JP7467982B2 - Abnormality determination device and abnormality determination method - Google Patents

Description

The present invention relates to an abnormality determination device and an abnormality determination method.

Patent Document 1 discloses a technology for monitoring and diagnosing a plant, taking into account the history of the plant's manufacturing, inspection, repair, operation, etc.

Japanese Patent Application Laid-Open No. 6-331507

However, according to the technology described in Patent Document 1, during plant maintenance and inspection, data on installed equipment and components is entered into a terminal by an operator, which can lead to cases where data is entered incorrectly. If data is entered incorrectly, there is a problem in that an accurate diagnosis of anomalies cannot be performed during plant operation. For example, even if an anomaly is determined to be occurring in the plant, it is not possible to distinguish whether the anomaly is due to erroneous data entry or an anomaly in the plant itself.

The present invention was made in consideration of the above problems, and its purpose is to provide an abnormality determination device and an abnormality determination method that can distinguish between erroneous data input and determine whether a device is normal or abnormal, even if erroneous data is input during maintenance and inspection of the device.

In order to solve the above-mentioned problems, an abnormality determination device and an abnormality determination method according to one aspect of the present invention obtain a result of determining whether or not a device is in a predetermined state based on device status data as a first determination result, and when the device is in the predetermined state, obtain maintenance history data of the device via an operation data input unit, obtain a result of determining whether or not there is an abnormality in the device based on the maintenance history data and the status data as a second determination result, and output the presence or absence of an abnormality in the device based on the first determination result and the second determination result.

According to the present invention, even if data is entered incorrectly during maintenance and inspection of the device, it is possible to distinguish between incorrect data entry and determine whether the device is normal or abnormal.

FIG. 1 is a block diagram showing the configuration of an abnormality determination device according to an embodiment of the present invention and a production robot that is the object of the abnormality determination device. FIG. 2 is a flowchart showing a processing procedure of the abnormality determination device according to one embodiment of the present invention. FIG. 3 is a flowchart showing a processing procedure of the abnormality determination device according to the first modified example of the present invention. FIG. 4 is a flowchart showing a processing procedure of an abnormality determination device according to a second modified example of the present invention.

Next, an embodiment of the present invention will be described in detail with reference to the drawings. In the description, the same parts will be given the same reference numerals and duplicate explanations will be omitted.

Figure 1 is a block diagram showing the configuration of the anomaly detection device according to this embodiment and the production robot that is the subject of the detection.

The abnormality determination device 100 is connected to the production robot 80 (device) wirelessly or by wire so as to be able to communicate with it. The abnormality determination device 100 determines whether or not the work performed by the production robot 80 is abnormal, or whether or not the production robot 80 itself is abnormal. The production robot 80 is, for example, a robot that has multiple moving parts and performs welding work on a vehicle body.

[Production robot configuration]
The production robot 80 includes a communication unit 82 , a measurement unit 84 , and a sensor 86 .

The sensor 86 detects the state of the production robot 80. The sensor 86 is composed of, for example, a torque sensor or an acceleration sensor. Alternatively, the sensor 86 may be, for example, a camera that captures images or videos, or may be an optical sensor, a sound sensor, an acceleration sensor, a vibration sensor, a pressure sensor, a strain sensor, an acoustic emission sensor, a temperature sensor, or a humidity sensor. Alternatively, the sensor 86 may be a position sensor that measures the position of the moving parts of the production robot 80, or a sensor that measures the braking/driving force and torque generated by the braking/driving parts.

The measurement unit 84 generates status data for the production robot 80 from data on the status of the production robot 80 detected by the sensor 86. The status data for the production robot 80 may be the data itself obtained by the sensor 86 (so-called raw data), or may be data obtained after the data obtained by the sensor 86 has been analyzed by an arithmetic circuit (not shown). The status data for the production robot 80 may also include information on the date and time when the data was obtained by the sensor 86 (a time stamp such as date and time).

The communication unit 82 transmits the status data of the production robot 80 generated by the measurement unit 84 to the abnormality determination device 100. The communication unit 82 may accumulate the status data and transmit it to the abnormality determination device 100 at regular intervals, or may transmit the status data generated by the measurement unit 84 to the communication unit 82 in real time.

In addition, the data transmitted by the communication unit 82 to the abnormality determination device 100 may include data such as the type, load, and frequency of work performed by the production robot 80, and may also include maintenance history data (information on the content of the maintenance work and the history of the maintenance points) indicating that maintenance work has been performed when maintenance work (repairs, part replacement, changing of lubricating oil, etc.) has been performed on the production robot 80.

[Configuration of the abnormality determination device]
The abnormality determination device 100 includes an input unit 102, a database 110, a control unit 120 (controller), and an output unit 130. The control unit 120 is connected to the input unit 102, the database 110, and the output unit 130 so as to be able to communicate with them. In addition, the output unit 130 is connected to a display unit 90 and an operation unit 95 (operation data input unit) so as to be able to communicate with them. The display unit 90 and the operation unit 95 may be included in the abnormality determination device 100 itself, or may be installed outside the abnormality determination device 100.

The input unit 102 is connected wirelessly or by wire to be able to communicate with the communication unit 82 of the production robot 80. The input unit 102 receives status data of the production robot 80 from the communication unit 82 of the production robot 80. That is, data output from a sensor that detects the status of the device is input to the input unit 102. In addition, the input unit 102 may receive data such as the type, load, and frequency of work performed by the production robot 80, as well as maintenance history data.

The data received by the input unit 102 is stored in the database 110. The data stored in the database 110 is read by the control unit 120 when executing the abnormality determination process. When storing the status data of the production robot 80, the database 110 may store data such as the type of work performed by the production robot 80, the load, and frequency of the work, and may also store data of the maintenance history of the production robot 80 in association with the data.

In addition, the input unit 102 is connected to the operation unit 95 described below, and accepts operations from users, such as supervisors and maintenance personnel of the production robot 80.

The control unit 120 determines whether there is an abnormality in the production robot 80 based on the status data of the production robot 80.

The output unit 130 is connected to the display unit 90, which will be described later, and outputs information generated by the control unit 120. In particular, the output unit 130 outputs information to notify supervisors and maintenance personnel of the production robot 80.

The display unit 90 displays the information received from the output unit 130. Furthermore, if the control unit 120 determines that the production robot 80 is abnormal, the display unit 90 issues an abnormality alarm based on a command from the control unit 120 to notify a supervisor or maintenance personnel that an abnormality has been detected. For example, the display unit 90 may be a display that displays figures and characters by combining multiple display pixels, or may be a rotating light, buzzer, etc. The display unit 90 is not limited to the examples given here.

The operation unit 95 (operation data input unit) is an input device that can be operated by a user, such as a supervisor or maintenance worker of the production robot 80. For example, the operation unit 95 is a keyboard, a mouse, a trackball, a touch panel, etc. The operation unit 95 is not limited to the examples given here.

[Configuration of control unit]
The control unit 120 (an example of a processing unit or controller) is a general-purpose microcomputer equipped with a CPU (Central Processing Unit), a memory, and an input/output unit. A computer program (abnormality determination program) is installed in the control unit 120 to function as part of the abnormality determination device 100 for determining abnormalities in the production robot 80. By executing the computer program, the control unit 120 functions as multiple information processing circuits (122, 124, 126, 128) equipped in the abnormality determination device 100.

In this embodiment, an example is shown in which multiple information processing circuits (122, 124, 126, 128) are realized by software. However, it is also possible to configure the information processing circuits (122, 124, 126, 128) by preparing dedicated hardware for executing each information process shown below. Also, the multiple information processing circuits (122, 124, 126, 128) may be configured by individual hardware. Furthermore, the information processing circuits (122, 124, 126, 128) may also be used as a control unit used to monitor or control the production robot 80.

The control unit 120 includes a first determination unit 122, a condition setting unit 124, a second determination unit 126, and a decision unit 128 as multiple information processing circuits (122, 124, 126, 128).

The first judgment unit 122 and the second judgment unit 126 read the status data of the production robot 80 from the database 110 and perform the process of abnormality judgment based on the read status data. The first judgment unit 122 and the second judgment unit 126 perform the process of abnormality judgment based on different judgment algorithms.

The result obtained by the abnormality determination by the first determination unit 122 is hereinafter referred to as the "first determination result," and the result obtained by the abnormality determination by the second determination unit 126 is hereinafter referred to as the "second determination result." Because the first determination unit 122 and the second determination unit 126 perform abnormality determination processing based on different determination algorithms, the first determination result and the second determination result do not necessarily match.

To perform the abnormality determination process, the first determination unit 122 and the second determination unit 126 calculate, for example, feature quantities that represent the characteristics of each state data based on each state data of the production robot 80.

The first and second determination units 122 and 126 may calculate, as features, statistics (e.g., mean, variance, skewness, kurtosis, etc.) or various analytical values (e.g., values of specific spectra after Fourier transform, ratios between spectra, etc.) designated in advance by the user, or combinations of these. The first and second determination units 122 and 126 may also calculate, as features, quantities that are automatically set using various feature selection methods such as principal component analysis, random forest, and XGBoost. The parameters for defining the "features" calculated by the first and second determination units 122 and 126 may be stored in the database 110.

The first judgment unit 122 and the second judgment unit 126 each have a first judgment model and a second judgment model that are generated by performing machine learning before performing the anomaly judgment process. Note that the first judgment model and the second judgment model are mutually different judgment models.

The first and second judgment models are generated in advance by performing machine learning based on "feature values" calculated based on past state data of the production robot 80. The first and second judgment models may be stored in the database 110.

The first judgment model and the second judgment model are models that classify the status data of the production robot 80 into at least two groups.

The two or more groups include a "normal range" group corresponding to the normal state of the production robot 80, and an "abnormal range" group corresponding to the abnormal state of the production robot 80. The first judgment unit 122 and the second judgment unit 126 perform abnormality judgment processing by determining whether the state of the production robot 80 is classified as "normal" or "abnormal" based on the calculated feature amount using the first judgment model and the second judgment model, respectively. Then, the first judgment unit 122 and the second judgment unit 126 obtain the results of the processing as a first judgment result and a second judgment result, respectively.

The first determination unit 122 and the second determination unit 126 may perform anomaly determination processing using a method for detecting changes in the correlation between multiple signals, such as a graphical model or sparse structure learning, or may calculate the degree of anomaly using a method that uses a probability distribution, such as kernel density estimation or density ratio estimation.

The first judgment unit 122 and the second judgment unit 126 may calculate, with a probability, which group the feature should be classified into, using the first judgment model and the second judgment model, respectively. Also, they may calculate which group the state data should be classified into depending on whether the difference between the feature of the data point representing the group and the feature calculated from the state data is equal to or less than a predetermined threshold.

In addition, the first judgment model and the second judgment model may be generated in advance before a series of abnormality judgment processes are performed, or may be updated sequentially while a series of abnormality judgment processes are performed.

When the first judgment unit 122 judges that the production robot 80 is abnormal, the condition setting unit 124 requests the user to check the maintenance history data of the production robot 80. Specifically, based on an instruction from the condition setting unit 124, the display unit 90 displays a message indicating that the maintenance history data of the production robot 80 needs to be checked.

In particular, if the first judgment unit classifies the area as "abnormal" in the classification of "normal" and "abnormal" groups and there is a possibility that the area has been affected by past maintenance, a request is made to check the maintenance history. In other words, if the first judgment unit classifies the area as "abnormal" and there is a change from "normal" to "abnormal" before and after past maintenance, or there is a change from "abnormal" to another "abnormal" before and after past maintenance, a request is made to check the maintenance history.

The condition setting unit 124 may display the reference maintenance history data on the display unit 90 together with a message indicating that the maintenance history data needs to be checked. This allows comparison with the reference information, and the user can accurately check the stored maintenance history data via the display unit 90, and furthermore, correct the stored maintenance history data via the operation unit 95.

When the user modifies the stored maintenance history data via the operation unit 95, the condition setting unit 124 sets or modifies the judgment conditions used in the abnormality judgment process performed by the second judgment unit 126 based on the modified maintenance history data.

Specifically, the condition setting unit 124 changes parameters for defining the "feature amount" calculated by the second judgment unit 126, thresholds for determining whether the state of the production robot 80 is classified as "normal" or "abnormal", etc.

The condition setting unit 124 sets or modifies the judgment conditions used in the abnormality judgment process performed by the second judgment unit 126 based on the maintenance history data, so that the second judgment unit 126 can be said to obtain the result of judging an abnormality in the production robot 80 based on the maintenance history data and the status data as the second judgment result.

The determination unit 128 determines whether or not there is an abnormality in the production robot 80 based on the first judgment result and the second judgment result. Information on the determined presence or absence of an abnormality in the production robot 80 is output to the display unit 90 via the output unit 130.

In addition, the decision unit 128 may control whether or not the first judgment unit 122 obtains a first judgment result and the second judgment unit 126 obtains a second judgment result, based on whether or not maintenance history data corresponding to maintenance work on the production robot 80 performed within a specified period in the past (within a period going back a specified period from the time when the abnormality processing is judged to the time when the abnormality processing is judged) is registered in the database 110.

Here, the specified period refers to a period of time, such as one week, one month, or one year. The specified period may be set in advance for each production robot 80, taking into consideration the timing and characteristics of the maintenance of the production robot 80, the work performed by the production robot 80, etc.

[Procedure for determining abnormality]
Next, an example of the procedure for the abnormality determination process according to this embodiment will be described. Fig. 2 is a flowchart showing the procedure for the abnormality determination device according to this embodiment. The abnormality determination process shown in Fig. 2 is started when an instruction to start abnormality diagnosis is received from a supervisor or maintenance person, or at a predetermined timing while the production robot 80 is running, and is repeatedly executed every time status data of the production robot 80 is input to the abnormality determination device 100.

In step S101, the control unit 120 reads status data of the production robot 80 from the database 110. The read status data is input to the first judgment unit 122 and the second judgment unit 126. The first judgment unit 122 performs an abnormality judgment process based on the input status data and obtains a first judgment result.

In step S105, if the first judgment result is normal (YES in step S105), the process proceeds to step S119, where the determination unit 128 determines that the production robot 80 is normal, and the output unit 130 outputs information indicating that the production robot 80 is normal.

On the other hand, if the first judgment result is not normal in step S105 (NO in step S105), the process proceeds to step S107, where the condition setting unit 124 requests the user to check the maintenance history data of the production robot 80.

If it is confirmed in step S109 that the maintenance history data is normal (YES in step S109), the process proceeds to step S117, where the determination unit 128 determines that the production robot 80 is abnormal, and the output unit 130 outputs information indicating that the production robot 80 is abnormal.

On the other hand, if it is determined in step S109 that the maintenance history data is not normal (NO in step S109), the process proceeds to step S111, where the condition setting unit 124 requests the user to correct the maintenance history data of the production robot 80.

In step S113, if the user modifies the stored maintenance history data, the condition setting unit 124 modifies the judgment conditions used in the abnormality judgment process performed by the second judgment unit 126 based on the modified maintenance history data. The second judgment unit 126 performs the abnormality judgment process based on the modified judgment conditions and the input status data, and obtains a second judgment result.

In step S115, if the second judgment result is normal (YES in step S115), proceed to step S119. On the other hand, if the second judgment result is not normal (NO in step S115), proceed to step S117.

After performing the processing in step S117 or step S119, the abnormality determination processing shown in FIG. 2 ends.

In this way, the processes from step S107 to step S115 check whether the maintenance history data is normal, and if necessary, correct the maintenance history data, thereby preventing the production robot 80 from being mistakenly determined to be abnormal due to incorrect input.

[Processing procedure for abnormality determination according to the first modified example]
Next, an example of a processing procedure for abnormality determination according to a first modified example of this embodiment will be described. Fig. 3 is a flowchart showing the processing procedure of the abnormality determination device according to the first modified example of the present invention. The abnormality determination processing shown in Fig. 3 differs from the abnormality determination processing shown in Fig. 2 in that step S203 is executed after step S101, but is otherwise the same processing.

In step S203, the decision unit 128 determines whether or not maintenance history data corresponding to maintenance work performed on the production robot 80 within a specified period in the past is registered in the database 110. If maintenance history data within the specified period is registered (YES in step S203), the process proceeds to step S107, where processing is performed based on whether or not the first determination result is normal.

On the other hand, if maintenance history data within the specified period has not been registered (NO in step S203), proceed to step S115 and perform processing based on whether the second judgment result is normal or not.

If maintenance history data within a specified period has not been registered, it is possible that at the time when the abnormality processing is judged, a specified period of time has passed since the maintenance work was performed, and the effects of various works will not be reflected in the status data. Therefore, a situation may arise in which it is not necessary to check whether the maintenance history data is normal. In anticipation of such a situation, if maintenance history data within a specified period has not been registered, processing based on whether the first judgment result is normal or not is omitted, reducing the user's confirmation work.

[Processing procedure for abnormality determination according to the second modified example]
Next, an example of a processing procedure for abnormality determination according to a second modified example of this embodiment will be described. Fig. 4 is a flowchart showing the processing procedure of the abnormality determination device according to the second modified example of the present invention. The abnormality determination processing shown in Fig. 4 differs from the abnormality determination processing shown in Fig. 2 in that step S303 is executed after step S101, but is otherwise the same processing.

In step S303, the second judgment unit 126 performs an abnormality judgment process based on the judgment conditions before correction and the input state data, and obtains a second judgment result. Then, if the second judgment result is normal (YES in step S303), the process proceeds to step S105. On the other hand, if the second judgment result is not normal (NO in step S303), the process proceeds to step S117.

Compared to the abnormality determination process shown in FIG. 2, the abnormality determination process shown in FIG. 4 performs processing based on whether the second determination result is normal before performing processing based on whether the first determination result is normal, thereby reducing the likelihood of an abnormality in the production robot 80 being overlooked.

[Effects of the embodiment]
As described in detail above, the abnormality judgment device and abnormality judgment method of this embodiment obtain a result of judging whether or not the device is in a predetermined state based on status data of the device as a first judgment result, and when the device is in the predetermined state, obtain maintenance history data of the device via the operation data input unit, obtain a result of judging whether or not there is an abnormality in the device based on the maintenance history data and the status data as a second judgment result, and output the presence or absence of an abnormality in the device based on the first judgment result and the second judgment result.

As a result, even if data is entered incorrectly during maintenance and inspection of the device, it is possible to distinguish between incorrect data entry and determine whether the device is normal or abnormal. In particular, even if maintenance history data relating to maintenance work different from the maintenance work actually performed is recorded, it is possible to detect that maintenance history data relating to different maintenance work has been recorded and correct the maintenance history data.

Then, based on the corrected maintenance history data, it is possible to perform anomaly determination processing in accordance with the corresponding maintenance work and to predict the occurrence of anomalies. As a result, it is possible to reduce false alarms and oversights in anomaly detection and prediction.

In addition, in the abnormality determination device and abnormality determination method according to this embodiment, when the first determination result indicates an abnormality in the device and maintenance history data corresponding to maintenance of the device within a predetermined period in the past exists, the maintenance history data of the device may be acquired via the operation data input unit. This makes it possible to narrow down the targets of abnormalities detected based on the first determination result obtained by the first determination unit to cases where there is a high possibility that maintenance work different from the maintenance work actually performed has been recorded, thereby improving the accuracy of abnormality determination based on the first determination result.

Furthermore, in the abnormality determination device and abnormality determination method according to this embodiment, when the first determination result indicates an abnormality in the device and maintenance history data corresponding to maintenance of the device within a specified period in the past does not exist, maintenance history data of the device may not be acquired via the operation data input unit.

If maintenance history data within a specified period has not been registered, it is possible that at the time when the abnormality processing is judged, a specified period of time has passed since the maintenance work was performed, and the effects of various works will not be reflected in the status data. Therefore, a situation may arise in which it is not necessary to check whether the maintenance history data is normal. In anticipation of such a situation, if maintenance history data within a specified period has not been registered, processing based on whether the first judgment result is normal or not is omitted, thereby reducing the burden of the checking work on the user.

In addition, in the abnormality determination device and abnormality determination method according to this embodiment, the first determination result and the second determination result may be generated based on different determination algorithms. This makes it possible to diagnose abnormalities in the maintenance history data and abnormalities in the facilities, equipment, machines, and devices separately, and to distinguish and determine abnormalities in the maintenance history data and abnormalities in the facilities, equipment, machines, and devices.

Furthermore, in the abnormality determination device and abnormality determination method according to this embodiment, after the display unit displays past maintenance history data, the operation data input unit may acquire operation data for correcting the displayed maintenance history data. This allows the user to accurately recognize the registered maintenance history data related to the abnormality detected by the first determination unit.

In addition, in the abnormality determination device and abnormality determination method according to this embodiment, when the display unit displays past maintenance history data, it may also display reference maintenance information. This allows the past maintenance history data to be compared with the reference maintenance information, and the user can accurately recognize the registered maintenance history data before correcting the maintenance history data.

Furthermore, in the abnormality determination device and abnormality determination method according to this embodiment, the specified state may be a state different from the normal state of the device. This makes it possible to detect only abnormalities that have occurred due to the influence of maintenance history data.

In the abnormality determination device and abnormality determination method according to this embodiment, the predetermined state may be a state in which the correlation between multiple state data of the device is different from the correlation between state data in the normal state of the device. This makes it possible to detect abnormalities with higher accuracy than by comparing each signal with the normal state.

Furthermore, in the abnormality determination device and abnormality determination method according to this embodiment, the normal state may be a state that is determined to be normal based on a normality determination model generated by learning state data when the device is normal.

When detecting anomalies by comparing with a uniformly defined normal state, there may be cases where an anomaly cannot be diagnosed using an appropriate model (for example, when the state data for normal cases is non-normally distributed data, and the threshold is uniquely determined using a normal distribution method). In contrast, the accuracy of anomaly detection can be further improved by diagnosing anomalies based on a normality determination model generated by learning state data for normal cases.

In addition, in the abnormality determination device and abnormality determination method according to this embodiment, the normal state may be updated. This allows the updated standard normal data to be used even when the contents of the regular maintenance work or the maintenance work that is usually performed are changed, making it possible to reduce false alarms and oversights.

Each of the functions described in the above embodiments may be implemented by one or more processing circuits. Processing circuits include programmed processors, electrical circuits, and even devices such as application specific integrated circuits (ASICs), and circuit components arranged to perform the described functions.

The contents of the present invention have been described above in accordance with the embodiments, but the present invention is not limited to these descriptions, and it will be obvious to those skilled in the art that various modifications and improvements are possible. The descriptions and drawings that form part of this disclosure should not be understood as limiting the present invention. Various alternative embodiments, examples, and operating techniques will be apparent to those skilled in the art from this disclosure.

The present invention naturally includes various embodiments not described here. Therefore, the technical scope of the present invention is determined only by the invention-specific matters related to the scope of the claims that are appropriate from the above description.

REFERENCE SIGNS LIST 100 Abnormality determination device 102 Input unit 110 Database 120 Control unit 122 First determination unit 124 Condition setting unit 126 Second determination unit 128 Determination unit 130 Output unit 80 Production robot 82 Communication unit 84 Measurement unit 86 Sensor 90 Display unit 95 Operation unit (operation data input unit)

Claims (10)

An abnormality determination device for determining an abnormality in an apparatus, comprising: an input unit to which data output from a sensor that detects a state of the apparatus is input; an operation data input unit to which data based on a user's operation is input; and a controller,
The controller:
acquiring status data of the device output from the sensor via the input unit;
acquiring a result of determining whether the device is in a predetermined state based on the state data as a first determination result;
When the first determination result indicates that the device is in the predetermined state, maintenance history data of the device is acquired via the operation data input unit;
Setting or modifying a judgment condition based on the acquired maintenance history data;
A result of determining whether or not the device is abnormal using the determination condition is acquired as a second determination result;
An abnormality determination device, comprising: an output indicating whether or not an abnormality exists in the device based on the first determination result and the second determination result. The abnormality determination device according to claim 1,
The controller:
An abnormality judgment device characterized by acquiring the maintenance history data of the device via the operation data input unit when the first judgment result indicates an abnormality of the device and maintenance history data corresponding to maintenance of the device within a specified period in the past exists. The abnormality determination device according to claim 1 or 2 ,
The anomaly determination device according to claim 1, wherein the first determination result and the second determination result are generated based on different determination algorithms. The abnormality determination device according to any one of claims 1 to 3 ,
Further, a display unit is provided,
An abnormality determination device, characterized in that after the display unit displays the past maintenance history data, the operation data input unit acquires operation data for correcting the displayed maintenance history data. The abnormality determination device according to claim 4 ,
The abnormality determination device according to claim 1, wherein when the display unit displays the past maintenance history data, the display unit also displays reference maintenance information. The abnormality determination device according to any one of claims 1 to 5 ,
The abnormality determination device, wherein the predetermined state is a state different from a normal state of the device. The abnormality determination device according to any one of claims 1 to 6 ,
The abnormality determination device, wherein the predetermined state is a state in which a correlation between a plurality of the status data of the device is different from a correlation between the status data in a normal state of the device. The abnormality determination device according to claim 6 or 7 ,
The abnormality determination device is characterized in that the normal state is a state that is determined to be normal based on a normality determination model generated by learning the state data when the device is normal. The abnormality determination device according to any one of claims 6 to 8 ,
The abnormality determination device is characterized in that the normal state is updated. An abnormality determination method for controlling an input unit to which data output from a sensor that detects a state of an apparatus is input, an operation data input unit to which data based on a user's operation is input, and a connected controller, comprising:
The controller:
acquiring status data of the device output from the sensor via the input unit ;
acquiring a result of determining whether the device is in a predetermined state based on the state data as a first determination result;
When the first determination result indicates that the device is in the predetermined state, acquiring maintenance history data of the device via an operation data input unit;
Setting or modifying a judgment condition based on the acquired maintenance history data;
A result of determining whether or not the device is abnormal using the determination condition is acquired as a second determination result;
An abnormality determination method, comprising: outputting the presence or absence of an abnormality in the device based on the first determination result and the second determination result.

Images