JP2021012096A - Abnormality detection device, abnormality detection method, method for controlling abnormality detection device, and control program - Google Patents

Abstract

To achieve an abnormality detection device that can detect an abnormality in a belt in a belt and pulley mechanism without adding a sensor.SOLUTION: An abnormality detection device (100) calculates an abnormality degree score from a first feature quantity that is the average value in a period of torque signals indicating the torque of a motor (220) that drives an endless belt (240), and a second feature quantity that is a variation parameter indicating the magnitude of variations in the period of the torque signals, and when the abnormality degree score exceeds a predetermined threshold, determines that the state of the endless belt is abnormal.SELECTED DRAWING: Figure 1

Description

The present invention relates to an abnormality detection device, an abnormality detection method, a control method of the abnormality detection device, and a control program for detecting an abnormality of a belt suspended between pulleys.

The belt and pulley mechanism, which is a mechanism for transmitting rotational motion between shafts, is widely used in industry. In particular, a method in which an endless belt suspended between a plurality of pulleys is driven by a pulley connected to a rotating shaft of a servomotor controlled by a control device such as a servo driver is a method in which the operation of the belt & pulley mechanism is precise. It is used in situations where control is required.

Japanese Unexamined Patent Publication No. 8-217302

In the belt & pulley mechanism, if the belt loosens due to deterioration over time, the required operation may not be possible. Therefore, Patent Document 1 proposes a technique for detecting looseness of a belt by providing a sensor that directly detects meandering of the belt during operation. However, in the present technology, it is necessary to provide a sensor capable of directly detecting the meandering of the belt for detecting the abnormality of the belt.

One aspect of the present invention is to realize an abnormality detecting device capable of detecting an abnormality of a belt in a belt & pulley mechanism driven by a motor without adding a sensor for directly observing the meandering of the belt. To do.

The present invention adopts the following configuration in order to solve the above-mentioned problems.

The abnormality detection device according to one aspect of the present invention includes a communication unit that receives a torque signal indicating torque of a motor that drives an endless belt suspended between pulleys by rotating a pulley, and a cycle of the torque signal. The first feature amount calculation unit that calculates the torque average, which is an average value, as the first feature amount, and the second feature, which calculates the fluctuation parameter indicating the magnitude of the fluctuation of the torque signal in the cycle as the second feature amount. When the amount calculation unit, the abnormality score calculation unit that calculates the abnormality score from the first feature amount and the second feature amount, and the abnormality score exceeds a predetermined threshold value, the state of the endless belt is abnormal. It is provided with an abnormality determination unit that determines that there is.

According to the above configuration, it is possible to realize an abnormality detection device capable of detecting an abnormality of the belt in the belt & pulley mechanism driven by the motor without adding a sensor for directly observing the meandering of the belt. Alternatively, the administrator of the belt & pulley mechanism can determine the criteria for determining the belt abnormality without considering what the operation of the belt & pulley mechanism in the cycle is. Alternatively, the biting of a foreign substance into the belt can be detected as a belt abnormality.

In the abnormality detection device according to the one aspect, the fluctuation parameter may have a configuration which is the maximum width or standard deviation in the cycle of the torque signal.

According to the above configuration, the above fluctuation parameters can be specifically defined. Alternatively, the biting of a foreign substance into the belt can be detected as a belt abnormality.

In the abnormality detection device according to the one aspect, the fluctuation parameter may have a configuration which is the maximum width or standard deviation of the high frequency component of the torque signal in the cycle of the torque signal.

According to the above configuration, the above fluctuation parameters can be specifically defined. Alternatively, the biting of a foreign substance into the belt can be detected as a belt abnormality.

In the abnormality detection device according to the one aspect, the abnormality degree score is based on the distribution of a large number of torque averages with respect to the torque signal acquired when the state of the endless belt is normal. The corrected torque average modified so as to increase the variation is used as the first feature amount, the fluctuation parameter is used as the second feature amount, and the first feature amount and the second feature amount in each cycle are numerous cycles in the two-dimensional space. It may have a configuration defined as a function of the first feature amount and the second feature amount based on the distribution of.

According to the above configuration, it is less dependent on the configuration of the belt & pulley mechanism, usage conditions, etc., and it is possible to flexibly construct a system with a belt abnormality detection function added. Alternatively, the administrator of the belt & pulley mechanism can determine the criterion for determining the belt abnormality without operating the belt & pulley mechanism in a state where the belt abnormality is intentionally generated. Alternatively, the administrator of the belt & pulley mechanism can construct a system that can detect a belt abnormality in a range of a required temperature having a range by acquiring torque signal data at a predetermined temperature.

The abnormality detection method according to one aspect of the present invention includes a step of receiving a torque signal indicating torque of a motor for driving an endless belt suspended between pulleys by rotating a pulley, and an average in the cycle of the torque signal. A step of calculating the torque average, which is a value, as a first feature amount, a step of calculating a fluctuation parameter indicating the magnitude of fluctuation of the torque signal in the cycle as a second feature amount, the first feature amount, and the first feature amount. 2. It includes a step of calculating an abnormality degree score from the feature amount, and a step of determining that the state of the endless belt is abnormal when the abnormality degree score exceeds a predetermined threshold value.

According to the above configuration, it is possible to realize an abnormality detection device capable of detecting an abnormality of the belt in the belt & pulley mechanism driven by the motor without adding a sensor for directly observing the meandering of the belt. Alternatively, the administrator of the belt & pulley mechanism can determine the criteria for determining the belt abnormality without considering what the operation of the belt & pulley mechanism in the cycle is. Alternatively, the biting of a foreign substance into the belt can be detected as a belt abnormality.

In the abnormality detection method according to the one aspect, the abnormality degree score is based on the distribution of a large number of torque averages with respect to the torque signal acquired when the state of the endless belt is normal. The corrected torque average modified so that the variation is expanded is used as the first feature amount, the fluctuation parameter is used as the second feature amount, and the first feature amount and the second feature amount in each cycle are numerous cycles in the two-dimensional space. It may have a configuration defined as a function of the first feature amount and the second feature amount based on the distribution of.

According to the above configuration, it is less dependent on the configuration of the belt & pulley mechanism, usage conditions, etc., and it is possible to flexibly construct a system with a belt abnormality detection function added. Alternatively, the administrator of the belt & pulley mechanism 200 can determine the determination criteria for the belt abnormality without operating the belt & pulley mechanism in a state where the belt abnormality is intentionally generated. Alternatively, the administrator of the belt & pulley mechanism 200 can construct a system capable of detecting a belt abnormality in a range of a required temperature having a range by acquiring torque signal data at a predetermined temperature.

The control method of the abnormality detection device according to one aspect of the present invention is the control method of the abnormality detection device according to one aspect of the present invention, in which the step of determining the abnormality degree score and the abnormality degree of the abnormality detection device are defined. The score calculation unit includes a step of setting the abnormality degree score.

The control program according to one aspect of the present invention causes the computer to execute each step described in the control method of the abnormality detection device according to the aspect of the present invention, thereby operating the computer as a control device of the abnormality detection device. It has a configuration.

According to the control method of the abnormality detection device according to one aspect of the present invention or the control program according to one aspect of the present invention, the abnormality of the belt in the belt & pulley mechanism driven by the motor can be directly detected by the meandering of the belt. It is possible to realize an abnormality detection device and an abnormality detection system that can detect without adding a sensor for observation. Alternatively, the administrator of the belt & pulley mechanism can determine the criteria for determining the belt abnormality without considering what the operation of the belt & pulley mechanism in the cycle is. Alternatively, the biting of a foreign substance into the belt can be detected as a belt abnormality.

Alternatively, it is less dependent on the configuration of the belt & pulley mechanism, usage conditions, etc., and it is possible to flexibly construct a system with a belt abnormality detection function added. Alternatively, the administrator of the belt & pulley mechanism can determine the criterion for determining the belt abnormality without operating the belt & pulley mechanism in a state where the belt abnormality is intentionally generated. Alternatively, the administrator of the belt & pulley mechanism can construct a system that can detect a belt abnormality in a range of a required temperature having a range by acquiring torque signal data at a predetermined temperature.

According to the abnormality detecting device according to one aspect of the present invention, the abnormality of the belt in the belt & pulley mechanism driven by the motor can be detected without adding a sensor for directly observing the meandering of the belt.

According to the abnormality detection method according to one aspect of the present invention, an abnormality detection device capable of detecting an abnormality of a belt in a belt & pulley mechanism driven by a motor without adding a sensor for directly observing meandering of the belt. Can be realized.

It is a figure which shows the whole structure including the abnormality detection apparatus which concerns on Embodiment 1 of this invention, and the belt & pulley mechanism which an abnormality detection apparatus monitors. It is a time chart which shows the torque signal in Embodiment 1 of this invention. The case where the belt is in a normal state and the case where the belt is loose are shown in comparison. It is a figure which shows the statistical data of the torque average in Embodiment 1 of this invention for four kinds of temperature times. The case where the belt is in a normal state and the case where the belt is loose (two ways) are shown in comparison. It is a time chart which shows the torque signal in Embodiment 1 of this invention. The case where the belt is in a normal state and the case where a foreign substance is caught in the belt are compared and shown. It is a figure which shows the statistical data of the maximum width of the torque average in Embodiment 1 of this invention for four kinds of temperature times. The case where the belt condition is normal and the case where foreign matter is caught (three ways) are compared and shown. It is a flowchart explaining the machine learning process in Embodiment 1 of this invention. It is a graph which shows the distribution for explaining the torque average and the corrected torque average in Embodiment 1 of this invention. It is a flowchart explaining the variation parameter processing (subroutine) in the machine learning processing in Embodiment 1 of this invention. It is a conceptual diagram for demonstrating that the abnormality degree score is calculated by the machine learning process in Embodiment 1 of this invention. It is a flowchart explaining the monitoring process executed by the abnormality detection apparatus which concerns on Embodiment 1 of this invention. It is a flowchart explaining the variation parameter processing (subroutine) in the machine learning processing in Embodiment 2 of this invention. It is a time chart which shows the high frequency component of the torque signal in Embodiment 2 of this invention. A comparison is shown between the case where the belt is in a normal state and the case where a foreign substance is caught. It is a figure which shows the statistical data of the maximum width of the high frequency component of a torque average in Embodiment 2 of this invention for four kinds of temperature times. The case where the belt condition is normal and the case where foreign matter is caught (three ways) are compared and shown.

Hereinafter, embodiments according to one aspect of the present invention will be described.

§1 Application example First, an example of a situation in which the present invention is applied will be described. The abnormality detection device according to the present embodiment is applied to a belt & pulley mechanism including an endless belt suspended between a plurality of pulleys and a motor for driving the endless belt by rotating the pulleys.

The abnormality detection device according to the present embodiment includes a communication unit that receives a torque signal indicating torque of a motor that drives an endless belt suspended between pulleys by rotating the pulleys.

Further, the abnormality detection device according to the present embodiment has a first feature amount calculation unit that calculates a torque average, which is an average value in the cycle of the torque signal, as a first feature amount, and a large fluctuation of the torque signal in the cycle. It is provided with a second feature amount calculation unit that calculates a variation parameter indicating the above as a second feature amount.

Further, the abnormality detection device according to the present embodiment includes an abnormality score calculation unit that calculates an abnormality score from the first feature amount and the second feature amount, and the endlessness when the abnormality score exceeds a predetermined threshold value. It is equipped with an abnormality determination unit that determines that the state of the belt is abnormal.

By providing the above configuration, the abnormality detection device according to the present embodiment can determine whether or not the state of the endless belt is abnormal based on the motor or the torque signal from the control device that controls the motor. it can.

Therefore, the abnormality of the belt in the belt & pulley mechanism driven by the motor can be detected without adding a sensor for directly observing the meandering of the belt. Moreover, since it is not necessary to modify the belt & pulley mechanism, it can be easily retrofitted to the existing belt & pulley mechanism.

Therefore, the administrator of the system provided with the belt & pulley mechanism can perform maintenance before the belt & pulley mechanism malfunctions by using the abnormality detection device according to the present embodiment. Therefore, it becomes possible to suppress the occurrence of malfunction of the belt & pulley mechanism.

Alternatively, the abnormality detection device according to the present embodiment can detect not only the looseness of the belt but also the biting of foreign matter into the belt as an abnormality by providing the above configuration. Alternatively, the administrator of the belt & pulley mechanism can determine the criteria for determining the belt abnormality without considering what the operation of the belt & pulley mechanism in the cycle is.

§2 Configuration example Below, a more specific example of the scene to which the present invention is applied will be described with reference to FIGS. 1 to 13. An example of a specific situation to which the present invention is applied may be factory automation (FA) at a production site. In this case, the anomaly detection device can be integrated into a programmable logic controller (PLC) as a central controller. In the following configuration example, a case where the PLC implements a function as an abnormality detection device will be described as an example.

[Embodiment 1]
<Overall configuration>
FIG. 1 is a schematic view showing an overall system configuration including an abnormality detection device 100 according to the first embodiment and a belt & pulley mechanism 200 monitored by the abnormality detection device 100. The belt & pulley mechanism 200 includes a servo driver 210, a servomotor 220, a pulley 221 connected to the rotating shaft of the servomotor 220 (motor), a movable mechanism 230, and a pulley 231 and a belt 240 connected to the rotating shaft of the movable mechanism 230. (Endless belt) is provided.

The servomotor 220 generates the driving force in the belt & pulley mechanism 200. The servo driver 210 controls the operation of the belt & pulley mechanism 200 by feedback-controlling the rotation of the servomotor 220. The servo driver 210 monitors the torque of the servo motor 220 when feedback-controlling the rotation of the servo motor 220.

The pulley 221 and the pulley 231 are rotating bodies on which an endless (ring-shaped) belt 240 is suspended. The pulley 221 connected to the rotation shaft of the servomotor 220 transmits the rotational movement of the rotation shaft of the servomotor 220 to the pulley 231 through the belt 240. The movable mechanism 230 is a mechanism provided with a movable portion that is operated by the rotational movement of the pulley 231. As a specific example of the movable mechanism 230, there is a movable part for executing processing, transportation, and inspection of a product or the like in a production or inspection process.

In FIG. 1, the belt 240 is drawn to be suspended between two pulleys 221, 231 but, as is known, may be suspended between three or more pulleys. The belt & pulley mechanism is a mechanism that is often used when rotating motions in a plurality of movable mechanisms are interlocked to operate, and the present invention can naturally be applied to such a configuration.

A computer 501 and an input / output device 502 are connected to the abnormality detection device 100. The input / output device 502 is a device that can receive input from the administrator to the abnormality detection device 100, which is also a PLC, and receive and display information from the abnormality detection device 100. The computer 501 may operate in the same manner as the input / output device 502. Further, the computer 501 may be a computer (control device of the abnormality detection device), which is also called a tool, provided with an application capable of supporting the change of the setting inside the abnormality detection device 100 which is a PLC.

<Configuration of abnormality detection device>
The abnormality detection device 100 includes a first communication unit 101 (communication unit), a recording unit 102, a control unit 103, a second communication unit 104, and an abnormality detection unit 110. The first communication unit 101 is a communication interface that receives the torque signal St output by the servo driver 210. The recording unit 102 is a memory for recording the torque signal St and various other information.

The control unit 103 is a functional block included in the abnormality detection device 100, which is a PLC, capable of controlling an external device and acquiring information from the external device to execute various information processing. The control unit 103 controls the servo driver 210 to cause the belt & pulley mechanism 200 to perform a required operation. Therefore, the control unit 103 exchanges information and acquires information regarding control with an external device such as the servo driver 210 through the first communication unit 101.

The second communication unit 104 is a communication interface that communicates with an external computer or the like. In the first embodiment, the computer 501 and the input / output device 502 are connected to the second communication unit 104 via an appropriate network.

The abnormality detection unit 110 detects an abnormality of the belt 240 in the belt & pulley mechanism 200 based on the torque signal St received by the first communication unit 101 or received by the first communication unit 101 and recorded in the recording unit 102. To execute. The abnormality detection unit 110 is provided with functional blocks of a first feature amount calculation unit 111, a second feature amount calculation unit 112, an abnormality degree score calculation unit 113, and an abnormality determination unit 114. The operations executed by each of these functional blocks will be described later.

<Torque signal when there is looseness>
In explaining the operation of the abnormality detection device 100 according to the first embodiment, first, the principle of the abnormality detection device 100 for detecting the abnormality of the belt 240 will be described.

FIG. 2 is a time chart showing an example of a torque signal St indicating the torque of the servomotor 220 output by the servo driver 210. In the time chart of FIG. 2, the vertical axis is an arbitrary unit. FIG. 2 shows a torque signal waveform when the belt 240 is normal, and a torque signal waveform when the belt & pulley mechanism 200 is operated by intentionally loosening the belt experimentally.

The torque signal St in each case shows periodic behavior, and FIG. 2 shows the three cycles. One cycle of the torque signal St corresponds to a processing operation for one unit executed by the belt & pulley mechanism 200 under the control of the abnormality detection device 100 which is a PLC.

Here, one unit corresponds to one product (product) to be processed if it is a product processing process or inspection process. In the processing for one unit, the servomotor 220 executes necessary operations such as start / stop of rotation, acceleration / deceleration, and reversal. Therefore, the torque signal St exhibits complicated behavior within one cycle. Further, since the processing operation for one unit is repeated for a large number of products to be processed, the torque signal St exhibits a periodic behavior.

As shown in FIG. 2, when comparing the case where the belt 240 is normal and the case where the belt 240 is loose with respect to the torque signal St, the magnitude of the torque is generally smaller in the latter case. If the average of the signal magnitudes is taken in each cycle of the torque signal St and it is taken as the torque average At, the magnitude of the torque average At is considered to be small when there is looseness.

FIG. 3 is a box plot showing statistical data obtained by acquiring torque average At for a large number of cycles. FIG. 3 shows a comparison of the results when the belt 240 has no looseness and is normal, and when the looseness is small and large. In addition, FIG. 3 also shows the results when the air temperature at the time of acquiring the torque signal St is 10 ° C, 20 ° C, 30 ° C, and 40 ° C.

Comparing at a constant temperature, it is clear that the torque average At becomes smaller as the looseness becomes larger than in the normal state. Therefore, from the result of FIG. 3, it is clear that the looseness of the belt 240 may be detected by monitoring the torque average At. However, it is also shown in FIG. 3 that the torque average At fluctuates with temperature. The torque average At increases as the temperature decreases, and the torque average At decreases as the temperature increases.

However, as shown in FIG. 3, the degree of decrease in the average torque At due to the occurrence of loosening is larger than the fluctuation caused by the temperature change in the normal range in the room such as the factory where the belt & pulley mechanism 200 is installed. ..

For example, a range of torque average At as shown by an arrow in FIG. 3 is determined to be a normal state, and a case outside the range is determined to be an abnormality. If it is assumed that the temperature at the location where the belt & pulley mechanism 200 is installed is within the range of 10 ° C to 40 ° C, the looseness of the belt & pulley mechanism 200 is large enough to cause malfunction (loose in FIG. 3). It is possible to detect a belt abnormality before (looseness larger than (large)) occurs. Further, in that case, it is shown that there is a possibility that erroneous detection that the belt is abnormal is not generated due to the fluctuation caused by the temperature change even though the belt is not loosened.

<Torque signal when foreign matter is caught>
In addition to the loosening of the belt 240, it is assumed that foreign matter is caught in the belt 240 as a cause of the belt abnormality that causes an abnormality in the operation of the belt & pulley mechanism 200.

FIG. 4 is a time chart showing a torque signal waveform when the belt 240 is normal and a torque signal waveform when the belt 240 is operated by intentionally causing foreign matter to bite into the belt 240. The signal waveform in the normal case is the same as in the case of FIG.

The torque signal St when a foreign substance is caught shows a characteristic waveform such that the torque fluctuates greatly for a moment from the case where the torque is normal. This phenomenon is caused by foreign matter coming into contact with the pulley. When a parameter (fluctuation parameter Dt) indicating the magnitude of torque fluctuation is calculated in each cycle of the torque signal St, it is considered that the fluctuation parameter Dt becomes larger when foreign matter is bitten than in the normal state.

FIG. 5 adopts the maximum width of the torque signal St in each cycle as the fluctuation parameter Dt, and shows the statistical data acquired in a large number of cycles as a box plot. The maximum width is the difference between the maximum value and the minimum value of the torque signal St in each cycle.

FIG. 5 shows a comparison of the results when the belt 240 is normal without foreign matter being caught and when the belt 240 is caught by foreign matter. As for the foreign matter, the results when the size is large, medium, and small are shown. In addition, FIG. 5 also shows the results when the air temperature at the time of acquiring the torque signal St is 10 ° C, 20 ° C, 30 ° C, and 40 ° C.

As is clear from FIG. 5, when the foreign matter is caught, the maximum width of the torque signal St is remarkably increased as compared with the normal case. Therefore, from the results of FIG. 5, it was shown that if the fluctuation parameter Dt such as the maximum width of the torque signal St is monitored, it is possible to detect an abnormality due to the biting of a foreign substance into the belt 240.

The fluctuation parameter Dt is not limited to the maximum width of the torque signal St, and even if a parameter indicating the magnitude of fluctuation of other torques such as the standard deviation of the torque signal St within the period is used, an abnormality is detected in the same manner. Is possible.

<Machine learning process>
As described above, if the torque signal St is monitored, there is a possibility that a belt abnormality due to loosening of the belt 240 or biting of a foreign object can be detected. However, the structure of the belt & pulley mechanism 200 is various, and the operation of the belt & pulley mechanism 200, which is controlled by PLC at a production site or the like, is also various. Therefore, in order to actually detect a belt abnormality of the belt & pulley mechanism 200 based on the above principle, it is necessary to deal with these various situations.

Therefore, in the abnormality detection device 100 according to the first embodiment, the belt abnormality determination standard is determined according to the usage conditions at the site where the belt & pulley mechanism 200 is actually used. At that time, the determination criteria are determined by the machine learning process using a large number of actually measured data of the belt & pulley mechanism 200 that actually monitors.

FIG. 6 is a flowchart showing a procedure (machine learning process) for determining a determination criterion in the abnormality detection device 100 by using such machine learning. When performing the machine learning process, the manager operates the belt & pulley mechanism 200 in a normal state under a predetermined temperature, and causes the recording unit 102 of the abnormality detection device 100 to record the torque signal St.

The computer 501 executes machine learning based on the torque signal St recorded in the recording unit 102, further controls the abnormality detection device 100, and causes the abnormality detection device 100 to set a specific determination model. It is a equipped computer. The procedure (machine learning process) for determining the determination criteria shown in the flowchart of FIG. 6 will be described in detail below.

Step S11: The computer 501 acquires the data of the torque signal St from the recording unit 102 of the abnormality detection device 100. The data of the torque signal St becomes data for machine learning.

Step S12: Subsequently, the computer 501 calculates one cycle of the operation of the belt & pulley mechanism 200 as a frame, and the average value of the torque signal St for each frame as the torque average At.

Step S13: Subsequently, the computer 501 calculates the average value Aa from the calculated large number of torque averages At.

Step S14: Subsequently, the computer 501 converts the torque average At data of each frame into the corrected torque average Am as follows. When the torque average At is the average value Aa, the corrected torque average Am is equal to the torque average At.

When the torque average At is smaller than the average value Aa, the difference from the average value Aa is converted into a value that becomes Em times. When the relationship between the torque average At and the corrected torque average Am is expressed by a mathematical formula, (Aa-At) x Em = Aa-Am in each frame.

When the torque average At is larger than the average value Aa, the difference from the average value Aa is converted into a value that becomes Ep times. When the relationship between the torque average At and the corrected torque average Am is expressed by a mathematical formula, (At-Aa) x Ep = Am-Aa in each frame.

FIG. 7 is a diagram conceptually showing the distribution of the torque average At and the distribution of the corrected torque average Am. The corrected torque average Am is obtained by expanding the variation in the distribution of the torque average At by Em times in a region smaller than the average value and by Ep times in a region larger than the average value.

Step S15: Subsequently, the computer 501 executes the variation parameter calculation process. The flowchart of the variation parameter calculation process (subroutine) in the first embodiment is shown in FIG. The variable parameter calculation process is as follows.

Step S21: The computer 501 calculates the maximum width of the torque signal St in each frame and sets it as the fluctuation parameter Dt. Next, the variation parameter calculation process ends, and the process proceeds to step S16.

Step S16: Subsequently, the computer 501 calculates the anomaly score So from the distribution data for each frame in the two-dimensional space in which the corrected torque average Am is the first feature amount and the variation parameter Dt is the second feature amount. ..

The distribution data for each frame should be regarded as the distribution under normal conditions. For the calculation of the anomaly score So, a machine learning algorithm for calculating the anomaly score, which is a value indicating the probability of being an abnormal value (outlier), can be used from the distribution of the feature amount in the normal state. As a specific example of such an algorithm, a known method such as an isolation forest (Isolation Forest) or a LOF (Local Outlier Factor) can be used.

Here, the corrected torque average Am as the first feature amount is an enlargement of the distribution of the actual torque average At acquired under a predetermined temperature. FIG. 9 is a conceptual diagram for explaining the relationship between the distribution data and the anomaly score for each frame. Graph 9A of FIG. 9 is a plot of data for each frame, with black dots for the torque average At and white dots for the modified torque average Am.

Graph 9B of FIG. 9 is a contour graph showing the anomaly score So with the first feature amount, which is the corrected torque average Am, on the horizontal axis and the second feature amount, which is the fluctuation parameter Dt, on the vertical axis. , Horizontal axis, vertical axis correspond to graph 9A. The situation in which the anomaly score is calculated based on the distribution data (corrected torque average Am and fluctuation parameter Dt) for each frame whose distribution is expanded in the horizontal axis direction is schematically shown.

According to the machine learning algorithm that calculates the anomaly score, the anomaly score So becomes smaller in the region where a lot of data for each frame exists, and the anomaly score So increases as it goes out of the region. 9 is schematically represented.

The fact that the corrected torque average Am as the first feature amount is an enlargement of the distribution of the actual torque average At acquired under a predetermined temperature is a certain temperature in the statistical data at the normal time shown in FIG. When the torque signal St is acquired in, it corresponds to expanding the range regarded as normal to the range indicated by the arrow. The coefficient Em and the coefficient Ep specified in step S14 are coefficients for expanding the variation of the data acquired at a certain temperature so as to correspond to the assumed indoor air temperature condition.

It is assumed that the temperature of the room in which the belt & pulley mechanism 200 is installed is in the range of 10 ° C to 40 ° C. For example, when the statistical data at the normal time is acquired at 10 ° C., the distribution of data smaller than the mean value Aa is greatly expanded (Em size), and the data is larger than the mean value Aa. The expansion of the distribution is small (Ep small).

Further, for example, when the acquisition of statistical data at normal times is performed at 30 ° C., the expansion of the distribution is reduced for data smaller than the mean value Aa (Em small), and for data larger than the mean value Aa. Increases the expansion of the distribution (large Ep).

As described above, the appropriate values of the coefficient Em and the coefficient Ep are adjusted according to the relationship between the assumed air temperature condition when the belt & pulley mechanism 200 operates and the air temperature at the time of data acquisition in the normal state. Therefore, it is preferable that the computer 501 prepares a reference value in advance in step S14 according to these values and selects the value by the administrator.

In this way, in step S16, the abnormality degree score is determined as a function of the first feature amount and the second feature amount. Since the anomaly score So is a function for the first feature amount and the second feature amount, its value may be expressed as a table for the first feature amount and the second feature amount.

In this way, the anomaly score So calculated as a value at a point in the two-dimensional space defined by the first feature amount and the second feature amount is used for monitoring the belt abnormality. For example, when the belt 240 is monitored, if the belt 240 is loosened significantly, the torque average At (first feature amount) decreases, which is represented by a point shown as "abnormality (looseness)" in Graph 9A of FIG. It becomes a state, a large abnormality score So is calculated, and an abnormality is detected. Further, for example, when the belt 240 is monitored, if a foreign substance is caught in the belt 240, the maximum width (second feature amount) of the torque signal St increases, and in the graph 9A of FIG. The state represented by the indicated points is obtained, a large abnormality score So is calculated, and the abnormality is detected.

Step S17: Computer 501 determines the value selected by the administrator as the threshold Th. The threshold value Th may be a specified value (default value).

Step S18: The computer 501 controls the abnormality detection device 100 to set the abnormality degree score So as a function for the first feature amount and the second feature amount in the abnormality degree score calculation unit 113. Further, the threshold value Th is set in the abnormality determination unit 114. Then the flow ends.

<Monitoring process>
The abnormality detection device 100 according to the first embodiment has a belt & pulley mechanism 200 according to a determination standard calculated based on a torque signal St acquired at the site where the belt & pulley mechanism 200 is used by the above machine learning process. Perform belt abnormality monitoring.

FIG. 10 is a flowchart showing a monitoring process executed by the abnormality detection device 100. The abnormality detection device 100 repeatedly executes the flow represented by steps S41 to S48 in FIG. 10 during the period during which monitoring is being executed. The flow iteration is performed at intervals equal to the frame.

Step S41: The abnormality detection device 100 acquires data for one cycle of the torque signal St from the servo driver 210 through the first communication unit 101. The abnormality detection device 100 may acquire data for one cycle of the torque signal St from the servo driver 210, which is sequentially recorded in the recording unit 102, by reading the data for one cycle from the recording unit 102.

Step S42: Subsequently, the first feature amount calculation unit 111 calculates the torque average At, which is the average of the torque signals St in the frame, and uses it as the first feature amount. In this way, at the time of monitoring, the torque average At is adopted as the first feature amount.

Step S43: Subsequently, the second feature amount calculation unit 112 calculates the variation parameter Dt in the frame and sets it as the second feature amount. In the first embodiment, the variation parameter Dt is the maximum width of the torque signal St in the frame.

Step S44: Subsequently, the abnormality degree score calculation unit 113 calculates the abnormality degree score So from the first feature amount and the second feature amount.

Step S45: Subsequently, the abnormality determination unit 114 determines whether or not the abnormality degree score So is equal to or less than the threshold value Th. If it is determined that the value is below the threshold value (So ≦ Th) (YES in S45), the process proceeds to step S46, and in other cases (NO in S45), the process proceeds to step S47.

Step S46: When the abnormality score So is equal to or less than the threshold Th, the abnormality determination unit 114 determines that the state of the belt is normal. Then the flow ends.

Step S47: If the abnormality score So is not equal to or less than the threshold Th, the abnormality determination unit 114 determines that the state of the belt is abnormal.

Step S48: Subsequently, the abnormality determination unit 114 notifies the input / output device 502 through the second communication unit that the state of the belt is abnormal. The input / output device 502 appropriately displays on the provided display that the state of the belt is abnormal, and notifies the administrator. Then the flow ends.

In the above flow, the abnormality detection device 100 may appropriately record the determination result of the state of the belt in the frame in the recording unit 102.

<Effect>
According to the abnormality detection device 100, the occurrence of a belt abnormality is detected based on the torque signal St. Therefore, a belt abnormality such as loosening of the belt 240 can be detected without newly providing the belt & pulley mechanism 200 with a sensor that directly detects the meandering of the operation of the belt 240. Therefore, the looseness of the belt 240 can be detected at low cost. Further, since the belt & pulley mechanism 200 does not need to be modified, it can be easily retrofitted to the existing belt & pulley mechanism 200.

According to the abnormality detection device 100, the setting of the belt abnormality determination standard is determined by machine learning based on the torque signal St acquired at the site where the belt & pulley mechanism 200 is actually used. Therefore, it is less dependent on the configuration and usage conditions of the belt & pulley mechanism 200, or the model of the servo driver 210, and it is possible to flexibly construct a system to which a belt abnormality detection function is added.

According to the abnormality detection device 100, the determination criteria for the belt abnormality are set by machine learning based on the torque signal St acquired for the belt & pulley mechanism 200 that operates normally. Therefore, the administrator of the belt & pulley mechanism 200 does not need to operate the belt & pulley mechanism 200 in a state where an abnormality such as loosening of the belt or biting into the belt is intentionally generated in order to perform machine learning. The work of setting the abnormality judgment criteria can be easily performed.

According to the abnormality detection device 100, the setting of the belt abnormality determination standard based on the torque signal St is performed by the above-mentioned feature amount calculated from the data for the cycle of the torque signal St. Therefore, the administrator of the belt & pulley mechanism 200 can set the judgment criteria without considering what the operation of the belt & pulley mechanism 200 is in the cycle, and the belt & pulley in setting the judgment criteria. There is no need to analyze the operation of the mechanism 200 in detail.

According to the abnormality detection device 100, the occurrence of a belt abnormality is detected based on the torque signal St. Therefore, as experimentally clarified by the present inventors, there are few false detections and looseness of the belt 240 is detected. become able to. Further, it becomes possible to detect the biting of foreign matter into the belt 240 at the same time.

Further, the determination standard for the belt abnormality based on the torque signal St is set based on the corrected torque average Am that expands the variation in the distribution of the torque average At calculated from the data for the period of the torque signal St. Therefore, the administrator of the belt & pulley mechanism 200 can detect a belt abnormality in a range of a required temperature range by acquiring torque signal St data under a predetermined temperature when performing machine learning. Can be constructed.

By using the abnormality detection device 100 according to the first embodiment, the administrator of the system provided with the belt & pulley mechanism can perform maintenance before the belt & pulley mechanism 200 malfunctions. Therefore, it is possible to suppress the occurrence of malfunction of the belt & pulley mechanism 200.

In the first embodiment, the maximum width in the cycle of the torque signal St was adopted as the fluctuation parameter Dt during the machine learning process and monitoring. However, even if a parameter related to fluctuation in the cycle of another torque signal St, for example, a standard deviation is used as the fluctuation parameter Dt, the abnormality detection device can be operated in the same manner. However, if the standard deviation is adopted, the amount of calculation increases, and from that point, it is preferable to adopt the maximum width.

[Embodiment 2]
Embodiment 2 of the present invention will be described below. For convenience of explanation, the same reference numerals are given to the members having the same functions as the members described in the above embodiment, and the description is not repeated. The second embodiment is the same except that the method of calculating the second feature amount is different from that of the first embodiment.

FIG. 11 is a flowchart showing a variation parameter calculation process (subroutine) of the machine learning process in the second embodiment. In the second embodiment, the details of the variation parameter calculation process are as follows.

Step S31: Computer 501 calculates the moving average waveform of the torque signal St. Here, the moving average waveform is a waveform in which the value of each measurement point of the torque signal St is replaced with the average value of the P points before and after the value.

Step S32: The computer 501 calculates a waveform obtained by subtracting the moving average waveform from the torque signal St as a high-frequency signal.

Step S33: The computer 501 calculates the maximum width of the high frequency signal of each frame and sets it as the fluctuation parameter Dt. Next, the variation parameter calculation process ends, and the process proceeds to step S16.

The high-frequency signal calculated as described above is a waveform obtained by extracting fine vibration components of the torque signal St according to the magnitude of the score P. FIG. 12 shows a high-frequency signal extracted from the torque signal St in the normal state shown in FIG. 4 and the torque signal St in the abnormal state in which foreign matter is caught in the belt 240.

FIG. 13 is a diagram showing a box plot of statistical data similar to that of FIG. 5 for the maximum width of each frame of the extracted high-frequency signal. Compared with FIG. 5, when the fluctuation parameter Dt based on the high frequency signal is adopted rather than based on the torque signal St, the difference in the feature amount is larger between the normal case and the abnormal case where foreign matter is caught. It has been shown that it can be clarified.

However, when a high-frequency signal is adopted, the administrator of the belt & pulley mechanism 200 needs to appropriately set the score P, which is a parameter for extracting high-frequency components, when executing machine learning. The point P needs to be set according to the operation of the servomotor 220 under the control of the servo driver 210 and the interval of the torque signal St data, which is a burden on the administrator of the belt & pulley mechanism 200.

Similarly, in the monitoring process executed by the abnormality detection device 100 in the second embodiment, as the second feature amount, the second feature amount is calculated by the second feature amount calculation unit 112 according to the processes from steps S31 to S33. Is calculated. That is, in step S43 in the flowchart of FIG. 10, the maximum width of the high frequency signal of the frame is set as the second feature amount.

The second embodiment also has the same effect as that of the first embodiment. Further, according to the second embodiment, there is a possibility that an abnormality caused by a foreign substance being caught in the belt can be detected more accurately.

In the second embodiment, the maximum width in the period of the high frequency signal was adopted as the fluctuation parameter Dt during the machine learning process and monitoring. However, even if other parameters related to the fluctuation in the period of the high frequency signal, for example, the standard deviation, are used as the fluctuation parameter Dt, the same operation can be performed. However, if the standard deviation is adopted, the amount of calculation increases, so it was preferable to adopt the maximum width at that point.

[Example of realization by software]
The functional blocks of the abnormality detection device 100 (particularly, the control unit 103 and the abnormality detection unit 110) may be realized by a logic circuit (hardware) formed in an integrated circuit (IC chip) or the like, or by software. You may.

In the latter case, the abnormality detection device 100 includes a computer that executes instructions of a program that is software that realizes each function. The computer includes, for example, one or more processors and a computer-readable recording medium that stores the program. Then, in the computer, the processor reads the program from the recording medium and executes it, thereby achieving the object of the present invention.

As the processor, for example, a CPU (Central Processing Unit) can be used. As the recording medium, a "non-temporary tangible medium", for example, a ROM (Read Only Memory) or the like, a tape, a disk, a card, a semiconductor memory, a programmable logic circuit, or the like can be used. Further, a RAM (Random Access Memory) for expanding the above program may be further provided.

Further, the program may be supplied to the computer via an arbitrary transmission medium (communication network, broadcast wave, etc.) capable of transmitting the program. It should be noted that one aspect of the present invention can also be realized in the form of a data signal embedded in a carrier wave, in which the above program is embodied by electronic transmission.

[Additional notes]
In the above embodiment, the operation of the belt & pulley mechanism 200 constituting the frame or the cycle of the torque signal St has been described as one cycle. However, the cycle constituting the frame in the application of the present invention is not limited to one cycle, and may be a plurality of cycles such as two cycles and three cycles, and the effect of the present invention is similarly exhibited in this case as well. ..

The present invention is not limited to the above-described embodiments, and various modifications can be made within the scope of the claims, and the embodiments obtained by appropriately combining the technical means disclosed in the different embodiments. Is also included in the technical scope of the present invention.

100 Anomaly detection device 101 1st communication unit 102 Recording unit 103 Control unit 104 2nd communication unit 110 Abnormality detection unit 111 1st feature amount calculation unit 112 2nd feature amount calculation unit 113 Abnormality score calculation unit 114 Abnormality judgment unit 200 belt & Pulley mechanism 210 Servo driver 220 Servo motor (motor)
221 231 Pulley 230 Movable mechanism 240 Belt (endless belt)
501 computer (control device for abnormality detection device)
502 I / O device

Claims (8)

A communication unit that receives a torque signal indicating torque of a motor that drives an endless belt suspended between pulleys by rotating the pulleys.
A first feature amount calculation unit that calculates the torque average, which is an average value in the cycle of the torque signal, as the first feature amount.
A second feature amount calculation unit that calculates a fluctuation parameter indicating the magnitude of fluctuation of the torque signal in the cycle as a second feature amount.
An abnormality score calculation unit that calculates an abnormality score from the first feature amount and the second feature amount,
An abnormality detection device including an abnormality determination unit that determines that the state of the endless belt is abnormal when the abnormality score exceeds a predetermined threshold value. The abnormality detection device according to claim 1, wherein the fluctuation parameter is the maximum width or standard deviation in the cycle of the torque signal. The abnormality detection device according to claim 1, wherein the fluctuation parameter is the maximum width or standard deviation of the high frequency component of the torque signal in the cycle of the torque signal. The abnormality score is
With respect to the torque signal acquired when the state of the endless belt is normal,
Based on the distribution of a large number of the torque averages, the modified torque average modified so as to expand the variation of the distribution is set as the first feature amount, and the fluctuation parameter is set as the second feature amount.
Claims 1 to 3 defined as a function of the first feature amount and the second feature amount based on the distribution of the first feature amount and the second feature amount for many periods in the two-dimensional space in each period. The abnormality detection device according to any one of the above items. A step of receiving a torque signal indicating torque of a motor that drives an endless belt suspended between pulleys by rotating the pulleys,
A step of calculating the torque average, which is an average value in the cycle of the torque signal, as the first feature amount, and
A step of calculating a fluctuation parameter indicating the magnitude of fluctuation of the torque signal in the cycle as a second feature amount, and
The step of calculating the abnormality score from the first feature amount and the second feature amount, and
An abnormality detection method comprising a step of determining that the state of the endless belt is abnormal when the abnormality degree score exceeds a predetermined threshold value. The abnormality score is
With respect to the torque signal acquired when the state of the endless belt is normal,
Based on the distribution of a large number of the torque averages, the modified torque average modified so as to expand the variation of the distribution is set as the first feature amount, and the fluctuation parameter is set as the second feature amount.
The fifth aspect of the present invention, which is defined as a function of the first feature amount and the second feature amount based on the distribution of the first feature amount and the second feature amount for a large number of periods in the two-dimensional space in each period. Abnormality detection method. The control method for the abnormality detection device according to claim 4.
The steps to determine the anomaly score and
A control method of an abnormality detection device, comprising: a step of setting the abnormality degree score in the abnormality degree score calculation unit of the abnormality detection device. A control program for operating the computer as a control device for an abnormality detection device by causing the computer to execute each step according to claim 7.

Images