JP4196975B2 - Crack detection method for drive mechanism - Google Patents

Description

  The present invention relates to a crack detection method in a mechanical system, particularly a machine tool having a linear motion shaft such as a mounting machine, a machining center, or an electric discharge machine.

Conventionally, in a crack diagnosis of a machine, a crack diagnosis using external sensor information such as an acceleration sensor that detects vibration and an AE (Acoustic Emission) sensor that detects an acoustic signal generated when a machine element is broken has been performed. It was done.
For example, a method and apparatus for diagnosing a crack in a rotating body detects axial vibrations by vibration sensors at bearings at both ends of the rotating body and simultaneously detects an AE signal by an AE sensor at the rotating body and the bearing part. It is characterized in that it has a form in which the occurrence position and depth of the crack are estimated, and an integral multiple component of the rotation period accompanying crack vibration is extracted by frequency analysis of the vibration signal, and the time is managed (for example, , See Patent Document 1).
JP 58-150859 A (7th page, Fig. 1)

  As described above, since the conventional crack diagnosis method can be realized by using an external sensor such as a vibration sensor or an AE sensor, by having an instrument that is not originally provided for the mechanical system to be diagnosed. There is a problem that the cost of equipment for diagnosis increases.

  On the other hand, the occurrence of cracks in machine elements has the potential for rapid progress due to local stress concentration in the crack, even in the case of minute cracks. Since it is appropriate to replace the member including the part, additional information such as a crack depth using an external sensor may not be necessarily required. In particular, in machine tools with linear motion axes such as mounting machines (mounters), machining centers, or electrical discharge machines, there is a problem that the rigidity of the mechanical system decreases due to cracks, leading to deterioration of hand positioning accuracy and machining accuracy. In general, this is dealt with by exchanging a member including a crack occurrence site.

  The present invention eliminates an increase in equipment cost associated with such an abnormality diagnosis, detects a crack generated in a mechanical system that affects positioning accuracy at an early stage, and prompts the user to replace the component. An object is to provide a crack detection method for a mechanism.

  According to the crack detection method of the present invention, when the motor drives a movable part with a motor torque that changes in accordance with a sine wave function of a predetermined period, a motor state quantity measurement is performed by measuring an instantaneous value of the motor torque and using it as time-series data. Step, FFT calculation step for performing Fourier transform on the time series data of the instantaneous value of the motor torque, and whether the result of the Fourier transform is the presence or absence of an increase in the 2f component, which is a frequency component twice the period of the sine wave function A 2f component determining step for determining and a crack determining step for determining that there is a crack in the movable part when the conditional branching means determines that there is a 2f component are provided.

  According to the crack detection method of the present invention, the presence or absence of an increase in 2f component due to the vertical asymmetry of the amplitude of the motor torque accompanying the occurrence of a crack is determined for each fixed interval using only the internal state quantity of the servo amplifier that is a motor control device. The diagnosis is performed only by the FFT of the series data, and an additional sensor for diagnosis is not required, so that the cost can be reduced.

Embodiment 1 FIG.
FIG. 1 is an example of a drive mechanism for positioning a machine tool to which a crack detection method of the present invention is applied, and a uniaxial linear motion link that converts rotational motion of a motor into linear motion via a ball screw. It is a figure which shows the structure of the drive mechanism which has this.
In FIG. 1, a linear guide 30 and a movable part 40 are connected to the servo motor 10 via a rotating part 20, and a hand part 50 is connected to the tip of the movable part 40.
The rotational motion of the servo motor 10 is converted into a linear motion by a ball screw (not shown) and transmitted to the movable portion 40. The movable portion 40 is supported by the linear guide 30 and linearly moves, so that the tip of the linear guide 30 is moved. The hand portion 50 moves linearly. The rotating unit 20 is supported by a rotation support unit 80 fixed to the base 60, but has a degree of freedom to rotate about plus or minus 30 degrees in a direction orthogonal to the linear motion of the hand part 50.
In this crack detection method, the hand portion 50 is screwed to the load 70 screwed to the base 60.
The output of the servo amplifier 90 that determines the driving force of the servo motor 10 is controlled by the personal computer 100. On the other hand, the rotational motion of the servo motor 10 is measured by the motor encoder 110, and the motor rotation angle or the like is fed back to the servo amplifier 90 and the personal computer 100.

FIG. 2 is a block diagram of a motor control system of the drive mechanism shown in FIG.
In FIG. 2, the personal computer 100 calculates the difference between the target position given from the position command unit 101 and the signal of the motor encoder 110 which is the position information of the motor fetched from the UP / DOWN counter 102, and the position deviation. As input to the position controller 103. The position controller 103 generates a speed command based on the integration of the position deviation and the position gain. This speed command is D / A converted by the speed command unit 104 and output to the servo amplifier 90.
In the servo amplifier 90, the speed controller 91, the current controller 92, the PWM generator 93, and the current detector 94 are controlled based on the difference between the speed command output from the speed command unit 104 and the differential value of the signal from the motor encoder 110. Thus, output control to the servo motor 10 is performed. The power supply voltage is input to the PWM generator 93, and the motor current is fed back from the current detector 94 to the current controller 92.
The motor torque detector 95 calculates motor torque from this motor current.

In the crack detection method for the drive mechanism of the present invention, the following dedicated operation is performed.
In the configuration shown in FIG. 1, a sine wave position command with a constant amplitude is given from the position command unit 101 with the hand portion 50 screwed to a load 70 screwed to the base 60, and the movable unit 40 is driven with a sine wave. It is something to be made.
By performing appropriate signal processing on the motor torque, which is the internal state quantity of the servo amplifier 90 extracted as an analog signal from the motor torque detector 95 during this dedicated operation, the mechanical element such as a ball screw or movable The abnormal state of the part, that is, the crack is to be detected.
In the following description, it is assumed that the motor torque is directly measured and used. However, it can be used in place of the motor torque as long as it is a state quantity that can know changes in the motor torque such as the motor current. It is.
In addition, since this operation places a load on the inside of the drive mechanism, particularly the ball screw and the movable portion 40, it is appropriate to use an elastic body for the load 70. However, the sine wave operation, which is a dedicated operation for crack detection, is to give a sine wave position command having a constant amplitude to the drive unit, and in order to perform this dedicated operation, the load state of the link end, that is, It does not depend on whether the load is heavy or light.

Here, the influence which the crack which arose in the movable part 40 exerts on the motor torque will be described.
As shown in FIG. 1, the movable direction of the movable part 40 in the drive mechanism is only the X-axis direction.
If the dedicated operation is performed by screwing the hand portion 50 and the load 70 screwed to the base 60, the position command for the hand portion 50 has a constant amplitude, and the load neutral point, that is, the reaction force is zero. Therefore, the amplitude of the motor torque waveform at that time is constant, and as a matter of course, its effective value is also constant as a change with time.
However, once a crack occurs in the movable portion 40, the rigidity increases in the direction in which the crack is compressed, and conversely, the rigidity decreases in the direction in which the crack is pulled.

FIG. 3 is a schematic diagram illustrating the direction dependency of rigidity when a crack occurs in the movable portion 40. FIG. 3A shows a case where the movable part 40 is driven in the positive direction of the X axis. In this case, the crack 41 is closed, and the rigidity of the link increases. On the other hand, FIG. 3B shows a case where the movable portion 40 is driven in the negative direction of the X-axis. In this case, the crack 41 spreads, and the rigidity of the link decreases.
Thus, when the crack 41 arises in the movable part 40, it has the direction dependence from which rigidity changes with a drive direction.

Next, characteristics of data at the time of crack occurrence obtained by operating the drive unit in a sine wave as a dedicated operation for crack detection will be described.
FIG. 4 shows the change over time of the effective value of the motor torque before and after the occurrence of cracks. In collecting data, the above-described dedicated operation is performed, and the dedicated operation is continuously performed in order to accelerate the generation of cracks.
Further, the effective value of the motor torque is calculated based on the time series data obtained from the motor torque detector 95, and the vertical axis divides the effective value of the motor torque by the rated torque of the servo motor 10. The horizontal axis shows the accumulated time from the start of the dedicated operation.

In the dedicated operation of this crack detection method, a sinusoidal wave position command having a constant amplitude is given, and the effective value of the motor torque is also constant if there is no abnormality in the machine elements and members.
According to FIG. 4, the effective value of the motor torque is almost flat up to about 80000 seconds in the dedicated operation accumulation time, and no abnormality is generated in the machine elements and members.

However, the effective value of the motor torque tends to decrease from about 80000 seconds after the dedicated operation accumulated time, and the effective value of the motor torque rapidly decreases after about 90000 seconds. There is a tendency for the value to saturate to a constant value.
As a result of investigating the periphery of the drive mechanism after the effective value of the motor torque decreased, a crack with the linear guide 30 as a boundary was found in the movable portion 40.

FIG. 5 shows the time series data of the instantaneous value of the motor torque shown in FIG. 4 in the upper stage, and shows the result of the FFT analysis on the time series data in the lower stage. FIG. FIG. 5B shows the result after the effective value is lowered.
According to FIG. 5A, the time-series data of the motor torque before the effective value of the motor torque is reduced has a constant triangular wave waveform whose amplitude is vertically symmetrical. The reason why such a triangular waveform is obtained is that it is a load torque depending on the stiffness that is directly affected by the load fixed to the hand as described above.
When this time-series data of the motor torque is subjected to an FFT analysis, it has a basic component (this is called a 1f component) that is the oscillation cycle of the movable part 40 and a component having a cycle that is three times (this is called a 3f component). I understand that.

On the other hand, according to FIG. 5B, the time series data of the motor torque after the decrease in the effective value of the motor torque has a triangular waveform similar to that before the decrease, but the asymmetry is different in the amplitude up and down. It turns out that it has sex. Further, when the time series data of the motor torque is subjected to an FFT analysis, it can be confirmed that in addition to the 1f component and the 3f component, a component having a period twice that of the basic component (referred to as a 2f component) is newly provided.
The increase in the 2f component due to the vertical asymmetry of the amplitude of the motor torque and the decrease in the effective value of the motor torque are due to the direction dependency of the rigidity when the crack is generated in the movable portion 40 described with reference to FIG. .

The crack detection method of the present invention is a method for determining that a crack has occurred when the occurrence and increase of the 2f component of the motor torque are detected by FFT analysis.
When a crack occurs, the effective value of the motor torque decreases. However, the evaluation method using this effective value may cause erroneous detection. For example, the friction that exists in the drive unit of a general drive mechanism changes even if the drive unit is normal, even if the lubricant temperature changes. When a sine wave position command having a constant amplitude is given, the effective value of the motor torque increases due to the increase in friction, and conversely, the effective value of the motor torque decreases as the friction decreases. Therefore, it is effective to determine the crack based on the generation and increase of the 2f component of the motor torque accompanying the asymmetry of the motor torque due to the crack.

Next, the content of the crack detection method of the drive mechanism of the present invention will be described based on the flowchart shown in FIG.
First, an instantaneous value of the motor torque from the start of the above-described dedicated operation is measured by the motor state quantity measurement step 121, and measurement data of a predetermined section is taken in a storage medium such as a hard disk as time series data 131. Note that the data sampling at that time is set sufficiently higher than the motion cycle of the sine wave drive and an appropriate antialiasing filter is incorporated in consideration of the observation band in order to prevent signal aliasing.

  Next, using the time series data 131 of the motor torque recorded in the motor state quantity measurement step 121, the time series FFT is calculated in the FFT calculation step 122. The analysis data 132, which is the calculation result of the FFT calculation step 122, is stored in a storage medium such as a hard disk, like the time series data 131 of the motor torque.

Next, it is determined whether or not the 2f component generated due to the vertical asymmetry of the motor torque is detected with reference to the analysis data 132 in the 2f component determination step 123. If there is no 2f component, the motor state quantity measurement is performed. Return to step 121.
On the other hand, if the 2f component of the motor torque is detected, it is determined that there is a crack, and the process proceeds to a crack determination step 124 for issuing an alarm to the user.

As described above, the crack detection method according to the first embodiment merely performs the FFT analysis on the time-series data for each fixed section using only the internal state quantity of the servo amplifier that is the motor control device, and the fluctuation of the friction of the machine element. Even if there is a decrease in the effective value due to, the presence or absence of a crack can be determined without erroneous detection.
Note that, as described above, the reduction in the effective value of the motor torque also occurs due to the variation in the friction of the machine elements constituting the positioning device. Even when the friction of the drive unit increases, the 2f component may increase. However, in the range of fluctuations in which no constant increase in the 2f component is observed, erroneous detection can be avoided by determining the detection of the 2f component when the 2f component continues to increase.

Embodiment 2. FIG.
In the first embodiment, it is determined that a crack has occurred based only on the FFT analysis result of the time series data of the motor torque. However, depending on conditions, there is a case where the friction constantly increases and the 2f component continues to increase even for reasons other than cracks. In this case, the second embodiment is a crack detection method that is free from false detection, and detects the 2f component detection and increase of the motor torque by FFT and sees a decrease in the effective value of the motor torque. This is a method for determining that a crack has occurred.

The content of the crack detection method of Embodiment 2 is demonstrated based on the flowchart shown in FIG.
First, the motor torque from the start of the dedicated operation is measured by the motor state quantity measuring step 121. The contents are the same as in the first embodiment.
Next, the effective value is calculated in the effective value calculating step 125 using the time series data 131 of the motor torque recorded in the motor state quantity measuring step 121. As in the first embodiment, the FFT calculation step 122 performs FFT calculation on the time series data 131. In the second embodiment, there is no restriction as to which of the two operations should be performed earlier. Further, the effective value data 133 calculated in the effective value calculation step 125 and the analysis data 132 in the FFT calculation step 122 are stored in a storage medium such as a hard disk, similar to the time series data 131 of the motor torque.

Next, it is determined whether or not the 2f component generated due to the vertical asymmetry of the motor torque is detected with reference to the analysis data 132 in the 2f component determination step 123. If the 2f component is not detected, the motor state The process returns to the quantity measurement step 121.
On the other hand, when the 2f component of the motor torque is detected, the process proceeds to the effective value determination step 126. Here, referring to the effective value data 133, the effective value of the motor torque stored in the past and the current motor torque are stored. The effective value is compared, and if it has decreased, it is determined that there is a crack, and the process proceeds to the crack determination step 124 where an alarm is given to the user, and the process ends. This determination is made when the slope between the immediately preceding effective value and the current effective value is greater than or equal to a certain threshold value, or when the gradient is below a threshold value that is a lower limit value of a predetermined effective value, etc. By the method.

  As described above, the 2f component associated with the vertical asymmetry of the motor torque also occurs when the effective value increases as compared with the past value, resulting in asymmetry. This symptom is mainly caused by an increase in friction of the mechanical elements constituting the positioning device. However, as described above, when a crack occurs, the effective value of the motor torque decreases. Therefore, the presence / absence of a crack can be determined by the effective value determination step 126.

  As described above, in the independent evaluation of the effective value or the FFT analysis result, even if the drive unit is normal, it may be erroneously detected due to the influence of friction that changes depending on the temperature. Since the combination evaluation of both, that is, the occurrence and increase of the 2f component due to FFT and the decrease in the effective value of the motor torque is detected, the occurrence of cracks is judged. It can be detected accurately without detection.

  As described above, according to the second embodiment, the drive mechanism crack detection method that gives a sine wave position command as a dedicated operation of the drive unit causes the vertical asymmetry due to the direction dependency of the mechanical rigidity of the motor torque. The detected 2f component is detected by FFT analysis, and the cause of this 2f component is determined to be a crack by lowering the effective value of the motor torque. It can be detected accurately.

  In the second embodiment, the crack detection method in which the effective value calculation is always performed is shown. This is because, considering the fact that there is a fluctuation in the friction of the drive unit due to temperature change as described above, it is more accurate to always calculate the effective value and determine the decrease compared to the previous value. . However, it is also possible to determine that the lower limit value of the effective value determined based on the data measured and stored in advance is set as a threshold value, and the threshold value is below the threshold value. In this case, it is more effective to prepare correction of data in consideration of conditions that cause fluctuations in effective values such as temperature in advance, or to perform correction at the time of determination.

  As described above, if the threshold value is set in advance, it is not necessary to always calculate the effective value, and calculation may be performed after the 2f component is detected by FFT analysis as shown in FIG. In this case, in the effective value determination step 126a, the current effective value of the effective value data 133 is compared with the preset threshold data 134.

Embodiment 3 FIG.
In the first and second embodiments, the crack detection method is described in which the effective value is compared with the past data when the 2f component is detected by the FFT analysis of the time series data of the motor torque. In Embodiment 2, a crack detection method is described in which the effective value of the motor torque is calculated, the obtained effective value is compared with past data, and determination is made by FFT when it is determined that the effective value is reduced.

FIG. 9 is a flowchart showing a crack detection method for the drive mechanism of the third embodiment.
The flowchart of FIG. 9 differs from that of FIG. 7 in that the effective value determination step 126 is provided before the 2f component determination step 123. This effective value determination step 126 gives a past effective value and a change amount of the effective value at the present time as a condition, and this determination includes the time between the immediately preceding effective value and the current effective value as described above. A method may be considered in which a determination is made when the gradient of change is equal to or greater than a certain threshold value, or a determination is made that the gradient is below a threshold value that is a predetermined lower limit value of the effective value.

When the time series data of the instantaneous value at which the motor torque decreases uniformly in the vertical direction is collected during the dedicated operation by the sine wave drive, the time series data is transferred from the effective value judgment step 126 to the 2f component judgment step 123. Since there is no vertical asymmetry, no 2f component is detected. Therefore, it can be determined that it is not due to the occurrence of cracks but due to factors such as drive unit friction.
On the other hand, when vertical asymmetry occurs in the series data at the instantaneous value and the process shifts from the effective value determination step 126 to the 2f component determination step 123, the 2f component is detected and it can be determined that a crack has occurred.

  As described above, the crack detection method in which the FFT analysis is always performed has been described. However, the FFT analysis is not always necessary, and as shown in FIG. 10, only the effective value of the motor torque is always calculated, and the motor torque With respect to the effective value, the calculation load can be reduced by performing the FFT calculation only when the above-described conditions are satisfied.

  In the crack detection method according to the third embodiment, cracks can be detected as in the second embodiment, but when the effective value calculation is always performed for purposes other than the detection of the crack, the calculation result is used. There are advantages you can do.

In the first to third embodiments, an example in which crack detection is performed by applying a sine wave drive as a dedicated operation has been described. However, a linear motion shaft such as a mounter, a machining center, or an electric discharge machine to be diagnosed. When applying to a machine tool with a, the machine tool itself is provided with a terminal dedicated to crack detection. This is a terminal for giving a sinusoidal drive as a dedicated operation in a state where the hand of an arbitrary linear motion shaft is fixed.
For example, when an arbitrary linear motion axis to be diagnosed always performs the dedicated operation, the crack detection method of the present invention can be applied as it is without interrupting the operation of the diagnosis object.
Also, even when the drive shaft does not perform the dedicated operation described above and operates in an arbitrary drive pattern, the operation is periodically interrupted and the hand is moved to the crack detection dedicated terminal described above to perform the dedicated operation. If the time series data of the motor torque can be collected by performing the above, it is possible to detect the presence or absence of a crack by applying the crack detection method of the present invention as it is.

On the other hand, when the diagnosis target is operating in an arbitrary drive pattern and the operation cannot be interrupted, the time series data of the motor torque during the work process is measured in advance for each work process constituting the drive pattern. Thus, an effective value may be obtained, and a lower limit value of the effective value in each work process may be determined based on the effective value. Then, when the effective value of the motor torque falls below this lower limit value in each work process, the operation of the machine tool is interrupted at that time and the hand is moved to the dedicated crack detection terminal to perform the dedicated operation. By doing so, the presence or absence of a crack can be detected by applying the crack detection method of the present invention.
In addition, it is possible to obtain a sufficient effect by subdividing a complicated work process and detecting the work process constituted by a simple operation by setting the lower limit as described above.

  Further, the present invention can also be applied to simple crack detection when a drive unit is incorporated in a machine tool to be newly started up.

  In the first to third embodiments, the crack detection method of the present invention has been described as being implemented by the DSP or CPU of the servo amplifier 90 having the motor torque detection unit 95. However, the personal computer 100 and other arbitrary calculations are described. It has been described that the time series data of the motor torque, the effective value calculation result, and the FFT analysis result can be stored in a storage medium such as a hard disk. It can be stored and used as well.

  Further, although a machine tool has been described as an example, the crack detection method of the present invention can be similarly applied to a drive mechanism capable of a dedicated operation by sinusoidal drive.

It is a child block diagram which shows an example of the drive mechanism of the machine tool which is the application object of the crack detection method of this invention. It is a block diagram which shows the motor control system of the drive mechanism of FIG. It is a schematic diagram which shows the direction dependence of mechanical system rigidity accompanying a crack generation. It is a figure which shows the time-dependent change of the effective value of the motor torque before and behind crack generation. It is a figure which shows the time series data of the instantaneous value of the motor torque before and behind crack generation, and its FFT analysis result. It is a figure which shows the crack detection method by Embodiment 1 of this invention. It is a figure which shows the crack detection method by Embodiment 2 of this invention. It is a figure which shows the crack detection method by Embodiment 2 of this invention. It is a figure which shows the crack detection method by Embodiment 3 of this invention. It is a figure which shows the crack detection method by Embodiment 3 of this invention.

Explanation of symbols

DESCRIPTION OF SYMBOLS 10 Motor 40 Movable part 41 Crack 121 Motor state quantity measurement step 122 FFT calculation step 123 2f component determination step 124 Crack determination step 125 Effective value calculation step 126, 126a Effective value determination step

Claims (5)

A motor state quantity measuring step in which an instantaneous value of the motor torque is measured as time-series data when the motor drives the movable part with a motor torque that changes according to a sine wave function of a predetermined period;
FFT calculation step for performing Fourier transform on the time series data of the instantaneous value of the motor torque,
A 2f component determination step for determining whether or not there is an increase in the 2f component, which is a frequency component twice as long as the result of the Fourier transform,
And a crack detection method for a drive mechanism, comprising: a crack determination step for determining that the movable part has a crack when the conditional branching means determines that the 2f component is present. A motor state quantity measuring step in which an instantaneous value of the motor torque is measured as time-series data when the motor drives the movable part with a motor torque that changes according to a sine wave function of a predetermined period;
An effective value calculating step for calculating an effective value of the motor torque from time-series data of instantaneous values of the motor torque;
FFT calculation step for performing Fourier transform on the time series data of the instantaneous value of the motor torque,
A 2f component determination step for determining whether or not there is an increase in the 2f component, which is a frequency component twice as long as the result of the Fourier transform,
When it is determined by the conditional branching means that the 2f component is present, the effective value of the motor torque is compared by comparing the current value with a value obtained by going back a predetermined time from the effective value of the motor torque. An effective value determination step for determining whether or not there is a decrease,
And a crack detection step for determining a crack in the movable portion when it is determined in the effective value determination step that there is a decrease in the effective value of the motor torque. Method. A motor state quantity measuring step in which an instantaneous value of the motor torque is measured as time-series data when the motor drives the movable part with a motor torque that changes according to a sine wave function of a predetermined period;
FFT calculation step for performing Fourier transform on the time series data of the instantaneous value of the motor torque,
A 2f component determination step for determining whether or not there is an increase in the 2f component, which is a frequency component twice as long as the result of the Fourier transform,
When it is determined by the conditional branching means that the 2f component is present, an effective value calculating step for calculating the effective value of the motor torque from the time series data of the instantaneous value of the motor torque. An effective value determination step for comparing the current value and the current effective value to determine whether or not the effective value of the motor torque has decreased,
And a crack detection step for determining a crack in the movable portion when it is determined in the effective value determination step that there is a decrease in the effective value of the motor torque. Method. A motor state quantity measuring step in which an instantaneous value of the motor torque is measured as time-series data when the motor drives the movable part with a motor torque that changes according to a sine wave function of a predetermined period;
An effective value calculating step for calculating an effective value of the motor torque from time-series data of instantaneous values of the motor torque;
FFT calculation step for performing Fourier transform on the time series data of the instantaneous value of the motor torque,
An effective value determination step of comparing the current value with a value at a time point back from the effective value of the motor torque to determine whether the effective value of the motor torque has decreased,
When it is determined in this effective value determination step that the effective value of the motor torque is decreased, 2f is determined from the result of the Fourier transform to determine whether or not the 2f component that is a frequency component twice the period is increased. Component determination step,
And a crack detection method for a drive mechanism, comprising: a crack determination step for determining that the movable part has a crack when the conditional branching means determines that the 2f component is present. A motor state quantity measuring step in which an instantaneous value of the motor torque is measured as time-series data when the motor drives the movable part with a motor torque that changes according to a sine wave function of a predetermined period;
An effective value calculating step for calculating an effective value of the motor torque from time-series data of instantaneous values of the motor torque;
An effective value determination step of comparing the current value with a value at a time point back from the effective value of the motor torque to determine whether the effective value of the motor torque has decreased,
When it is determined in this effective value determination step that the effective value of the motor torque has decreased, an FFT calculation step for performing Fourier transform on the time series data of the instantaneous value of the motor torque,
2f component determination step for determining whether or not there is an increase in 2f component, which is a frequency component twice as long as the result of the Fourier transform,
And a crack detection method for a drive mechanism, comprising: a crack determination step for determining that the movable part has a crack when the conditional branching means determines that the 2f component is present.

Images