JP7263913B2 - Anomaly detection device and anomaly detection method - Google Patents

Description

The present invention relates to an anomaly detection device and an anomaly detection method for detecting an anomaly in a device in which periodic disturbance occurs.

2. Description of the Related Art Devices for detecting abnormalities such as failures and malfunctions occurring in devices are used in various fields. By detecting an abnormality, it becomes possible to repair, replace, or troubleshoot the device at an early stage, and it is possible to realize normal operation of the device. For example, if it detects an abnormality in a manufacturing apparatus, it is possible to prevent the manufacturing of defective products and the decrease in production efficiency.

By the way, depending on the device, even if the device is operated stably, there are some devices in which the values measured by the sensor periodically fluctuate. For example, in a heat treatment apparatus in which a plurality of objects to be treated are successively introduced and subjected to heat treatment, when an object to be treated at room temperature is introduced into the heat treatment apparatus being heated, the temperature inside the heat treatment apparatus decreases. In such a situation, when trying to detect an abnormality by measuring the temperature in the heat treatment apparatus with a temperature sensor, a disturbance occurs when the object to be treated is periodically introduced into the heat treatment apparatus and the temperature drops. In other words, in a heat treatment apparatus, periodic disturbance may occur that is unrelated to the abnormality of the apparatus itself.

On the other hand, it is difficult to distinguish such periodic disturbances from anomalies occurring in the device. Therefore, there is a possibility that such a periodic disturbance may be erroneously detected as an abnormality by a conventional abnormality detection device, making accurate abnormality detection difficult.

For example, Patent Literature 1 describes a technique for easily obtaining useful information for temperature measurement and control. Using the information obtained by using the method described in Patent Document 1, it is also possible to detect abnormalities such as failures and malfunctions occurring in the apparatus.

Japanese Patent No. 4052089

However, even if the information obtained by using the method described in Patent Document 1 is used, it is difficult to distinguish periodic disturbances occurring in the device from abnormalities and to detect periodic disturbances unrelated to abnormalities. , there was a risk of erroneous detection of abnormalities.

SUMMARY OF THE INVENTION It is an object of the present invention to provide an anomaly detection device and an anomaly detection method for detecting an anomaly without erroneously detecting an anomaly in a device in which periodic disturbance occurs.

An anomaly detection device according to the present disclosure is an anomaly detection device that detects an anomaly in a system in which periodic disturbance occurs, and is an average value of the operation amount that operates the system for a predetermined period determined in consideration of the periodic disturbance. An average calculation unit that calculates an amount average value and a sensor average value that is an average value of measured values for a predetermined period obtained by a sensor provided in the system by manipulating the operation amount, and divides the sensor average value by the operation amount average value. A steady-state gain calculation unit that obtains the obtained value as a steady-state gain to be used for estimating an abnormality, and a detection unit that detects an abnormality by comparing the steady-state gain with a predetermined threshold value.

ADVANTAGE OF THE INVENTION The abnormality detection apparatus and the abnormality detection method which concern on this invention can detect abnormality, without erroneously detecting periodic disturbance as abnormality.

1 is a block diagram illustrating an abnormality detection device according to Embodiment 1; FIG. 4 is a flowchart for explaining an abnormality detection method according to Embodiment 1; 9 is a flowchart for explaining an abnormality detection method according to Embodiment 2; FIG. 4 is a flowchart for explaining abnormality detection processing in the flowchart of FIG. 3; FIG. 7 is a graph showing an example of the passage of time (with disturbance) of target values and measured values; 7 is a graph showing an example of the time course of the operation amount and disturbance (with disturbance); It is an example of the gain (with disturbance) obtained from the case of the manipulated variable and the measured value. It is an example of a steady-state gain (with disturbance) obtained from the manipulated variable average value and the temperature average value. 7 is a graph showing an example of the passage of time (without disturbance) of target values and measured values; FIG. 11 is a graph showing an example of time course of operation amount and disturbance (no disturbance); FIG. It is an example of the gain (no disturbance) obtained from the case of the manipulated variable and the measured value. It is an example of a steady-state gain (no disturbance) obtained from the manipulated variable average value and the temperature average value. FIG. 7 is a block diagram illustrating an abnormality detection device according to Modification 1; 9 is a flowchart for explaining an abnormality detection method according to Modification 1; FIG. 11 is a block diagram illustrating an abnormality detection device according to Modification 2;

Hereinafter, embodiments of the present invention will be described with reference to the drawings. An anomaly detection device according to the present invention detects an anomaly of a device without erroneously detecting periodic disturbance as an anomaly in a device that generates periodic disturbance. In addition, in the following description, the same reference numerals are given to the same configurations, and the description thereof is omitted.

Here, "periodic disturbance" is a signal, load, noise, etc., generated from the outside in a device, etc., that is subject to abnormality detection by the abnormality detection device, and it is not related to the presence or absence of abnormality, and it occurs even in normal conditions. be.

In the following embodiments, an abnormality detection device will be described using an example of detecting an abnormality in a heat treatment device that performs heat treatment on objects to be processed that are periodically introduced. Further, the abnormality detection device according to the embodiment will be described using a specific example of preventing erroneous detection without detecting periodic disturbance due to introduction of a workpiece into the heat treatment apparatus as an abnormality.

Here, the "target value SP" is assumed to be the target temperature when heating for heat treatment in the heat treatment apparatus.
Also, the "manipulated variable MV" is an index representing an output amount or the like for manipulating the heating amount of the heater in the heat treatment apparatus.
Also, the "measured value PV" is assumed to be the temperature actually measured by the sensor as a result of the operation of the manipulated variable MV in the heat treatment apparatus.

[Embodiment 1]
As shown in FIG. 1, an abnormality detection apparatus 1 according to Embodiment 1 is connected to a heat treatment apparatus 2 to be subjected to abnormality detection.

In the heat treatment apparatus 2, a workpiece W, which is an object to be processed, is put into a chamber (not shown), and the workpiece W is subjected to heat treatment. For example, the work W is a semiconductor wafer, and the heat treatment device 2 is a device for processing the wafer installed in a chamber.
As shown in FIG. 1, the heat treatment apparatus 2 includes a temperature control section 21, a heater 22, and a temperature sensor 23. As shown in FIG.

The temperature control unit 21 controls the heater 22 to bring the inside of the chamber to the set target value SP. Specifically, the temperature control unit 21 outputs a signal specified by the manipulated variable MV as a signal for controlling the heater 22 . Here, the target value SP may be input according to the target work W, or may be registered in a storage unit (not shown) in the heat treatment apparatus 2 in advance.

The heater 22 is heated according to the signal input from the temperature control unit 21, and adjusts the workpiece W to the temperature specified by the target value SP.

The temperature sensor 23 measures the temperature of the workpiece, which is the object to be processed, placed in the chamber. Here, the temperature measured by the temperature sensor 23 is the measured value PV. Also, the measured value PV measured by the temperature sensor 23 is output to the temperature controller 21 .
The temperature controller 21 controls the inside of the chamber to the target value SP by performing feedback control using the measured value PV.

<Abnormality detection device>
As shown in FIG. 1 , the anomaly detection device 1 can be implemented by an information processing device including a control unit 11 , a storage unit 12 and a communication interface (I/F) 13 . Although illustration is omitted, the abnormality detection device 1 may include an input unit used for inputting information and an output unit used for outputting information.

The control unit 11 is a controller such as a CPU that controls the entire abnormality detection device 1 . For example, by reading and executing a program (not shown) stored in the storage unit 12, the control unit 11 operates as shown in FIG. The processing as the gain calculation unit 114, the detection unit 115, and the output processing unit 116 can be executed.

The storage unit 12 is storage means such as RAM and ROM for storing necessary data and programs. The storage unit 12 stores, for example, operation amount data 121, measured value data 122, average value data 123, and setting data 124 by executing processing. Although not shown, the storage unit 12 may store data necessary for abnormality detection in advance.

The communication I/F 12 is means for realizing data transmission/reception with an external device. The communication method used by this communication I/F 12 is not limited.

The setting unit 111 sets various conditions for abnormality detection when the heat treatment apparatus 2 starts operating. The setting unit 111 sets, for example, the time (processing time for one workpiece) required from the loading of the workpiece W corresponding to the cycle of the periodic disturbance to the loading of the next workpiece W, the number of workpieces W to be processed by the heat treatment apparatus 2, the number of abnormalities Set values such as thresholds used for judgment. Also, the setting unit 111 causes the storage unit 12 to store these values as the setting data 124 .

The storage processing unit 112 acquires values such as the operation amount MV used for controlling the heater 22 and the measured value PV measured by the temperature sensor 23 from the heat treatment apparatus 2 and stores them in the storage unit 12 . For example, the storage unit 12 associates the operation amount MV with the time and stores the operation amount data 121 in the storage unit 12 . Further, the storage unit 12 associates the measured value PV with the time and stores it in the storage unit 12 as the measured value data 122 .

The average calculation unit 113 calculates the average value of the manipulated variable MV for manipulating the temperature of the heat treatment apparatus 2 and the average value of the measured values PV obtained by measuring the temperature obtained in the heat treatment apparatus 2 by manipulating the manipulated variable. Calculate the temperature average value that is

Here, the average calculator 113 obtains an average value of the manipulated variable MV and the measured value PV for a predetermined period determined according to the disturbance cycle. This predetermined period is preferably longer than the disturbance period. If the period is more than twice the disturbance period, it is difficult to detect, so the period is preferably less than or equal to twice the disturbance period. By obtaining the average value using the data of the period determined according to the disturbance period in this manner, the influence of the period disturbance can be eliminated, and the frequency of occurrence of erroneous detection can be reduced.

Specifically, the average calculation unit 113 reads out the operation amount data 121 of the storage unit 12 and obtains the average value of the operation amount MV for a predetermined period as the operation amount average value MVave. Further, the average calculation unit 113 reads out the measured value data 122 of the storage unit 12 and obtains the average value of the measured values PV for a predetermined period as the average temperature value PVave. Furthermore, the average calculation unit 113 outputs the calculated operation amount average value MVave and temperature average value PVave to the steady gain calculation unit 114 as average value data 123 . Also, the average calculation unit 113 may cause the storage unit 12 to store the average value data 123 .

For example, the average calculator 113 calculates the average value at the timing when the amount of change in the measured values becomes equal to or greater than a predetermined amount of change. Specifically, when periodic disturbance occurs, the measured values often change. Therefore, when the change in the measured value of the temperature sensor 23 during a certain sampling period after the start of the measurement object becomes larger than a predetermined value, the average value is calculated from the manipulated variable data 121 and the measured value data 122 in the subsequent predetermined period. calculate. When the change in the measured value is large is when the amount of change in the predetermined period is greater than or equal to the set value (for example, when there is a change of 1° C. or more in 5 seconds). As a result, the average value can be obtained at the timing when the period disturbance is likely to occur without the need for a trigger signal from the outside, and when an abnormality occurs, the abnormality can be detected.

Also, the average calculation unit 113 calculates an average value at a predetermined timing, for example. Specifically, the average value is calculated from the manipulated variable data 121 and the measured value data 122 during the processing period of the previous workpiece W at the timing when a predetermined period has passed since the new workpiece W was loaded. As a result, the average value can be obtained at the timing of the end of the processing period for each work, and an abnormality can be detected when an abnormality occurs. It should be noted that the predetermined period includes the timing immediately after the injection.

Steady-state gain calculation section 114 obtains a value obtained by dividing temperature average value PVave obtained by average calculation section 113 by manipulated variable average value MVave as a steady-state gain K (=PVave/MVave) used for estimating abnormality.

Steady-state gain calculator 114 may use temperature average value PVave obtained by subtracting predetermined offset value PV0 to obtain steady-state gain K (=(PVave-PV0)/MVave). Here, the "offset value PV0" is a value that identifies the influence of the surrounding environment on the measured value PV, and is excluded from the measured value to enable accurate abnormality detection.

For example, the steady-state gain calculator 114 can use an initial value measured by the temperature sensor 23 before the heater 22 is controlled by the temperature controller 21 as the offset value PV0. As a result, the ambient temperature of the workpiece W, that is, the temperature inside the chamber, can be assumed as the initial value, and the steady-state gain K excluding the influence thereof can be obtained. In this case, for example, the setting unit 111 acquires the initial values in advance and stores them in the storage unit 12 as the setting data 124 . Also, the steady-state gain calculator 114 uses the initial value of the setting data 124 as the offset value PV0.

In addition, steady-state gain calculation section 114 may use a value set by the user as offset value PV0. As a result, the steady-state gain K can be obtained using a value set by the user in consideration of the characteristics of the heat treatment apparatus 2 . In this case, for example, the setting unit 111 stores the offset value PV0 set by the user as the setting data 124 in the storage unit 12 .

A detection unit 115 detects an abnormality by comparing the steady-state gain K obtained by the steady-state gain calculation unit 114 with a predetermined threshold value. For example, the detection unit 115 determines that the heat treatment apparatus 2 is normal when the steady-state gain K is within the range of the upper threshold value and the lower threshold value stored in the storage unit 12 . Further, the detection unit 115 determines that the heat treatment apparatus 2 is abnormal when the steady-state gain K is larger than the upper limit threshold or smaller than the lower limit threshold.

The steady-state gain K has been averaged over the effects of periodic disturbances. Also, the upper and lower thresholds are set in consideration of periodic disturbances. Therefore, the detection unit 115 can prevent erroneous detection of periodic disturbance as abnormal.

Here, the detection unit 115 can use a threshold selected from a plurality of thresholds preset in the storage unit 12 according to the type and state of the disturbance. For example, the threshold used for a predetermined period immediately after the occurrence of the disturbance can be different from the threshold used after the predetermined period has passed since the occurrence of the disturbance. Further, for example, different threshold values can be used depending on the type of work W to be thrown. In this way, by using an optimum threshold among a plurality of thresholds set according to the type and state of the disturbance, erroneous detection of abnormality can be reduced.

The output processing unit 116 outputs the abnormality detection result by the detection unit 115 to an external output device. This output device is a device that outputs an abnormal signal, and for example, a display that displays an abnormal signal with text or an image, a speaker that outputs an abnormal signal with a beep sound, or a notification of an abnormal signal by lighting or blinking of light. lamps and the like.

As described above, the abnormality detection device 1 according to Embodiment 1 can prevent erroneous detection due to the influence of periodic disturbance when detecting an abnormality in a device in which periodic disturbance occurs.

<Abnormality detection method>
An abnormality detection method in the abnormality detection device 1 will be described with reference to the flowchart shown in FIG.

When the heat treatment apparatus 2 starts operating, the setting unit 111 performs initial setting (S101).
For example, the setting unit 111 sets the following values as initial settings.
"Disturbance period Pd": Interval from the input of the previous work to the input of the next work (for example, 100 (seconds))
"Number of final heat treatments n": Number of heat treatments finally performed in the heat treatment apparatus 2 (for example, 200 (times))
"Number of times of heat treatment i": Number of times of heat treatment that has been started so far in heat treatment apparatus 2 (initial value 0 (times) is set)
"Upper limit threshold Kmax": Upper limit value compared with steady gain K in abnormality detection "Lower limit threshold Kmin": Lower limit value compared with steady gain K in abnormality detection

When a signal indicating that the workpiece W is loaded into the heat treatment apparatus 2 is input (YES in S102), the setting unit 111 sets the timer t (S103). The timer t measures the elapsed time after the workpiece W is loaded. Specifically, the setting unit 111 resets the value of the timer t (sets it to 0 (seconds)), and then starts counting the timer t.

Also, the setting unit 111 increments the heat treatment number i (S104). This makes it possible to identify the number of workpieces W that have been processed so far.

The storage processing unit 112 starts acquiring the manipulated variable MV and the measured value PV from the heat treatment apparatus 2 (S105). The storage processing unit 112 stores the acquired operation amount MV in the storage unit 12 as the operation amount data 121 . The storage processing unit 112 also stores the acquired measurement value PV in the storage unit 12 as the measurement value data 122 .

When the setting unit 111 determines that the value of the timer t is equal to or greater than the disturbance period Pd (YES in S106), the abnormality detection device 1 executes the abnormality detection process (broken line part in the flowchart of FIG. 2). Specifically, the average calculation unit 113 calculates the operation amount average value MVave, which is the average value of the operation amount MV during the period from 0 to Pd of the timer t, and calculates the average value of the measurement values PV during the period. A certain temperature average value PVave is calculated (S107).

The steady-state gain calculator 114 calculates a steady-state gain K using the manipulated variable average value MVave and the temperature average value PVave obtained in step S107 (S108).

The detection unit 115 compares the steady-state gain K obtained in step S108 with the upper threshold value Kmax, and determines that there is an abnormality when the steady-state gain K is equal to or greater than the upper threshold value Kmax (S109).
Further, when the steady-state gain K is equal to or less than the upper threshold value Kmax (NO in S109), the detection unit 115 compares the steady-state gain K with the lower limit threshold value Kmin. S110).

When the detection unit 115 detects an abnormality (YES in S109 or YES in S110), the output processing unit 116 outputs an abnormality signal (S111).

The setting unit 111 determines whether or not the number of heat treatments i has reached the final number of heat treatments n. That is, when the number of heat treatments i has not reached the final number of heat treatments n, the processing in the heat treatment device 2 is not completed, and the abnormality detection device 1 continues to determine abnormality.

On the other hand, when the number of heat treatments i reaches the final number of heat treatments n, the abnormality detection device 1 terminates the abnormality determination process (NO in S112). That is, since all the workpieces W that need to be treated have been heat-treated by the heat treatment device 2, the abnormality detection device 1 does not need to continue the determination of abnormality.

In the example described above with reference to FIG. 2, the average value of the manipulated variable MV and the measured value PV obtained during the period of the disturbance period Pd after the wafer is loaded is calculated. However, it is not limited to this. For example, as described above, the average value may be obtained at the timing when the measured value PV reaches a predetermined amount of change. Alternatively, an average value may be obtained from the manipulated variable MV and the measured value PV for a period longer than the disturbance period Pd.

As described above, the abnormality detection device 1 according to Embodiment 1 can prevent erroneous detection due to the influence of periodic disturbance when detecting an abnormality in a device in which periodic disturbance occurs.

[Embodiment 2]
Since the abnormality detection device according to the second embodiment has the same configuration as the abnormality detection device 1 described above with reference to FIG. 1, it will be described with reference to FIG. In the anomaly detection device 1 according to the second embodiment, the average operation unit 113 does not calculate the operation amount average value MVave and the temperature average value PVave every time the disturbance period Pd elapses, but always calculates the operation amount average value. This differs from the first embodiment in that a moving average is calculated as the value MVave and the average temperature value PVave. Specifically, when acquiring the manipulated variable MV and the measured value PV in the sampling cycle, the average calculator 113 obtains the manipulated variable average value MVave and the temperature average value PVave in the sampling cycle. Therefore, the steady-state gain calculation unit 114 also calculates the steady-state gain K continuously. Further, the detection unit 115 also always performs abnormality detection.

<Abnormality detection method>
An abnormality detection method in the abnormality detection device 1 will be described with reference to the flowcharts shown in FIGS. 3 and 4. FIG. Since the flowchart of FIG. 3 is the same as the flowchart described above with reference to FIG.

When the setting unit 111 determines that the value of the timer t is equal to or greater than the disturbance period Pd, the abnormality detection device 1 executes abnormality detection processing (S200). Here, the abnormality detection device 1 continuously executes the abnormality detection process using the operation amount MV and the measured value PV newly acquired in step S105. The processing of step S200 will be described using the flowchart shown in FIG. Here, the reason why the abnormality detection process is started when the value of the timer t becomes equal to or greater than the disturbance period Pd is that until then, the operation amount MV and the measurement value PV sufficient to obtain the average value are acquired. This is because they are not.

As shown in FIG. 4, in the abnormality detection process, the average calculation unit 113 obtains the average value of the operation amount MV for a predetermined period stored in the storage unit 12 as the operation amount average value MVave, and calculates the average operation amount MVave. An average value is obtained as a temperature average value PVave (S201). Here, an average value is calculated continuously. Therefore, the average value calculated by the average calculation unit 113 is a moving average.

The steady-state gain calculator 114 calculates a steady-state gain K using the manipulated variable average value MVave (moving average) and the temperature average value PVave (moving average) obtained in step S201 (S202).

The detection unit 115 compares the steady-state gain K obtained in step S202 with the upper limit threshold value Kmax, and determines that there is an abnormality when the steady-state gain K is equal to or greater than the upper limit threshold value Kmax (S203).
Further, when the steady-state gain K is equal to or less than the upper threshold value Kmax (NO in S203), the detection unit 115 compares the steady-state gain K with the lower limit threshold value Kmin. ).

When the detection unit 115 detects an abnormality (YES in S203 or YES in S204), the output processing unit 116 outputs an abnormality signal and terminates the abnormality detection process (S205).

If no abnormality is detected (NO in S204), the abnormality detection process is terminated without outputting an abnormality signal.
Further, when the abnormality detection process is completed, the process proceeds to step S113 shown in FIG.

Specifically, the abnormality detection process shown in FIG. 4 is executed once per sampling cycle. As a result, the average calculator 113 obtains the moving average as described above.

The setting unit 111 determines whether or not the termination condition is met (S113). Here, the termination condition is, for example, the case where the number of times of heat treatment i reaches the final number of times of heat treatment n and the disturbance period Pd has elapsed after the n-th workpiece W is introduced.

If the termination condition is not met (NO in S113), the timer t is reset, the number of heat treatments i is incremented, and the process returns to step S201 to repeat the process of S201 (S114).

On the other hand, if the termination condition is satisfied, the abnormality detection device 1 terminates the abnormality determination process (YES in S113). That is, the abnormality detection device 1 continues the abnormality detection process until the heat treatment number i reaches the final heat treatment number n and the processing in the heat treatment device 2 is completed.

Although not shown in FIG. 3, the setting unit 111 increments the number of times of heat treatment i each time it acquires a signal indicating that the workpiece W has been loaded, thereby continuing the process until the heat treatment in the heat treatment apparatus 2 is completed. be done.

As described above, the abnormality detection device 1 according to the second embodiment can prevent erroneous detection due to the influence of periodic disturbance when detecting an abnormality in a device in which periodic disturbance occurs.

<<Comparative example>>
Using the graphs of FIGS. 5A to 6D, the case where the gain K' is obtained using the manipulated variable MV and the measured value PV, and the manipulated variable average value MVave (moving average) and the temperature average value PVave (moving average) are used. A case where the steady-state gain K is obtained by

(In case of periodic disturbance)
FIG. 5A is a graph showing the passage of time between the target value SP and the measured value PV measured by the temperature sensor 23, where the horizontal axis is time (seconds) and the vertical axis is temperature (° C.). As shown in FIG. 5A, with respect to the target value SP, the temperature rises for a certain period from the start of heating by the heater 22, and the measured value PV is not stable. After that, after a certain amount of time has passed, the measured value PV stabilizes near the target value SP.

FIG. 5B is a graph showing the time course of the operation amount MV and the disturbance when the horizontal axis is time (seconds) and the vertical axis is the operation amount (%). FIG. 5B shows an example of the manipulated variable MV and disturbance when the measured value PV as described above in FIG. 5A is obtained. In FIG. 5B, the disturbance is shown as a value converted to the manipulated variable MV. The “value converted to the manipulated variable” is a value that excludes the influence of disturbance on the manipulated variable MV. In FIG. 5B, the manipulated variable MV gradually decreases until the temperature rises and the measured value PV stabilizes, as described above with reference to FIG. 5A. Then, when the measured value PV is stabilized, in the heat treatment apparatus 2, since the workpieces W are successively loaded and heat-treated, disturbance occurs at the timing of loading the workpieces W. As shown in FIG. 5B, this disturbance is a periodic disturbance. Specifically, the occurrence of this disturbance is an increase in temperature. Further, as shown in FIG. 5B, the manipulated variable MV also fluctuates according to the occurrence of periodic disturbance.

FIG. 5C is an example of the gain K' (=PV/MV) obtained in the case of the manipulated variable MV and the measured value PV shown in FIGS. 5A and 5B. In FIG. 5C, the horizontal axis is time (seconds) and the vertical axis is the value of the gain K'. FIG. 5C also shows an upper limit threshold Kmax and a lower limit threshold Kmin to be compared with the gain K'. As shown in FIG. 5B, the manipulated variable MV is affected by the periodic disturbance. Therefore, as shown in FIG. 5C, the manipulated variable MV affected by the periodic disturbance and the gain K′ obtained from the measured value PV are greatly affected by the periodic disturbance, and the range of the upper threshold value Kmax and the lower threshold value Kmin , resulting in erroneous detection of anomalies.

FIG. 5D shows a steady-state gain K (= PV /MV) is an example. In FIG. 5D, the horizontal axis is time (seconds), and the vertical axis is the steady-state gain K value. As shown in FIG. 5B, the manipulated variable MV is affected by the periodic disturbance. However, the steady-state gain K is obtained by averaging the manipulated variable average value MVave and the temperature average value PVave in which the influence of the periodic disturbance is mitigated. Therefore, as shown in FIG. 5D, the gain K′ obtained by the manipulated variable average value MVave and the temperature average value PVave is within the range of the upper threshold value Kmax and the lower threshold value Kmin as shown in FIG. No detection occurs.

5C and 5D show the steady-state gain K obtained even when the steady-state gain calculation section 114 uses the moving average obtained as the average value in the average calculation section 113. FIG. Although illustration is omitted, in the case of the example described above in Embodiment 1, the steady-state gain calculator 114 calculates the steady-state gain K will ask for Therefore, when using the average value, a value plotted at each disturbance period Pd is obtained as the gain K instead of a continuous value as shown in FIGS. 5C and 5D.

(Without periodic disturbance)
Similar to FIG. 5A, FIG. 6A shows the passage of time between the target value SP and the measured value PV measured by the temperature sensor 23 when the horizontal axis is time (seconds) and the vertical axis is temperature (° C.). It is a graph showing.

FIG. 6B is a graph showing the time course of the operation amount MV and the disturbance when the horizontal axis is time (seconds) and the vertical axis is the operation amount (%). FIG. 6B shows an example of the manipulated variable MV and disturbance when the measured value PV as described above with reference to FIG. 6A is obtained. FIG. 6B is an example with no disturbance. Therefore, as shown in FIG. 6B, the manipulated variable MV is stable.

FIG. 6C is an example of the gain K' (=PV/MV) obtained in the case of the manipulated variable MV and the measured value PV shown in FIGS. 6A and 6B. In FIG. 6C, the horizontal axis is time (seconds) and the vertical axis is the value of the gain K'. As shown in FIG. 6B, since the manipulated variable MV is not affected by the disturbance, the gain K′ obtained by the manipulated variable MV and the measured value PV is, as shown in FIG. Desired.

FIG. 6D shows a steady-state gain K (= PV /MV) is an example. In FIG. 5D, the horizontal axis is time (seconds), and the vertical axis is the steady-state gain K value. As shown in FIG. 6B, the manipulated variable MV is not affected by disturbance, and a steady-state gain K as shown in FIG. 6D is obtained.

[Modification 1]
FIG. 7 shows an example of an abnormality detection device 1A according to Modification 1. As shown in FIG.
The abnormality detection device 1A differs from the abnormality detection device 1 described above with reference to FIG. Further, the heat treatment apparatus 2A, which is the abnormality detection target of the abnormality detection apparatus 1A, differs from the heat treatment apparatus 2A described above with reference to FIG. (not shown) has an ambient temperature sensor 24 for measuring the temperature of the ambient environment, and is different in that the influence of the ambient temperature is corrected.

In the abnormality detection device 1A, the steady-state gain calculator 114 uses the offset value based on the measured value of the temperature sensor that measures the ambient temperature. As a result, the steady-state gain K excluding the influence of the ambient temperature can be obtained.

Specifically, when the storage processing unit 112 of the abnormality detection device 1A acquires the manipulated variable MV and the measured value PV and stores them in the storage unit, it also acquires the measured value of the ambient temperature sensor 24 and sets it as the offset value PV0, It is stored as offset value data 125 together with the time.

When calculating the average operation amount value MVave and the average temperature value PVave, the average calculation unit 113 calculates the offset average value PV0ave using the offset value PV0 in the same period.

When calculating the steady-state gain K, the steady-state gain calculation unit 114 calculates the steady-state gain K (=(PVave-PV0ave) by dividing the value obtained by subtracting the temperature mean value PVave from the offset mean value PV0ave by the manipulated variable mean value MVave. /MVave). Other configurations are the same as those of the abnormality detection device 1 described above with reference to FIG.

<Abnormality detection method>
An abnormality detection method in the abnormality detection device 1A will be described with reference to the flowchart shown in FIG. The same reference numerals are assigned to the same processes as in the flowchart described above with reference to FIG. 2, and the description thereof is omitted.

When starting to acquire the manipulated variable MV and the measured value PV, the storage processing unit 112 also starts acquiring the measured value of the ambient temperature sensor 24 as the offset value PV0 (S305). The storage processing unit 112 also stores the acquired offset value PV0 as the offset value data 125 in the storage unit 12 .

When calculating the manipulated variable average value MVave and the temperature average value PVave, the average calculator 113 also calculates an offset average value PV0ave that is the average value of the offset values PV0 (S307).

Further, the steady-state gain calculation unit 114 calculates the steady-state gain K using the offset average value PV0ave (S308). Specifically, a value obtained by subtracting the temperature average value PVave from the offset average value PV0ave is divided by the manipulated variable average value MVave to obtain the steady-state gain K (=(PVave−PV0ave)/MVave).

As described above, in the anomaly detection device 1A according to Modification 1, when detecting an anomaly in a device in which periodic disturbance occurs, if the ambient temperature is not stable, the effect of the ambient temperature is corrected, and then the period It is possible to prevent erroneous detection due to the influence of disturbance.
If the environmental temperature or the like is stable, the ambient temperature sensor 24 and its setting function can be omitted by obtaining the steady-state gain K without using the offset value.

[Modification 2]
FIG. 9 is an example of an abnormality detection device 1B according to Modification 2. As shown in FIG. The abnormality detection device 1B differs from the abnormality detection device 1 described above with reference to FIG. Therefore, in the abnormality detection device 1B, instead of obtaining the average of the operation amount MV and the measured value PV, the steady-state gain calculation unit 114 calculates the steady-state gain K (=PV/MV) from the operation amount MV and the measured value PV in advance. . Further, the average calculation unit 113 obtains the average of the steady-state gain K obtained by the steady-state gain calculation unit 114 for a predetermined period as the gain average value Kave. Other configurations and processes are the same as those of the anomaly detection device 1 described above with reference to FIG. 1, so description thereof is omitted.

The abnormality detection device 1B according to Modification 2 can also prevent erroneous detection due to the influence of periodic disturbance when detecting an abnormality in a device in which periodic disturbance occurs.

<others>
In each of the above embodiments and modifications, an example of detecting an abnormality in a heat treatment apparatus has been described, but an abnormality in a motor can also be detected using the present invention.
For example, in the case of a motor used in an escalator, a change in the weight of an escalator occupant may be treated as a disturbance, and the motor current (MV) and speed (PV) fluctuations caused by the change may be used to detect an abnormality by the method described above. Here, the "disturbance period" uses the "arrival time of the passenger from the entrance to the exit".
Further, by using the frictional characteristics of the motor speed reducer as a disturbance and using the fluctuations in the motor current (MV) and speed (PV), an abnormality can be detected by the method described above. Here, "time required for one rotation after deceleration" is used as the "disturbance period".

INDUSTRIAL APPLICABILITY The anomaly detection device and anomaly detection method of the present disclosure are useful for detecting an anomaly in devices in which periodic disturbances occur.
useful for

1 Abnormality detection device 11 Control unit 111 Setting unit 112 Storage processing unit 113 Average calculation unit 114 Steady-state gain calculation unit 115 Detection unit 116 Output processing unit 12 Storage unit 121 Operation amount data 122 Measured value data 123 Average value data 124 Setting data 13 Communication Interface 2 Heat treatment equipment W Work

Claims (12)

An anomaly detection device for detecting an anomaly in a system in which a periodic disturbance occurs,
A manipulated variable average value that is an average value of the manipulated variable that operates the system for a predetermined period determined in consideration of the periodic disturbance, and a measurement of the predetermined period obtained by the sensor provided in the system by manipulating the manipulated variable. an average calculation unit for obtaining a sensor average value that is the average value of the values;
a steady-state gain calculation unit that obtains a value obtained by dividing the sensor average value by the operation amount mean value as a steady-state gain to be used for estimating an abnormality;
a detection unit that detects an abnormality by comparing the steady-state gain with a predetermined threshold;
Abnormality detection device. An anomaly detection device for detecting an anomaly in a heat treatment apparatus in which periodic disturbance occurs,
A manipulated variable average value that is an average value of manipulated variables for manipulating the temperature of the heat treatment apparatus for a predetermined period determined in consideration of the periodic disturbance; an average calculation unit that obtains an average temperature value that is an average value of measured values of temperature;
a steady-state gain calculation unit that obtains a value obtained by dividing the temperature average value by the manipulated variable mean value as a steady-state gain to be used for estimating an abnormality;
a detection unit that detects an abnormality by comparing the steady-state gain with a predetermined threshold;
Abnormality detection device. The average calculation unit calculates the average value for a period determined according to the period of the periodic disturbance that occurs periodically in the heat treatment apparatus at the timing when the amount of change in the measured value becomes equal to or greater than a predetermined amount of change. The abnormality detection device according to claim 2, wherein an average value of the manipulated variable and the measured value is calculated. The averaging unit calculates the manipulated variable and the operation amount for a period determined according to the period of periodic disturbances that occur periodically in the heat treatment apparatus at the timing when a predetermined period has passed since the object to be processed was put into the heat treatment apparatus. The abnormality detection device according to claim 2, wherein an average value of measured values is calculated. The abnormality detection device according to any one of claims 2 to 4, wherein the average calculation unit continuously calculates a moving average of the operation amount and the measured value as the operation amount average value and the temperature average value. 6. The abnormality detection according to any one of claims 2 to 5, wherein the average calculation unit obtains a moving average using the manipulated variable and the measured value for a period longer than a period of a predetermined periodic disturbance generated in the heat treatment apparatus. Device. 7. The steady-state gain calculation unit determines a value obtained by dividing a value obtained by subtracting the mean temperature value from a measured value obtained in advance as an initial value by the manipulated variable as the steady-state gain. An anomaly detection device as described. 7. The steady-state gain calculator according to any one of claims 2 to 6, wherein the steady-state gain is obtained by dividing a value obtained by subtracting the average temperature value by an offset value input by a user by the manipulated variable. anomaly detection device. The steady-state gain calculation unit acquires the environmental temperature of the object to be processed installed in the heat treatment apparatus, and divides a value obtained by subtracting the average temperature value by the environmental temperature, by the manipulated variable, and obtains the steady-state gain The abnormality detection device according to any one of claims 2 to 6. The detection unit receives a signal specifying whether a work is put into or taken out of the heat treatment apparatus, and detects a plurality of The anomaly detection device according to any one of claims 2 to 9, wherein a threshold to be used is selected from the thresholds of . The detection unit receives a signal specifying the type of disturbance generated based on the work loaded into the heat treatment apparatus, and selects a threshold to be used from among a plurality of thresholds according to the type specified by the signal. The abnormality detection device according to any one of claims 2 to 9. An anomaly detection method for detecting an anomaly in a system in which a periodic disturbance occurs,
A step of obtaining an average value of the operation amount that is the average value of the operation amount for operating the system for a predetermined period determined in consideration of the periodic disturbance;
a step of obtaining a sensor average value, which is an average value of measured values for the predetermined period obtained by the sensor provided in the system by manipulating the manipulated variable;
a step of obtaining a value obtained by dividing the sensor average value by the operation amount average value as a steady-state gain to be used for estimating an abnormality;
detecting an abnormality by comparing the steady-state gain with a predetermined threshold;
An anomaly detection method comprising:

Images