, the fault classification device 1 acquires a data set in which a fault unit and a component are correlated, and a fault history in which a fault event and a countermeasure component are correlated. It then compares the component with the countermeasure to correlate one or more event pieces of information with the fault unit. Because the fault classification device 1 can group fault events into fault units in this way, it is possible to output fault units based on an event and to specify the location of a mechanism unit that is the cause of the fault, even if a machine controller does not generate an alarm. As a result, it is easy to specify fault locations, and the time required for fault tracking is reduced.

[0041] Since the feature classification device 1 extracts a fault keyword that characterizes the event information of faults for respective fault units, the correlation between an event and a fault unit becomes clear and it is possible to specify the causes of faults more reliably.

[0042] Since the fault classification device 1 receives a recovery request containing a search keyword, it is also possible to search the fault history database 21 to specify a fault unit and output an associated message, such as a countermeasure procedure. Because the fault classification device 1 extracts feature data contained in a recovery request as a search keyword, it is possible to perform an efficient recovery process based on a keyword match, thereby increasing user convenience and more reliably specifying and outputting the causes of faults.

[0043] Although embodiments of the present invention have been described above, the present invention is not limited to these embodiments. Furthermore, the advantageous effects described in the embodiments of the present invention are only examples of the most desirable effects produced by the present invention, so that the advantageous effects of the present invention are not limited to those described in the embodiments of the present invention.

[0044] The error classification procedure is implemented by software in the error classification device 1. When the error classification procedure is implemented by software, the programs constituting the software are installed on a computer (the error classification device 1). Furthermore, these programs can be recorded on a removable medium and distributed to users, and can be distributed by downloading them over a network to a user's computer. EXPLANATION OF REFERENCE MARKS 1 fault classification device 10 Control unit 11 Error history recording unit 12 error unit recording unit 13 Correlation Unit 14 Feature extraction unit 15 Recovery receiving unit 16 Recovery Execution Unit 20 storage units 21 Error history database

Claims

Fault classification device (1) for detecting a location that is the cause of a fault in a mechanism unit of a multi-component machine, comprising: a fault unit detection unit (12) that captures a data record in which a fault unit of a machine is correlated with one or more components, wherein the fault unit is a group of one or more components whose causal fault locations cannot be separated from one another during the design of the machine; a fault history detection unit (11) that captures a fault history containing an event of faults that have occurred in the past and countermeasure components for eliminating the faults; and a correlation unit (13) that stores information about one or more events correlated with the fault unit in a fault history database (21) by matching the component with the countermeasure component;a recovery receiving unit (15) that receives a recovery request containing a search keyword; and a recovery execution unit (16) that recovers information about the event based on the search keyword and outputs the fault unit correlated with the information about the event in order to specify the location in the mechanism unit of the machine that caused the fault. Fault classification device (1) according to claim 1, further comprising: a feature extraction unit (14) which extracts a fault keyword contained in the information about the event as feature data of each of the fault units correlated by the correlation unit (13), wherein the correlation unit (13) stores one or more parts of the feature data in correlation with the fault unit. Fault classification device (1) according to claim 1, wherein the recovery receiving unit (15) extracts feature data contained in the recovery request as the search keyword. Fault classification method for detecting a location that is causally responsible for a fault in a mechanism unit of a machine with multiple components, in order to cause a computer to (1) perform: a fault unit detection step to capture a data set in which a fault unit of a machine is correlated with one or more components, wherein the fault unit is a group of one or more components whose causal fault locations cannot be separated from each other in the design of the machine; a fault history detection step to capture a fault history that includes events of faults that have occurred in the past and countermeasure parts to eliminate the fault;and a correlation step of storing information about one or more events in correlation with the fault unit by matching the component with the countermeasure component in a fault history database (21); a recovery receive step of receiving a recovery request containing a search keyword; and a recovery execution step of recovering the information about the event based on the search keyword and outputting the fault unit correlated with the information about the event in order to specify the location in the mechanism unit of the machine that caused the fault.; A fault classification program for detecting a location that is the cause of a fault in a mechanism unit of a multi-component machine, causing a computer to (1) perform: a fault unit capture step to capture a data record in which a fault unit of a machine is correlated with one or more components, wherein the fault unit is a group of one or more components whose causal fault locations cannot be separated from each other in the design of the machine; a fault history capture step to capture a fault history that includes events of faults that have occurred in the past and countermeasure parts to eliminate the fault;and a correlation step of storing information about one or more events in correlation with the fault unit by matching the component with the countermeasure component in a fault history database (21); a recovery receive step of receiving a recovery request containing a search keyword; and a recovery execution step of recovering the information about the event based on the search keyword and outputting the fault unit correlated with the information about the event in order to specify the location in the mechanism unit of the machine that caused the fault.;

Images