Abnormality detection assistance device, abnormality detection assistance method, and abnormality detection assistance program

Abstract

The present invention accurately determines an abnormality occurring in a control system. This abnormality detection assistance device comprises: a storage unit that stores waveform data at the time of an abnormality in association with an operation method of a device corresponding to the waveform data at the time of the abnormality for each abnormality cause regarding a system including a device to be controlled; an identification unit that acquires measurement data when an abnormality probability equal to or greater than a prescribed threshold is calculated when an abnormality probability calculated by a first abnormality detection algorithm becomes equal to or greater than the prescribed threshold on the basis of measurement data measured during the operation of the system, and identifies the cause of the abnormality; and an output unit that outputs the identified cause of the abnormality.

Description

Anomaly detection support device, anomaly detection support method, and anomaly detection support program

[0001] The present disclosure relates to an anomaly detection support device, an anomaly detection support method, and an anomaly detection support program.

[0002] Various control systems installed in substrate manufacturing equipment or peripheral devices are equipped with anomaly detection algorithms that detect abnormalities when they occur. The anomaly detection algorithms determine whether an abnormality has occurred based on measurement data measured while the control system is in operation, for example.

[0003] On the other hand, since the occurrence of an abnormality is caused by changes over time in various characteristics within the control system, an operating control system controls the controlled object so as to minimize the impact of these changes. For this reason, in an operating control system, the occurrence of an abnormality is difficult to detect, and the accuracy of the judgment of the anomaly detection algorithm and the accuracy of identifying the cause of the abnormality may decrease.

[0004] US Patent Application Publication No. 2023 / 0195074

[0005] The present disclosure accurately determines an abnormality occurring in a control system.

[0006] An anomaly detection support device according to one aspect of the present disclosure has, for example, the following configuration: a storage unit that stores, for each anomaly cause, waveform data at the time of an anomaly and an operation method of the device corresponding to the waveform data at the time of the anomaly, in association with each other, for a system including an apparatus to be controlled, an identification unit that, when the anomaly probability calculated by a first anomaly detection algorithm based on measurement data measured while the system is operating becomes equal to or greater than a predetermined threshold, acquires the measurement data at which the anomaly probability equal to or greater than the predetermined threshold is calculated, and identifies the anomaly cause, and an output unit that outputs the identified anomaly cause.

[0007] According to the present disclosure, it becomes possible to accurately determine abnormalities occurring in a control system.

[0008] FIG. 1 is a diagram illustrating an application example of an anomaly detection support device. FIG. 2 is a diagram illustrating an overview of processing performed by the anomaly detection support device. FIG. 3 is a diagram illustrating an example of the hardware configuration of the anomaly detection support device. FIG. 4 is a diagram illustrating an example of the functional configuration of the anomaly detection support device. FIG. 5 is an example of a list stored in a list storage unit. FIG. 6 is an example of a flowchart illustrating the flow of monitoring processing performed by the anomaly detection support device. FIG. 7 is an example of a flowchart illustrating the flow of an anomaly detection support processing performed by the anomaly detection support device. FIG. 8 is an example of a model of a controlled object. FIG. 9 is a diagram illustrating an example of data that can be acquired by the anomaly detection support device while a control system is in operation. FIG. 10 is a first diagram illustrating an example of data acquired by the anomaly detection support device when the control system is not in operation. FIG. 11 is a second diagram illustrating an example of data acquired by the anomaly detection support device when the control system is not in operation. FIG. 12 is a third diagram illustrating an example of data acquired by the anomaly detection support device when the control system is not in operation. FIG. 13 is a first diagram illustrating an example of another control block. FIG. 14 is a second diagram illustrating an example of another control block.

[0009] Hereinafter, each embodiment will be described with reference to the accompanying drawings. In this specification and drawings, components having substantially the same functional configurations are designated by the same reference numerals, and redundant description will be omitted.

[0010] [First embodiment] <Application example of an anomaly detection support device> First, an application example of an anomaly detection support device according to a first embodiment will be described. Fig. 1 is a diagram showing an application example of an anomaly detection support device. As shown in Fig. 1, an anomaly detection support device 120 according to the first embodiment is applied to a control system 110.

[0011] The control system 110 to which the anomaly detection support device 120 according to the first embodiment is applied may have any configuration. In the example of Fig. 1 , the control system 110 includes an anomaly detection device 111, a measurement data acquisition device 112, a measurement device 113, a control device 114, a measurement device 115, and an operation device 116.

[0012] The anomaly detection device 111 acquires measurement data measured by the measurement device 113 or the measurement device 115 via the measurement data acquisition device 112. The anomaly detection device 111 calculates the probability that an anomaly has occurred in the control system 110 (referred to as the "first anomaly probability") based on the measurement data measured while the control system 110 is in operation.

[0013] The first abnormality probability calculated by the abnormality detection device 111 is notified to the abnormality detection support device 120 .

[0014] When calculating the first anomaly probability based on the measurement data measured while the control system 110 is in operation, the anomaly detection device 111 calculates the first anomaly probability using, for example, a machine learning model that has been trained in advance based on the measurement data measured when an anomaly occurs. Alternatively, the anomaly detection device 111 may calculate the first anomaly probability using, for example, a physical model that is constructed to detect the occurrence of an anomaly based on the measurement data.

[0015] Furthermore, when a machine learning model is used, any algorithm may be adopted as the first anomaly detection algorithm. For example, an algorithm such as a convolutional neural network (CNN) or a one-dimensional convolutional neural network (1D-CNN) may be adopted as the first anomaly detection algorithm. Alternatively, an algorithm such as a long short-term memory (LSTM), a CNN+LSTM, or a transformer may be adopted as the first anomaly detection algorithm.

[0016] The measurement data acquisition device 112 acquires measurement data measured by the measurement device 113 or the measurement device 115 and notifies the anomaly detection device 111 and the anomaly detection support device 120 of the data.

[0017] The measurement device 113 refers to a measurement device, among various measurement devices in the control system 110, that is not involved in the feedback control of the operation device 116. The measurement data measured by the measurement device 113 is notified to the measurement data acquisition device 112.

[0018] The measuring device 115 refers to a measuring device involved in the feedback control of the operating device 116, among the various measuring devices in the control system 110. Therefore, if the control device 114 is a control device that does not perform feedback control on the operating device 116, the control system 110 does not need to have the measuring device 115. On the other hand, if the control device 114 is a control device that performs feedback control on the operating device 116, the control system 110 has the measuring device 115, and the measurement data measured by the measuring device 115 is notified to the measurement data acquisition device 112 and the control device 114.

[0019] When performing feedback control on the operating device 116, the control device 114 acquires measurement data from the measuring device 115 and generates an operating signal based on the acquired measurement data and target data. The control device 114 also transmits the generated operating signal to the operating device 116. When the control system 110 is not operating, the control device 114 generates the operating signal based on the acquired measurement data and target data, using the target data included in the operating method notified by the anomaly detection support device 120. When the control system 110 is not operating, the control device 114 also notifies the anomaly detection support device 120 of the generated operating signal.

[0020] In this embodiment, "when control system 110 is not operating" specifically refers to the following time. For example, if control system 110 is one of multiple control systems mounted on a maintenance robot (an example of a peripheral device) that performs maintenance on a substrate manufacturing apparatus, "when control system 110 is not operating" refers to the idle time of control system 110 itself. In other words, "when control system 110 is not operating" refers to the idle time of each control system mounted on the maintenance robot, rather than the idle time of the entire maintenance robot on which control system 110 is mounted.

[0021] When the control device 114 does not perform feedback control on the operation device 116, it transmits a predetermined operation signal to the operation device 116. When the control system 110 is not in operation, the control device 114 transmits to the operation device 116 an operation signal included in the operation method notified by the anomaly detection support device 120.

[0022] The operating device 116 is a device to be controlled within the control system 110 and operates based on an operating signal sent from the control device 114 .

[0023] The anomaly detection support device 120 monitors the first anomaly probability notified by the anomaly detection device 111 while the control system 110 is operating, and if the first anomaly probability is greater than or equal to a predetermined threshold, it determines that an anomaly may have occurred in the control system 110.

[0024] The abnormality detection support device 120 acquires measurement data from the measurement data acquisition device 112 when it determines that an abnormality may have occurred, and thereby identifies the type of abnormality that may have occurred (the cause of the abnormality).

[0025] The anomaly detection support device 120 notifies the control device 114 of an operation method of the operation device 116 that corresponds to the identified cause of the anomaly and that is for making it clear that an anomaly has occurred in the control system 110. The anomaly detection support device 120 also displays the identified cause of the anomaly and the operation method of the operation device 116 that corresponds to the identified cause of the anomaly. Note that the operation method of the operation device 116 includes an operation method of the operation device 116 during operation and an operation method of the operation device 116 for making it clear that an anomaly has occurred when the operation device 116 is not in operation (an operation method with stricter operating conditions than the operation method of the operation device 116 during operation).

[0026] Based on the notified operation method, the abnormality detection support device 120 acquires from the control device 114 the operation signal when the control device 114 operates the operating device 116 when the control system 110 is not in operation, and acquires measurement data from the measurement data acquisition device 112.

[0027] The anomaly detection support device 120 calculates the probability that an anomaly has occurred in the control system 110 (referred to as the "second anomaly probability") based on either or both of the operation signal acquired from the control device 114 and the measurement data acquired from the measurement data acquisition device 112. If the second anomaly probability is equal to or greater than a predetermined threshold, the anomaly detection support device 120 determines that an anomaly has occurred in the control system 110 and displays the calculated probability in association with the identified cause of the anomaly.

[0028] The anomaly detection support device 120 acquires measurement data measured by the control device 114 operating the operating device 116 based on the notified operation method, and calculates the second anomaly probability based on the acquired measurement data using a pre-trained machine learning model. Alternatively, the anomaly detection support device 120 may calculate the second anomaly probability using, for example, a physical model constructed to detect the occurrence of an anomaly based on the measurement data.

[0029] Furthermore, when a machine learning model is used, any algorithm may be adopted as the second anomaly detection algorithm. For example, algorithms such as a convolutional neural network (CNN) or a one-dimensional convolutional neural network (1D-CNN) may be adopted as the second anomaly detection algorithm. Alternatively, algorithms such as a long short-term memory (LSTM), a CNN+LSTM, or a transformer may be adopted as the second anomaly detection algorithm.

[0030] In this way, the anomaly detection support device 120 according to the first embodiment executes a process (monitoring process) to monitor whether an anomaly may have occurred in the control system 110 while the control system 110 is in operation. Furthermore, when the anomaly detection support device 120 according to the first embodiment determines that an anomaly may have occurred, it: identifies the cause of the anomaly using measurement data obtained when it determines that an anomaly may have occurred; and, when the control system 110 is not in operation, executes an anomaly detection support process using an operation method corresponding to the identified cause of the anomaly, thereby making it apparent that an anomaly has occurred in the control system 110.

[0031] As a result, the anomaly detection support device 120 can reliably determine whether or not an anomaly has occurred in the control system 110. In other words, the anomaly detection support device 120 makes it possible to accurately determine whether an anomaly has occurred in the control system 110.

[0032] Note that "determining that an abnormality may have occurred" also includes detecting a sign of an abnormality, because it is possible to detect a sign of an abnormality depending on how the threshold for the first abnormality probability is set. Similarly, "determining that an abnormality has occurred" also includes detecting a sign of an abnormality, because it is possible to detect a sign of an abnormality depending on how the threshold for the second abnormality probability is set.

[0033] Setting the threshold value to detect signs of anomalies has the advantage of making it possible to prevent anomalies from occurring in the control system 110. However, if the threshold value for the first anomaly probability is set to detect signs of anomalies, it is conceivable that the anomaly detection support process will be executed more frequently. For this reason, it is effective to optimize the timing of setting the threshold value for the first anomaly probability to detect signs of anomalies, for example, by limiting it to before scheduled maintenance.

[0034] <Outline of Processing by Anomaly Detection Support Device> Next, an outline of processing by the anomaly detection support device 120 will be described. Fig. 2 is a diagram for explaining an outline of processing by the anomaly detection support device.

[0035] 2, reference numeral 200 indicates the operating status of the control system 110. Also, reference numeral 210 indicates that a "monitoring process" is executed by the anomaly detection support device 120 while the control system 110 is operating. The example of reference numeral 210 indicates that the first anomaly probability notified by the anomaly detection device 111 while the control system 110 is operating has been continuously below the threshold value.

[0036] As shown by the symbol 220 in Figure 2, if the first abnormality probability notified by the anomaly detection device 111 while the control system 110 is operating is continuously below the threshold, the anomaly detection support device 120 will not execute the anomaly detection support process when the control system 110 is not operating immediately thereafter.

[0037] On the other hand, in Figure 2, symbol 230 indicates a case where, as a result of monitoring processing being performed by the anomaly detection support device 120 while the control system 110 is in operation, the first anomaly probability notified by the anomaly detection device 111 includes a first anomaly probability that is equal to or greater than the threshold value.

[0038] In this case, the abnormality detection support device 120 determines that an abnormality may have occurred in the control system 110, identifies the cause of the abnormality based on the measurement data, and identifies the operating method of the operating device 116 in accordance with the identified cause of the abnormality.

[0039] Furthermore, the abnormality detection support device 120 displays the identified cause of the abnormality and the operating method of the operating device 116 corresponding to the identified cause of the abnormality.

[0040] As a result, as shown in FIG. 2, the abnormality detection support device 120 executes the "abnormality detection support process", for example, immediately after the control system 110 is out of operation.

[0041] 2 , reference numeral 240 indicates the state in which the anomaly detection support process is executed by the anomaly detection support device 120. As indicated by reference numeral 240, the anomaly detection support device 120 causes the control system 110 to perform an operation based on the identified operation method, thereby calculating a second anomaly probability and determining whether an anomaly has occurred in the control system 110.

[0042] In addition, if the calculated second abnormality probability is equal to or greater than a predetermined threshold and it is determined that an abnormality has occurred in the control system 110, the abnormality detection support device 120 displays the calculated second abnormality probability in association with the identified cause of the abnormality.

[0043] This allows maintenance to be performed on the abnormal location corresponding to the cause of the abnormality identified by the abnormality detection support device 120 when the maintenance period begins (for example, the maintenance period for the entire maintenance robot) (symbol 250).

[0044] When the maintenance period ends, the control system 110 resumes operation as indicated by reference numeral 260 in Fig. 2. This causes the abnormality detection support device 120 to perform the monitoring process again.

[0045] <Hardware Configuration of the Anomaly Detection Support Device> Next, the hardware configuration of the anomaly detection support device 120 will be described. Fig. 3 is a diagram showing an example of the hardware configuration of the anomaly detection support device. As shown in Fig. 3, the anomaly detection support device 120 includes a processor 301, a memory 302, an auxiliary storage device 303, a connection device 304, a communication device 305, and a drive device 306. Note that the various pieces of hardware included in the anomaly detection support device 120 are connected to each other via a bus 307.

[0046] The processor 301 has various computing devices such as a CPU (Central Processing Unit), a GPU (Graphics Processing Unit), etc. The processor 301 reads various programs (for example, an anomaly detection support program) into the memory 302 and executes them.

[0047] The memory 302 has a main storage device such as a read-only memory (ROM) or a random access memory (RAM). The processor 301 and the memory 302 form a so-called computer, and the processor 301 executes various programs read onto the memory 302, causing the computer to realize various functions.

[0048] The auxiliary storage device 303 stores various programs and various information used when the various programs are executed by the processor 301. A list storage unit 410, which will be described later, is realized in the auxiliary storage device 303.

[0049] The connection device 304 is a connection device that connects to external devices (such as an operation device 311 and a display device 312).

[0050] The communication device 305 is a communication device for transmitting and receiving various information between the anomaly detection device 111 , the measurement data acquisition device 112 , the control device 114 and the anomaly detection support device 120 .

[0051] The drive device 306 is a device for setting the recording medium 313. The recording medium 313 here includes media that record information optically, electrically, or magnetically, such as CD-ROMs, flexible disks, magneto-optical disks, etc. The recording medium 313 may also include semiconductor memories that record information electrically, such as ROMs, flash memories, etc.

[0052] The various programs to be installed in the auxiliary storage device 303 are installed, for example, by setting the distributed recording medium 313 in the drive device 306 and reading the various programs recorded on the recording medium 313 by the drive device 306. Alternatively, the various programs to be installed in the auxiliary storage device 303 may be installed by being downloaded from a network (not shown) via the communication device 305.

[0053] <Functional Configuration of Anomaly Detection Support Device> Next, the functional configuration of the anomaly detection support device 120 will be described. Fig. 4 is a diagram showing an example of the functional configuration of the anomaly detection support device. As described above, an anomaly detection support program is installed in the anomaly detection support device 120. By executing this program in the anomaly detection support device 120, the anomaly detection support device 120 functions as a first anomaly probability monitoring unit 401, an anomaly cause identification unit 402, an operation method output unit 403, a second anomaly probability calculation unit 404, and a display unit 405.

[0054] The first anomaly probability monitoring unit 401 acquires the first anomaly probability notified from the anomaly detection device 111 while the control system 110 is in operation, and monitors whether the acquired first anomaly probability is equal to or greater than a predetermined threshold. If the first anomaly probability is determined to be equal to or greater than the predetermined threshold, the first anomaly probability monitoring unit 401 notifies the anomaly cause identification unit 402.

[0055] When the anomaly cause identification unit 402 is notified by the first anomaly probability monitoring unit 401 that the first anomaly probability is equal to or greater than a predetermined threshold, the anomaly cause identification unit 402 determines that there is a possibility that an anomaly has occurred.

[0056] The anomaly cause identification unit 402 acquires the measurement data obtained when it is determined that an anomaly may have occurred from the measurement data acquisition device 112. When the anomaly cause identification unit 402 acquires the measurement data obtained when it is determined that an anomaly may have occurred, it refers to the list storage unit 410 and searches for waveform data similar to the acquired measurement data.

[0057] The list storage unit 410 stores a list in which the causes of anomalies, waveform data used for similarity determination, and operation methods of the operating device 116 corresponding to the causes of anomalies are previously described. The waveform data here includes at least one of time-series data, data indicating the shape of the time-series data, and feature quantities indicating the shape of the time-series data. The anomaly cause identification unit 402 identifies the cause of anomaly and the operation method associated with the waveform data searched for in the list. The anomaly cause identification unit 402 also notifies the operation method output unit 403 of the identified cause of anomaly and operation method.

[0058] The operation method output unit 403 is an example of an output unit, and when it is notified of the cause of the abnormality and the operation method by the abnormality cause identification unit 402, it notifies the display unit 405 of the cause of the abnormality and the operation method, and also notifies the control device 114 of the operation method.

[0059] Based on the notified operation method, the second anomaly probability calculation unit 404 acquires, from the measurement data acquisition device 112, measurement data obtained when the control device 114 operates the operating device 116 while the control system 110 is not in operation. Alternatively, based on the notified operation method, the second anomaly probability calculation unit 404 acquires, from the control device 114, an operation signal obtained when the control device 114 operates the operating device 116 while the control system 110 is not in operation.

[0060] The second anomaly probability calculation unit 404 calculates a second anomaly probability based on either or both of the operation signal acquired from the control device 114 and the measurement data acquired from the measurement data acquisition device 112, and determines whether the calculated second anomaly probability is equal to or greater than a predetermined threshold. If the calculated second anomaly probability is equal to or greater than the predetermined threshold, the anomaly detection support device 120 determines that an anomaly has occurred in the control system 110, and notifies the display unit 405 of the second anomaly probability in association with the identified cause of the anomaly.

[0061] When the display unit 405 is notified of the cause of the abnormality and the operation method by the operation method output unit 403, it displays them on the display device 312. Furthermore, when the display unit 405 is notified of the cause of the abnormality and the second abnormality probability by the second abnormality probability calculation unit 404, it displays them on the display device 312.

[0062] <Example of List> Next, a description will be given of the list stored in the list storage unit 410. Fig. 5 shows an example of a list stored in the list storage unit.

[0063] As shown in FIG. 5, the list 500 includes information items such as "number," "cause of abnormality," "waveform data," and "diagnosis operation."

[0064] In the "number" field, an identifier for identifying waveform data used to determine whether it is similar to measurement data when it is determined that an abnormality may have occurred is stored for each cause of the abnormality.

[0065] The "cause of abnormality" field stores the type of abnormality that can occur in the control system 110. The example in Fig. 5 shows that "abnormality A" and "abnormality B" are stored as the types of abnormality.

[0066] For example, suppose that control system 110 is an arm motion control system that is one of multiple control systems mounted on a maintenance robot that maintains substrate manufacturing equipment, and operating device 116 is a system including a motor, gears, belts, bearings, etc. In this case, abnormality A includes, for example, abnormalities such as deterioration of the gears and belts, or a decrease in torsional rigidity (change in vibration frequency) due to belt breakage, etc. Furthermore, abnormality B includes, for example, abnormalities such as depletion of grease in the bearings, an increase in the viscous friction coefficient (decrease in operating speed) due to grease abnormalities, etc.

[0067] "Waveform data" stores waveform data used to determine whether the data is similar to the measurement data when it is determined that an abnormality may have occurred in the control system 110. The example in Fig. 5 shows that for the abnormality cause = "abnormality A", the following are stored: 10% of the waveform data for abnormality A, 30% of the waveform data for abnormality A, and 50% of the waveform data for abnormality A.

[0068] Note that the 10% waveform data for abnormality A refers to the waveform of measurement data in a state where torsional rigidity has decreased by 10% when, for example, abnormality A is "deterioration of the gear and belt, or a decrease in torsional rigidity due to a break in the belt, etc." Similarly, the 30% and 50% waveform data for abnormality A refer to the waveforms of measurement data in states where torsional rigidity has decreased by 30% and 50%, respectively. However, these waveform data may be waveforms of measurement data obtained by actual measurement, or waveforms of measurement data obtained by simulation.

[0069] Furthermore, the example of FIG. 5 shows that for the abnormality cause = "abnormality B", the following have been stored: - 110% waveform data for abnormality B; - 130% waveform data for abnormality B; - 150% waveform data for abnormality B.

[0070] Note that the 110% waveform data for abnormality B refers to the waveform of measurement data when the viscous friction coefficient has increased to 110% when, for example, abnormality B = "an increase in the viscous friction coefficient due to grease depletion in the bearing, a grease abnormality, etc." Similarly, the 130% and 150% waveform data for abnormality B refer to the waveforms of measurement data when the viscous friction coefficient has increased to 130% and 150%, respectively. However, these waveform data may be waveforms of measurement data obtained by actual measurement, or may be waveforms of measurement data obtained by simulation.

[0071] An operation method according to the cause of the abnormality is stored in the "diagnostic operation." The example of FIG. 5 shows that when the abnormality cause is "abnormality A," the diagnostic operation is "operation method A," and when the abnormality cause is "abnormality B," the diagnostic operation is "operation method B." As described above, the operation method is an operation method of the operating device 116 for making it apparent that an abnormality has occurred in the control system 110 (i.e., an operation method of the operating device 116 suitable for diagnosing whether or not an abnormality has occurred). In the first embodiment, the operation method includes items such as: a) a control method for the operating device 116 when not in operation (whether or not to perform feedback control); b) input data to be input to operate the operating device 116 when not in operation (a waveform of target data when feedback control is performed when not in operation, and a waveform of an operation signal to operate the operating device 116 when feedback control is not performed when not in operation); and c) data to be used to calculate the second abnormality probability (an operation signal or measurement data when feedback control is performed when not in operation, and measurement data when feedback control is not performed when not in operation).

[0072] For example, suppose that the control system 110 performs speed feedback control during operation, and abnormality B (an increase in the viscous friction coefficient (a decrease in the operating speed)) is identified as the cause of the abnormality that causes a change in the torque for maintaining the operating speed of the operating device 116. In this case, as operating method B, for example, performing speed feedback control on the operating device 116 when not in operation, inputting target data with a ramp waveform as target data for the operating speed in the speed feedback control, and using a torque signal, which is an operating signal, to calculate the second abnormality probability. By notifying the operating method, it is possible to make it apparent in the control system 110 that an abnormality has occurred in the viscous friction coefficient (operating speed).

[0073] Furthermore, suppose that the control system 110 performs speed feedback control during operation, and that abnormality A (a decrease in torsional rigidity (a decrease in vibration frequency)) is identified as the cause of the abnormality that causes a change in the vibration frequency of the operating device 116. In this case, operation method A may be, for example, the following: - not performing feedback control on the operating device 116 when not in operation; - inputting an operation signal (torque signal) in a band that includes the vibration frequency (i.e., a step-like waveform) as an operation signal that operates the operating device 116; - using the operating speed, which is measurement data, to calculate the second abnormality probability. By notifying the operation method, it is possible to make it apparent in the control system 110 that an abnormality has occurred in the torsional rigidity (vibration frequency).

[0074] Furthermore, although not shown in the list 500, it is assumed that anomaly C includes, for example, a decrease in the moment of inertia (decrease in operating acceleration) due to the falling of a part of a machine component (such as a bolt). Then, it is assumed that the control system 110 performs velocity feedback control during operation, and that the anomaly C (decrease in the moment of inertia (decrease in operating acceleration)) is identified as the cause of the anomaly that causes a change in the operating acceleration of the operating device 116. In this case, the operating method C may be, for example, performing velocity feedback control on the operating device 116 when not in operation, inputting target data with a parabolic waveform (e.g., a waveform in which acceleration gradually increases) as target data for the operating speed in the velocity feedback control, and using a torque signal, which is an operating signal, to calculate the second anomaly probability. By notifying the operating method, it is possible to make it apparent in the control system 110 that an anomaly has occurred in the moment of inertia (operating acceleration).

[0075] In the above explanation, abnormality C is a decrease in the moment of inertia (a decrease in the motion acceleration), but it may also be a fluctuation in the moment of inertia (a fluctuation in the motion acceleration), including an increase in the moment of inertia (an increase in the motion acceleration).

[0076] Furthermore, in the above explanation, the fluctuation in the moment of inertia (fluctuation in the operating acceleration) is assumed to occur due to the falling of a part of a mechanical component (such as a bolt), but the events that cause the fluctuation in the moment of inertia (fluctuation in the operating acceleration) are not limited to this.

[0077] Examples of events other than the falling of a part of a machine part (such as a bolt) include the adhesion of foreign matter to a machine part, peeling of paint from a machine part, and the loss of part of a machine part. Alternatively, if the maintenance robot has a multi-joint arm, for example, a situation may arise in which the values ​​of sensors measuring the angles of each joint deviate from the true values, causing the posture of the arm estimated from the angles of each joint to be controlled to a posture different from the original posture. Such events can also cause fluctuations in the moment of inertia (fluctuations in operational acceleration).

[0078] In this way, the operating method can be appropriately determined by, for example, selecting data in which the characteristics of the cause of the abnormality appear as item c), and selecting, as items a) and b), a control method and input data that can make it clear that an abnormality has occurred in the data selected in item c).

[0079] <Processing Flow by Anomaly Detection Support Device> Next, the flow of processing by the anomaly detection support device 120 (monitoring processing performed while the control system 110 is in operation, and anomaly detection support processing performed when the control system 110 is not in operation) will be described.

[0080] (1) Flow of Monitoring Process Fig. 6 is an example of a flowchart showing the flow of monitoring process by the anomaly detection support device. The monitoring process shown in Fig. 6 is started when the control system 110 starts operation.

[0081] In step S601, the anomaly detection support device 120 reads the list 500 from the list storage unit 410.

[0082] In step S602, the abnormality detection support device 120 acquires measurement data from the measurement data acquisition device 112 of the control system 110.

[0083] In step S603, the anomaly detection assistance device 120 acquires the first anomaly probability from the anomaly detection device 111 of the control system 110, and monitors whether the acquired first anomaly probability is equal to or greater than a threshold value.

[0084] In step S604, if the anomaly detection support device 120 determines that the acquired first anomaly probability is not equal to or greater than the threshold value (NO in step S604), the process proceeds to step S608.

[0085] On the other hand, in step S604, if the anomaly detection support device 120 determines that the acquired first anomaly probability is equal to or greater than the threshold value (YES in step S604), the process proceeds to step S605.

[0086] In step S605, the anomaly detection support device 120 searches the list 500 based on the measurement data when it determines that the acquired first anomaly probability is greater than or equal to the threshold value, thereby identifying waveform data similar to the measurement data.

[0087] In step S606, if the abnormality detection support device 120 determines that the waveform data is not similar to any of the waveform data stored in the list 500 (NO in step S606), the abnormality detection support device 120 proceeds to step S608.

[0088] On the other hand, in step S606, if the abnormality detection support device 120 determines that the waveform data is similar to any of the waveform data stored in the list 500 (YES in step S606), the abnormality detection support device 120 proceeds to step S607.

[0089] In step S607, the anomaly detection support device 120 sorts the similar waveform data in descending order of similarity, and displays the causes of the anomaly and the operating methods associated with the waveform data sorted in descending order of similarity.

[0090] In step S608, the anomaly detection support device 120 determines whether the control system 110 is in operation. If it is determined in step S608 that the control system 110 is in operation (YES in step S608), the process returns to step S602. On the other hand, if it is determined in step S608 that the control system 110 is not in operation (NO in step S608), the monitoring process ends.

[0091] (2) Flow of the Anomaly Detection Support Process Fig. 7 is an example of a flowchart showing the flow of the anomaly detection support process performed by the anomaly detection support device. The anomaly detection support process shown in Fig. 7 is executed when the control system 110 is not operating.

[0092] In step S701, the anomaly detection support device 120 notifies the control device 114 of the control system 110 of an operation method according to the cause of the anomaly. As a result, the control device 114 of the control system 110 operates the operation device 116 based on the notified operation method.

[0093] In step S702, the anomaly detection support device 120 acquires data to be used in calculating the second anomaly probability. Specifically, the anomaly detection support device 120 acquires an operation signal or measurement data when the control device 114 performs feedback control on the operation device 116. Furthermore, the anomaly detection support device 120 acquires measurement data when the control device 114 does not perform feedback control on the operation device 116. Furthermore, the anomaly detection support device 120 calculates the second anomaly probability based on the acquired operation signal or measurement data.

[0094] In step S703, the anomaly detection support device 120 determines whether the calculated second anomaly probability is equal to or greater than a threshold value.

[0095] In step S703, if it is determined that the second abnormality probability is less than the threshold value (NO in step S703), the process proceeds to step S706.

[0096] On the other hand, if it is determined in step S703 that the second abnormality probability is equal to or greater than the threshold value (YES in step S703), the process proceeds to step S704.

[0097] In step S704, the abnormality detection support device 120 identifies the cause of the abnormality (identifies the cause of the abnormality corresponding to the operation method notified in step S701).

[0098] In step S705, the anomaly detection support device 120 displays the identified anomaly cause and the corresponding second anomaly probability.

[0099] In step S706, the anomaly detection support device 120 determines whether the control device 114 of the control system 110 has operated the operation device 116 based on all of the operation methods displayed in step S608 of Fig. 6. If it is determined in step S706 that there is an operation method for which the operation device 116 has not been operated (NO in step S706), the process returns to step S701. On the other hand, if it is determined that the operation device 116 has been operated based on all of the operation methods (YES in step S706), the anomaly detection support process ends.

[0100] In addition, if the cause of the abnormality is identified in step S704, the abnormality detection support process may be terminated without operating the operating device 116 based on all of the operating methods (i.e., without performing the process of step S706).

[0101] <Details of Controlled Object> Next, details of the operating device 116 that is the controlled object will be described. Here, it is assumed that the control system 110 is an arm operation control system that is one of multiple control systems mounted on a maintenance robot, and that the operating device 116 is a system that has a motor, gears, belts, bearings, etc. In the case of a system that has a motor, gears, belts, bearings, etc., the operating device 116 can be generalized and expressed as a two-inertia system model. Figure 8 is an example of a model of the controlled object.

[0102] As indicated by reference numeral 810 in Fig. 8, the model of the two-inertia system can be roughly divided into a drive side, a load side, and a transmission part, each of which can be expressed using the equations of motion indicated by reference numeral 810. Details of the variables included in each equation of motion indicated by reference numeral 820 are as indicated by reference numeral 830.

[0103] <Specific example of data acquired by the abnormality detection support device> Next, a specific example of data acquired by the abnormality detection support device 120 by executing an abnormality detection support process on a control system 110 (here, an arm motion control system) having an operation device 116 that controls the two-inertia system model shown in Figure 8 will be described.

[0104] In describing the specific examples, first, as a comparative example, a specific example of data that can be acquired by the anomaly detection support device 120 when the anomaly detection support process is not executed will be described. That is, a specific example of data that can be acquired by the anomaly detection support device 120 while the control system 110 is operating will be described (see (1) below).

[0105] Next, a specific example of data acquired by the anomaly detection support device 120 when the anomaly detection support device 120 executes an anomaly detection support process while the control system 110 is not operating will be described (below (2)).

[0106] Next, we will explain specific examples of data that the anomaly detection support device 120 acquires when it performs an anomaly detection support process using a different operating method when the control system 110 is not in operation ((3) and (4) below).

[0107] (1) During Operation of the Control System First, a specific example of data that can be acquired by the anomaly detection support device 120 during operation of the control system 110 will be described. Fig. 9 is a diagram showing an example of data that can be acquired by the anomaly detection support device during operation of the control system.

[0108] In FIG. 9, reference numeral 910 denotes an example of a control block of the control system 110 that performs velocity feedback control on the operating device 116 (a two-inertia system model) that is the object of control.

[0109] As indicated by the reference numeral 910, the control device 114 of the control system 110 has a difference calculation unit 911. The difference calculation unit 911 receives input of target data for the rotation speed of the drive side and also receives input of measured data for the rotation speed of the drive side. The difference calculation unit 911 also calculates the difference between the target data and the measured data and inputs the difference to the control unit 912.

[0110] The control unit 912 generates an operation signal (torque signal) for rotating the drive motor in accordance with the input difference, and inputs the signal to the drive motor.

[0111] While the control system 110 is operating, the control device 114 performs speed feedback control on the operation device 116 under a control block indicated by reference numeral 910. Reference numeral 920 indicates an example of data that can be acquired by the abnormality detection support device 120 while the control system 110 is operating.

[0112] Among these, the example indicated by the reference numeral 921 is rotation speed measurement data that can be acquired by the anomaly detection support device 120, and indicates measurement data when no anomaly occurs (normal state) and measurement data when an anomaly occurs (abnormal state). Note that the example indicated by the reference numeral 921 also indicates, for reference, the target rotation speed data input when the control device 114 performs speed feedback control. Furthermore, the state in which an anomaly occurs here refers to a state in which torsional rigidity has decreased.

[0113] When the control device 114 performs speed feedback control, the measurement data follows the target data regardless of whether an abnormality occurs, as indicated by the reference numeral 921. In other words, there is no difference between the measurement data when an abnormality occurs and the measurement data when no abnormality occurs, and it can be said that it is not easy to determine whether an abnormality occurs based on the measurement data of the rotation speed while the control system 110 is operating.

[0114] In addition, the example of symbol 922 is an operating signal (torque signal) that can be acquired by the abnormality detection support device 120, and shows a torque signal (normal) when no abnormality has occurred and a torque signal (abnormal) when an abnormality has occurred.

[0115] In the control block indicated by the reference numeral 910, the operation signal (torque signal) is a signal output by the control unit 912 based on the target rotation speed data and the measured rotation speed data, and there is a slight difference between when an abnormality occurs and when no abnormality occurs. However, the difference between the two is small, and as with the measured data, it can be said that it is not easy to determine whether an abnormality has occurred based on the operation signal while the control system 110 is running.

[0116] As described above, during operation of the control system 110, the control device 114: performs speed feedback control on the operating device 116; and receives target data indicated by reference numeral 921 (target data corresponding to the operating scene is received). Therefore, the occurrence of an abnormality is unlikely to be apparent in the data obtainable by the anomaly detection support device 120. In contrast, the anomaly detection support device 120 according to the first embodiment identifies an operation method for making the occurrence of an abnormality apparent for each cause of the abnormality and notifies the control device 114 of the operation method. As a result, the occurrence of an abnormality is apparent in the data obtainable by the anomaly detection support device 120 when the control system 110 is not operating. Below, in (2) to (4), specific examples of data obtainable when the anomaly detection support device 120 notifies the control device 114 of each operation method when the control system 110 is not operating will be described.

[0117] (2) When the control system is not operating, part 1 As a first example of an operation method for making the occurrence of an abnormality apparent, here, a specific example of data acquired when the anomaly detection support device 120 notifies the control device 114 of the following operation method: - performs speed feedback control on the operating device 116, - inputs target data with a stepped waveform as target data for the rotational speed, and - uses the measurement data (rotational speed) or the operation signal (torque signal) to calculate the second abnormality probability will be described. Figure 10 is a first diagram showing an example of data acquired by the anomaly detection support device when the control system is not operating.

[0118] In FIG. 10, reference numeral 1010 denotes an example of a control block of the control system 110 that performs velocity feedback control on the operating device 116 (two-inertia system model) that is the control target, and is the same as reference numeral 910 in FIG.

[0119] When the above operation method is notified, the control device 114 performs speed feedback control on the operation device 116 under the control block indicated by reference numeral 1010 while the control system 110 is not in operation, using target data with a stepped waveform as an input. Reference numeral 1020 indicates an example of data acquired by the abnormality detection support device 120 at this time.

[0120] Among these, the example indicated by the reference numeral 1021 is measurement data of the rotation speed acquired by the anomaly detection support device 120, and indicates measurement data when no anomaly occurs (normal state) and measurement data when an anomaly occurs (abnormal state). Note that the example indicated by the reference numeral 1021 also indicates, for reference, the target data (step-like waveform) of the rotation speed input by the control device 114 during speed feedback control. Furthermore, the state in which an anomaly occurs here refers to a state in which torsional rigidity has decreased.

[0121] As shown by reference numeral 1021, even when the control device 114 performs speed feedback control, if a step-like waveform is input as target data, the difference in measurement data between when an abnormality occurs and when no abnormality occurs becomes large. In other words, the occurrence of an abnormality becomes apparent in the data acquired by the abnormality detection support device 120, and the abnormality detection support device 120 can reliably determine whether or not an abnormality has occurred.

[0122] In addition, the example of symbol 1022 is an operation signal (torque signal) acquired by the abnormality detection support device 120, and shows an operation signal when no abnormality has occurred (normal) and an operation signal when an abnormality has occurred (abnormal).

[0123] As with the rotational speed measurement data, when a step-like waveform is input as target data for the operation signal (torque signal), the difference in measurement data between when an abnormality occurs and when no abnormality occurs becomes large. In other words, the occurrence of an abnormality becomes apparent in the data acquired by the abnormality detection support device 120, allowing the abnormality detection support device 120 to reliably determine whether or not an abnormality has occurred.

[0124] (3) When the control system is not operating, part 2 As a second example of an operation method for making the occurrence of an abnormality apparent, here, a specific example of data acquired when the anomaly detection support device 120 notifies the control device 114 of the following operation method: - does not perform speed feedback control on the operation device 116, - inputs an operation signal (torque signal) with a step waveform as the operation signal (torque signal), - uses the measurement data (rotation speed) to calculate the second abnormality probability will be described. Figure 11 is a second diagram showing an example of data acquired by the anomaly detection support device when the control system is not operating.

[0125] In FIG. 11, reference numeral 1110 denotes an example of a control block of the control system 110 when velocity feedback control is not performed on the operating device 116 (two-inertia system model) that is the object of control.

[0126] As indicated by the reference numeral 1110, the control device 114 of the control system 110 inputs a predetermined operation signal (torque signal) to the drive-side motor. When the control system 110 is not operating, the predetermined operation signal (torque signal) input by the control device 114 to the drive-side motor is generated based on the operation method notified by the abnormality detection support device 120 (as described above, the operation signal (torque signal) has a step-like waveform).

[0127] Furthermore, when the control device 114 is notified of the operation method, it controls the operating device 116 under the control block indicated by reference numeral 1110 when the control system 110 is not in operation. Reference numeral 1120 indicates an example of data acquired by the anomaly detection support device 120 at this time.

[0128] Among these, the example indicated by the reference numeral 1121 is the measurement data of the rotation speed acquired by the anomaly detection support device 120, and indicates the measurement data when no anomaly occurs (normal state) and the measurement data when an anomaly occurs (abnormal state). The measurement data when an anomaly occurs (abnormal state) is shown divided into the following cases: - when the cause of the anomaly is anomaly A ("deterioration of the gear and belt, or reduction in torsional rigidity due to belt breakage, etc."), and - when the cause of the anomaly is anomaly B ("increase in the viscous friction coefficient due to bearing grease depletion, anomaly, etc.").

[0129] Furthermore, the example of reference numeral 1122 indicates: - the error between the measurement data of the rotation speed when no abnormality occurs (normal) and the measurement data of the rotation speed when abnormality A occurs (abnormal); - the error between the measurement data of the rotation speed when no abnormality occurs (normal) and the measurement data of the rotation speed when abnormality B occurs (abnormal).

[0130] As indicated by reference numeral 1122, when the cause of the abnormality is abnormality A ("deterioration of the gear and belt, or reduction in torsional rigidity due to breakage of the belt, etc."), the occurrence of the abnormality becomes more apparent by notifying the control device 114 of the following operation method: - speed feedback control is not performed on the operating device 116; - an operation signal (torque signal) with a step waveform is input as the operation signal (torque signal); - the measurement data (rotation speed) is used to calculate the second abnormality probability. In other words, according to the above operation method, the abnormality detection support device 120 can more reliably determine whether or not abnormality A has occurred.

[0131] (4) When the control system is not operating, part 3 As a third example of an operation method for making the occurrence of an abnormality apparent, here, a specific example of data acquired when the anomaly detection support device 120 notifies the control device 114 of the following operation method: - performs speed feedback control on the operation device 116, - inputs target data with a ramp waveform as target data for the rotation speed, and - uses the operation signal (torque signal) to calculate the second abnormality probability will be described. Figure 12 is a third diagram showing an example of data acquired by the anomaly detection support device when the control system is not operating.

[0132] In FIG. 12, reference numeral 1210 denotes an example of a control block of the control system 110 that performs velocity feedback control on the operating device 116 (two-inertia system model) that is the control target, and is the same as reference numeral 910 in FIG.

[0133] When the above operation method is notified, the control device 114 performs speed feedback control on the operation device 116 under the control block indicated by reference numeral 1210 while the control system 110 is not in operation, using target data with a ramp waveform as an input. Reference numeral 1220 indicates an example of data acquired by the anomaly detection support device 120 at this time.

[0134] Among these, the example of reference numeral 1221 indicates an operation signal acquired by the anomaly detection support device 120, and indicates an operation signal when no anomaly has occurred (normal operation) and an operation signal when an anomaly has occurred (abnormal operation). The operation signal when an anomaly has occurred (abnormal operation) is shown divided into the following cases: - when the cause of the anomaly is an anomaly A ("deterioration of the gear and belt, or reduction in torsional rigidity due to belt breakage, etc."), and - when the cause of the anomaly is an anomaly B ("increase in the viscous friction coefficient due to bearing grease depletion, anomaly, etc.").

[0135] Furthermore, the example of reference numeral 1222 indicates: - the error between the operation signal when no abnormality occurs (normal) and the operation signal when abnormality A occurs (abnormal); - the error between the operation signal when no abnormality occurs (normal) and the operation signal when abnormality B occurs (abnormal).

[0136] As shown by reference numeral 1222, when the cause of the abnormality is abnormality B ("an increase in the viscous friction coefficient due to bearing grease depletion, an abnormality, etc."), the occurrence of the abnormality becomes more apparent by notifying the control device 114 of the following operation method: - performing speed feedback control on the operating device 116, - inputting target data with a ramp waveform as target data for the rotational speed, - using the operating signal (torque signal) to calculate the second abnormality probability. In other words, according to the above operation method, the abnormality detection support device 120 can more reliably determine whether or not abnormality B has occurred.

[0137] <Summary> As is clear from the above explanation, the anomaly detection support device 120 according to the first embodiment: For the control system 110 including the operating device 116 to be controlled, the waveform data at the time of an anomaly and the operating method of the operating device 116 corresponding to the waveform data at the time of the anomaly are associated with each cause of anomaly and stored in the list 500. When the anomaly probability calculated by the first anomaly detection algorithm based on the measurement data measured while the control system 110 is in operation becomes equal to or exceeds a predetermined threshold, the measurement data at which the calculated anomaly probability is equal to or exceeds the predetermined threshold is acquired, and the cause of the anomaly is identified. The identified anomaly cause is displayed.

[0138] In this way, the anomaly detection support device 120 according to the first embodiment identifies the cause of an anomaly by using waveform data during operation for each cause of an anomaly. As a result, according to the first embodiment, the cause of an anomaly can be identified with high accuracy even when the control system 110 is in operation.

[0139] Furthermore, the anomaly detection support device 120 according to the first embodiment: Determines an operation method during non-operation to make the occurrence of an anomaly apparent based on the identified cause of the anomaly.

[0140] As a result, according to the first embodiment, it becomes possible to determine an abnormality while the control system 110 is operating under an operating method for making the occurrence of an abnormality apparent when the system is not in operation, and it becomes possible to accurately determine an abnormality that occurs in the control system 110.

[0141] Second Embodiment In the first embodiment described above, the control system 110 is an arm motion control system mounted on a maintenance robot, and the operating device 116 to be controlled is a system including a motor, gears, belts, bearings, etc. However, the control system 110 and operating device 116 to which the anomaly detection support device 120 is applied are not limited to these. For example, the control system 110 may be a chemical liquid discharge control system mounted on a substrate manufacturing apparatus, and the operating device 116 to be controlled may be a system including a pump, a control valve (an example of an operating element), piping, etc. The second embodiment will be described below, focusing on the differences from the first embodiment described above.

[0142] FIG. 13 is a first diagram showing an example of another control block. Specifically, FIG. 13 shows an example of a control block of a control system 110 (a chemical liquid discharge control system mounted on a substrate manufacturing apparatus) that performs discharge rate feedback control on an operating device 116 (a system including a pump, a control valve, piping, etc.) that is the control target. As indicated by reference numeral 1310, the control device 114 of the control system 110 has a difference calculation unit 911. The difference calculation unit 911 accepts input of target data for the discharge rate and also accepts input of measured data for the discharge rate. The difference calculation unit 911 also calculates the difference between the target data and the measured data, and inputs the difference to a control unit 912.

[0143] The control unit 912 generates an operation signal of an operation amount (valve opening degree) in accordance with the input difference, and inputs the signal to the regulator valve.

[0144] While the control system 110 is in operation, the control device 114 performs discharge rate feedback control on the operating device 116 under the control block indicated by the reference numeral 1310. Therefore, regardless of whether an abnormality has occurred, the measurement data follows the target data, and it is not easy to determine whether an abnormality has occurred based on the measurement data of the discharge rate while the control system 110 is in operation. Here, abnormalities in the liquid chemical discharge control system include, for example, an increase in the kinetic viscosity of the liquid chemical (a decrease in the discharge rate), etc.

[0145] In response to this, as in the first embodiment, the anomaly detection support device 120 identifies an operation method for the operating device 116 (here, a regulator valve) for making the occurrence of an abnormality apparent, and notifies the control device 114. As a result, the anomaly detection support device 120 can reliably determine whether or not an abnormality has occurred when the control system 110 is not operating. Note that, as in the first embodiment, the operation method notified by the anomaly detection support device 120 at this time includes items such as: a) a control method for the operating device 116 when not operating (whether or not to perform discharge rate feedback control), b) input data to be input to operate the operating device 116 when not operating (a waveform of target data when discharge rate feedback control is performed when not operating, and a waveform of an operation signal to operate the operating device 116 when discharge rate feedback control is not performed when not operating), and c) data to be used to calculate the second abnormality probability (an operation signal or measurement data when discharge rate feedback control is performed when not operating, and measurement data when discharge rate feedback control is not performed when not operating).

[0146] For example, suppose an increase in kinetic viscosity is identified as the cause of an abnormality that causes a change in the operating amount (valve opening) for maintaining the discharge rate of the operating device 116. In this case, the operating method can be appropriately determined, as in the first embodiment, by, for example, selecting data that exhibits the characteristics of the abnormality cause as item c), and selecting, as items a) and b), control methods and input data that can reveal the occurrence of an abnormality in the data selected in item c). Specifically, the abnormality detection support device 120 notifies the control device 114 of the operating method, which includes: performing discharge rate feedback control on the operating device 116 when the operating device 116 is not in operation; inputting target data for a predetermined discharge rate (e.g., a maximum discharge rate) as target data for the discharge rate feedback control; and using the measured data (discharge rate) to calculate the second abnormality probability. This can reveal the increase in the kinetic viscosity of the chemical liquid. In other words, the abnormality detection support device 120 can reliably determine whether the kinetic viscosity of the chemical liquid is increasing.

[0147] In this way, the abnormality detection support device 120 inputs a waveform that maintains the discharge rate at a high value as target data. As a result, even when the control device 114 performs discharge rate feedback control, the abnormality detection support device 120 can make it apparent in the control system 110 that the kinetic viscosity of the chemical liquid is increasing. As a result, the abnormality detection support device 120 can reliably determine whether the kinetic viscosity of the chemical liquid is increasing.

[0148] [Third Embodiment] In the above first and second embodiments, the control system 110 to which the anomaly detection support device 120 is applied is described as a control system that performs feedback control during operation. However, the control system 110 to which the anomaly detection support device 120 is applied is not limited to a control system that performs feedback control during operation, and may be a control system that does not perform feedback control during operation. For example, the control system 110 may be a plasma control system mounted on a substrate manufacturing apparatus, and the operating device 116 to be controlled may be a system that includes a microwave output device or the like.

[0149] In the first embodiment, when feedback control is not performed, a step-like waveform is exemplified as the waveform of the operation signal that operates the operating device 116 (FIG. 11). However, when feedback control is not performed, the waveform of the operation signal that operates the operating device 116 is not limited to a step-like waveform.

[0150] The third embodiment will be described below, focusing on the differences from the first and second embodiments.

[0151] Fig. 14 is a second diagram showing an example of another control block. Specifically, Fig. 14 shows an example of a control block of a control system 110 (a plasma control system mounted on a substrate manufacturing apparatus) that performs plasma control on an operating device 116 (a system including a microwave output device, etc.) to be controlled. As indicated by reference numeral 1410, the control device 114 of the control system 110 has a control unit 912 and inputs a microwave waveform as an operating signal to the microwave output device. The microwave output device inputs pulse-modulated microwaves into a chamber to perform plasma processing.

[0152] The control system 110 also has a high-frequency power supply and a matching unit, and during plasma processing, the RF signal generated by the high-frequency power supply undergoes impedance matching processing in the matching unit, and the matched RF signal is input to the chamber.

[0153] During operation of the control system 110, the control device 114 inputs microwaves of a predetermined frequency as an operating signal to the microwave output device. When plasma processing is performed repeatedly, deposits accumulate inside the chamber, and it is known that the signal strength (frequency characteristics) of the matched RF signal at each frequency changes due to the deposition.

[0154] However, when the control system 110 is operating, simply inputting microwaves of a predetermined frequency as an operating signal makes it difficult for the above-mentioned matched RF signal to reveal the state of adhesion of deposits inside the chamber.

[0155] Therefore, similarly to the first embodiment, the anomaly detection support device 120 notifies the control device 114 of an operation method of the operating device 116 (here, the microwave output device) for making the adhesion state of the deposit inside the chamber apparent. Specifically, the anomaly detection support device 120 notifies the control device 114 of the operation method, for example: - not performing feedback control on the operating device 116, - inputting the waveform of an operating signal (microwave) that sweeps a predetermined frequency band as the waveform of the operating signal (microwave) that operates the operating device 116, - using the matched RF signal to calculate the second anomaly probability.

[0156] This enables the abnormality detection support device 120 to reveal the adhesion status of deposits within the chamber while the control system 110 is not operating, thereby enabling the adhesion status of deposits within the chamber to be reliably determined.

[0157] Fourth Embodiment In the above embodiments, it has been described that it is determined that an abnormality may have occurred in the control system 110 when the first abnormality probability is equal to or greater than a predetermined threshold. However, the determination method is not limited to this, and, for example, it may be configured to determine whether an abnormality may have occurred by taking into account changes in the first abnormality probability over time. In this case, the anomaly detection support device 120 may display a graph showing changes in the first abnormality probability over time (a graph in which the horizontal axis represents time and the vertical axis represents the value of the first abnormality probability).

[0158] Furthermore, in the above embodiments, the second abnormality probability calculation unit 404 has been described as calculating the second abnormality probability based on either or both of the operation signal acquired from the control device 114 and the measurement data acquired from the measurement data acquisition device 112. However, the data used to calculate the second abnormality probability is not limited to the operation signal or the measurement data, and may be data calculated based on the operation signal or data calculated based on the measurement data. Data calculated based on the operation signal or data calculated based on the measurement data refers to, for example, a differential signal from the operation signal when no abnormality occurs or differential data from the measurement data when no abnormality occurs. Note that the operation signal when no abnormality occurs or the measurement data when no abnormality occurs may be stored in advance in, for example, the list 500.

[0159] Furthermore, in each of the above embodiments, when it is determined in the monitoring process that an abnormality may occur, the abnormality detection support process is performed when the control system is not in operation immediately after the abnormality detection support process. However, the timing of performing the abnormality detection support process is not limited to this. Furthermore, whether or not to perform the abnormality detection support process may be determined based on instructions from an operator. Furthermore, when it is determined that an abnormality may occur for multiple abnormality causes, the abnormality detection support process may be performed for some of the abnormality causes when the control system 110 is not in operation. For which abnormality cause the abnormality detection support process is performed may be determined based on instructions from an operator.

[0160] In the above embodiments, the abnormality detection support process is described as transitioning to a maintenance period when it is determined that an abnormality has occurred. However, the timing of transitioning to a maintenance period is not limited to this. Furthermore, whether or not to transition to a maintenance period may be determined by an operator based on the determined cause of the abnormality.

[0161] Although the above embodiments do not mention the details of maintenance, for example, if the gears and belts deteriorate or if an abnormality such as a broken belt occurs, new gears and belts are procured and replaced. Also, if an abnormality such as the grease running out in the bearings or a grease abnormality occurs, new grease is procured and the bearings are greased up.

[0162] The present invention is not limited to the configurations described in the above embodiments, but may be combined with other elements, etc. These aspects can be changed without departing from the spirit of the present invention, and can be appropriately determined depending on the application form.

[0163] This application claims priority based on Japanese Patent Application No. 2024-058759, filed on April 1, 2024, the entire contents of which are incorporated herein by reference.

[0164] 110: Control system 111: Anomaly detection device 113, 115: Measuring device 114: Control device 116: Operating device 120: Anomaly detection support device 401: First anomaly probability monitoring unit 402: Anomaly cause identification unit 403: Operating method output unit 404: Second anomaly probability calculation unit 405: Display unit 500: List 911: Difference calculation unit 912: Control unit

Claims

1. An anomaly detection support device having: a storage unit that stores, for each cause of an anomaly in a system including a device to be controlled, waveform data at the time of an anomaly and an operating method of the device corresponding to the waveform data at the time of an anomaly, in association with each other; an identification unit that, when the probability of an anomaly calculated by a first anomaly detection algorithm based on measurement data measured while the system is in operation becomes equal to or greater than a predetermined threshold, acquires the measurement data when the calculated probability of an anomaly equals or exceeds the predetermined threshold, and identifies the cause of the anomaly; and an output unit that outputs the identified cause of the anomaly.

2. The anomaly detection support device according to claim 1, wherein the output unit outputs the identified cause of the anomaly in association with an operating method of the device corresponding to the waveform data at the time of the anomaly.

3. The anomaly detection support device according to claim 1, wherein the method of operating the device corresponding to the waveform data at the time of the anomaly is a method of operating the device when the system is not in operation.

4. The anomaly detection support device according to claim 3, further comprising a calculation unit that calculates an anomaly probability using a second anomaly detection algorithm based on measurement data measured by operating the device to be controlled using an operating method of the device corresponding to the waveform data at the time of the anomaly.

5. The anomaly detection support device according to claim 3, further comprising a calculation unit that calculates an anomaly probability using a second anomaly detection algorithm based on an operation signal when the controlled device is operated using an operation method for the device corresponding to the waveform data at the time of the anomaly.

6. The anomaly detection support device according to claim 1, wherein the identification unit identifies the cause of the anomaly by calculating the similarity between the waveform data stored in the storage unit and the acquired measurement data.

7. The anomaly detection support device according to claim 3, wherein the operating method of the device corresponding to the waveform data at the time of the anomaly is different from the operating method of the device while the system is running.

8. The anomaly detection support device according to claim 7, wherein the method of operating the device corresponding to the waveform data at the time of the anomaly includes operating the device using a control method different from the control method of the device while the system is running.

9. The anomaly detection support device according to claim 7, wherein the method of operating the device corresponding to the waveform data at the time of the anomaly includes controlling the device using target data having a waveform different from target data for controlling the device while the system is running, or operating the device using an operating signal having a waveform different from an operating signal for operating the device while the system is running.

10. The abnormality detection support device according to claim 9, wherein the storage unit stores an operating method for operating the device to be controlled by an operating signal in a band including the vibration frequency in association with the cause of the abnormality corresponding to the waveform data at the time of the abnormality.

11. The anomaly detection support device according to claim 9, wherein the method of operating the device corresponding to the waveform data at the time of the anomaly includes performing feedback control on the device to be controlled using target data of a waveform different from the target data when performing feedback control on the device to be controlled while the system is in operation.

12. The abnormality detection support device according to claim 11, wherein the storage unit stores an operating method for performing feedback control of the controlled device using target data for the operating speed of the controlled device in a ramp-shaped waveform as input in response to an abnormal cause that causes a change in the torque for maintaining the operating speed of the controlled device, in association with the abnormal cause.

13. The abnormality detection support device according to claim 11, wherein the storage unit stores an operating method for performing feedback control of the controlled device using target data for the operating speed of a parabolic waveform as input in association with an abnormal cause that causes a change in the operating acceleration of the controlled device.

14. The abnormality detection support device described in claim 11, wherein the storage unit stores an operating method for performing feedback control of the controlled device using target data for a predetermined discharge amount as input in response to an abnormal cause that causes a change in the operating amount of the operating terminal for maintaining the discharge amount of the controlled device, in association with the abnormal cause.

15. The abnormality detection support device according to claim 9, wherein the storage unit stores an operating method for operating the controlled device by an operating signal that sweeps a predetermined frequency band in association with an abnormality cause that causes a change in frequency characteristics within a chamber in which plasma processing is performed by the controlled device.

16. An anomaly detection support method performed by a computer, comprising the steps of: storing, for each cause of an anomaly in a storage unit, waveform data at the time of an anomaly and an operation method of the device corresponding to the waveform data at the time of an anomaly; acquiring the measurement data when the anomaly probability calculated by a first anomaly detection algorithm based on measurement data measured while the system is operating, exceeds a predetermined threshold, and identifying the cause of the anomaly; and outputting the identified cause of the anomaly.

17. An anomaly detection support program that causes a computer to execute the following steps for a system including a device to be controlled: storing, for each cause of an anomaly, waveform data at the time of an anomaly and the device operating method corresponding to the waveform data at the time of an anomaly in a storage unit in association with each other; acquiring the measurement data when the anomaly probability calculated by a first anomaly detection algorithm based on measurement data measured while the system is operating, exceeds a predetermined threshold, and identifying the cause of the anomaly; and outputting the identified cause of the anomaly.

Images