JP5516682B2 - Vibration control device for feedback control system and motor control device provided with vibration detection device - Google Patents

Description

  The present invention relates to a vibration detection device that detects vibration in a motor control device that performs feedback control of a load machine, and a motor control device that includes the vibration detection device.

  Normally, in a motor control device that performs feedback control, the load position or speed is made to follow the target command at high speed and with high accuracy, so adjustment of the position or speed control unit gradually increases the corresponding control gain and monitors its vibration state. While done. At the time of this adjustment, if the control gain is excessively set, the control system may oscillate due to mechanical system resonance and dead time, and sound may be generated or the machine may be destroyed in some cases. Therefore, in order to make a safe adjustment, it is necessary to detect the vibration state of the control system early before oscillation.

  In the first prior art, a speed observer is configured, and an oscillation / vibration state is detected based on a difference signal between an actual speed and an estimated speed (see, for example, Patent Document 1).

  In the second prior art, each vibration amplitude value of the torque command within the detection range and each time between the lower limit peak and the upper limit peak in the vibration are detected, and an energy value for each vibration is calculated from the detected value, The number of vibrations of the torque command in the detection range is set to N, each integrated value of vibration energy values from the first time to each number is calculated, the sum of the respective integrated values is calculated, and this is calculated from the first time to the Nth time. A comparison is made with a value obtained by multiplying the integrated value by N / 2, and if the sum of the integrated values is smaller, it is detected as oscillation (see, for example, Patent Document 2).

  FIG. 6 is a control block diagram showing an embodiment of the first prior art. In the figure, 101 is the speed control, 102 is the motor and load, 111 is the motor and load inertia of the observer, 112 is the observer speed control gain, 114 is the integral, 113 is the observer speed control integral, and 103 is the observer speed control. The whole. When it is observed that the difference between the estimated speed ωob of the observer and the actual speed ω exceeds a certain detection level, oscillation / vibration is detected.

  FIG. 7 is a control block diagram showing an embodiment of the second prior art. In the figure, 201 is a speed controller, 202 is a current controller, 203 is a motor, 204 is a mechanical load, 205 is an encoder, 206 is a differentiator, 207 is an oscillation detector, 208 is a memory, 209 is a subtractor, Focusing on the fact that the amplitude value of the torque command increases when oscillation occurs, the vibration energy value is expressed by the amplitude value of the torque command Tref and the time from the peak to the peak of the amplitude value (the time from the lower limit peak to the upper limit peak). It is defined by a weighted function en (i) to detect oscillation. The oscillation detector 207 receives the torque command Tref to detect oscillation. When oscillation is detected, the oscillation signal is output to the memory 208. The memory 208 receives the oscillation signal and stores it in advance from the speed controller 201. The pre-oscillation control parameter is written in the control parameter of the speed controller 201 and changed.

Japanese Patent Laying-Open No. 2004-248386 (page 2-3, FIG. 3) Japanese Patent Laying-Open No. 2004-187439 (page 3-5, FIG. 1)

  In the first prior art, vibration is detected based on the difference between the estimated speed ωob of the observer using the model to be controlled and the actual speed ω, so that the parameter of the controlled object, particularly the value of the moment of inertia of the motor or load is If there is a deviation, the difference between the observer's estimated speed ωob and the actual speed ω in the transient process becomes large even if vibration does not occur, so that there is a problem that oscillation and vibration cannot be detected correctly. Here, the deviation of the value of the moment of inertia occurs when there is an identification error or a load fluctuation, and is present in most cases in a motor control device that controls a motor.

  In the second prior art, the oscillation state is determined based on whether the average value of vibration energy within a certain detection time range is smaller than the vibration energy at the time of determination. There was a point.

  The present invention has been made in view of such problems, and even when the value of the moment of inertia of the motor or load deviates, the vibration state is accurately detected, and abnormalities are quickly detected to detect noise, damage to mechanical systems, etc. And a vibration detection device and a vibration detection device of a feedback control system capable of accurately detecting a vibration peak even for a signal in which noise is superimposed or a signal in an oscillation state An object is to provide a motor control device.

  In order to solve the above problem, the present invention is configured as follows.

  The present invention is a vibration detection device for a control device that controls a motor connected to a load, the control device including a controller that controls the motor by feedback controlling the control target, and the vibration detection device. The device generates vibrations based on the upper and lower limit peak values obtained by subtracting the immediately preceding lower limit peak value from the upper limit peak value at the detection time of the upper limit peak value of the output signal of the control device and their occurrence times. A vibration state determination unit for determining is provided.

  In the present invention, the vibration state determination unit may generate vibration based on the occurrence time of the upper and lower limit peaks obtained by subtracting the detection time of the lower limit peak value of the output signal from the detection time of the upper limit peak value of the output signal. Is to judge.

  Further, the present invention, the vibration state determination unit, the vibration intensity value calculated based on the upper and lower limit peak values and their occurrence time, the vibration moving average intensity value calculated by moving average the vibration intensity value, The vibration moving average intensity value is a vibration duration that is a time width exceeding a preset vibration intensity minimum value, and a preset value corresponding to each of the vibration intensity value, the vibration moving average intensity value, and the vibration duration time. The generation of the oscillation of the output signal of the control device is determined based on the comparison with the threshold value.

  Further, the present invention is a vibration detection device of a control device that controls a motor connected to a load, the control device comprising a controller that controls the motor by feedback controlling the control target, The vibration detection device defines a peak detection interval of the output signal of the control device by a left-hand interval with a horizontal axis as a time axis, a right-hand interval, and a center point included in both of the intervals, and When the value of the output signal is larger than the value of the output signal in the left section and the right section, the center point is determined as the upper limit peak, and occurrence of vibration is determined.

  Further, the present invention is a vibration detection device of a control device that controls a motor connected to a load, the control device comprising a controller that controls the motor by feedback controlling the control target, The vibration detection device defines a peak detection interval of the output signal of the control device by a left-hand interval with a horizontal axis as a time axis, a right-hand interval, and a center point included in both of the intervals, and When the value of the output signal is smaller than the value of the output signal in the left section and the right section, the center point is determined as the lower limit peak, and occurrence of vibration is determined.

  Further, the present invention is a vibration detection device of a control device that controls a motor connected to a load, the control device comprising a controller that controls the motor by feedback controlling the control target, The vibration detection device defines a peak detection interval of the output signal of the control device by a left interval, a right interval and a center point included in both intervals with the horizontal axis as a time axis. Output signal value C, maximum value MaxL and minimum value MinL of the left section, maximum value MaxR and minimum value MinR of the right section, difference ΔL between the maximum value MaxL and the minimum value MinL, maximum value MaxR and the minimum If the difference ΔR of the value MinR is defined and both the ΔL and ΔR are larger than the threshold value and the maximum values MaxL, MaxR and the output signal value C are equal, the center point is determined as the upper limit peak, and the maximum value MaxL, One of MaxR If the minimum value MinL, MinR and the output signal value C are greater than the signal value C, the center point is determined as the lower limit peak, and the output signal value C must be equal to both the minimum value MinL and MinR. For example, it is determined that there is no upper limit peak or lower limit peak in the peak detection section, and occurrence of vibration is determined.

  In the present invention, the peak detection interval of the output signal of the control device is set as a set of N elements in advance, and peak detection is performed in the peak detection interval that is one element of the set, and at the center point Until the value of the output signal is determined to be the upper limit peak or the lower limit peak, the N peak detection intervals are sequentially determined to determine the occurrence of vibration.

According to the present invention, even when there is no model to be controlled, it is possible to accurately detect the vibration state even if the value of the moment of inertia of the motor or the load is shifted. Further, it is possible to quickly detect a damped vibration and an unstable vibration that are large due to mechanical resonance, and it is also possible to quickly detect an unstable vibration due to a noise and a control system delay. Furthermore, the vibration frequency can be detected using the time interval between peaks.
In addition, according to the present invention, it is possible to quickly detect an abnormality before the complete oscillation is reached and to prevent noise or damage to the mechanical system.

  In addition, according to the present invention, accurate peak detection can be performed even if noise is superimposed on the output signal or vibration components including a wide range of frequency components are present in the output signal.

  In addition, according to the present invention, an abnormal stop can be accurately stopped based on the determination of occurrence of oscillation or vibration by the vibration detection device.

1 is a block diagram showing the configuration of a feedback system vibration detection apparatus in Embodiment 1 of the present invention. Waveform diagram showing a difference signal between the output signal and the estimated output signal in the first embodiment The figure which shows the Vpp signal calculated by step 102 in Example 1. FIG. The figure which shows P signal calculated by step 103 in Example 1. The flowchart which shows the process sequence which judges the vibration state in Example 1. FIG. Control block diagram showing an embodiment of the first prior art Control block diagram showing an embodiment of the second prior art The figure explaining the method of searching the amplitude peak of the output vibration component signal 15 The flowchart which shows the process sequence which searches the amplitude peak of the output vibration component signal 15 The block diagram which shows the structure of the vibration detection apparatus of the feedback system in Example 2 of this invention. Waveform diagram showing output signal 4 in vibration state 7 is a flowchart illustrating a processing procedure for determining a vibration state in the second embodiment. The figure which showed amplitude peak Vp of the output signal 4, and time tp when a peak value appears The figure which shows Vpp calculated at step 202 in Example 2. The figure which shows App calculated at step 203 in Example 2. The figure which shows Av calculated at step 205 The figure which shows Tv calculated at step 207 The block diagram which shows the structure of the vibration detection apparatus of the feedback system in Example 3 of this invention. Waveform diagram showing output signal 4 in the third embodiment Flowchart showing a processing procedure for detecting a peak in the third embodiment. Flowchart showing peak detection processing executed in step 403 The flowchart which shows the peak judgment process implemented at step 503

  Hereinafter, embodiments of the present invention will be described with reference to the drawings.

  FIG. 1 is a block diagram showing a configuration of a feedback system vibration detection apparatus in Embodiment 1 of the present invention. In the figure, 1 is a command signal, 12 is a feedback control device, comprising a controller 2 and a control object 3, 4 is an output signal, 5 is an output feedback signal, 13 is a model part of the feedback control device, The model unit 6 of the controller 2 and the model unit 7 of the controlled object 3 are provided, 8 is an estimated output signal, 9 is an estimated output feedback signal, 10 is a difference signal between the output signal 4 and the estimated output signal 8, and 14 is a low-pass filter. , 15 is an output vibration component signal, and 11 is a vibration state determination unit. That is, vibration detection is performed based on a difference signal between the output signal 4 of the feedback control device 12 and the estimated output signal 8 of the model unit 13 of the feedback control device.

  Here, when this feedback system vibration detection device is applied to a motor control device, when the feedback control device 12 is a position control system, the command signal 1 is a position command, the controller 2 is a position control or speed controller, When the feedback control device 12 is a speed control system, the command signal 1 is a speed command, the controller 2 is a speed controller, and the output signal 4 is a speed. The detection signal or the position difference signal and the estimated output signal 8 are speed estimation signals.

  FIG. 2 is a waveform diagram showing a difference signal between the output signal 4 and the estimated output signal 8 in FIG. In the figure, the horizontal axis is the time axis, the vertical axis is the signal amplitude, and 10 is the difference signal between the output signal 4 and the estimated output signal 8. In the case where the feedback system vibration detection device in FIG. 1 is applied to a motor control device, this waveform diagram shows feedback control when the control gain is adjusted and the gain of the controller 2 is increased so that the motor drives the load. The apparatus 12 vibrates and the vibration state appears in the difference signal 10.

  In general, when the control gain is increased until the feedback control system is in an oscillating state, the output signal 4 substantially follows the command signal 1 in the low and medium frequency region of the command signal 1, so the model to be controlled The output signal 4 and the estimated output signal 8 that follow the same command signal 1 are components in the low / medium frequency region even if there is some error in the unit 7 (for example, a deviation in the value of the moment of inertia of the motor or load). Are almost the same. Therefore, the difference signal 10 contains almost only high-frequency vibration components. Further, the difference signal 10 is passed through the low-pass filter 14 and a noise component having a higher frequency than the vibration component is cut, and then input to the vibration state determination unit 11. Here, the time constant of the low-pass filter 14 is set so that the cutoff frequency of the filter is higher than the vibration frequency.

  Thus, in order not to be affected by the model error of the control target (for example, the deviation of the value of the moment of inertia of the motor or load), it is necessary to use the vibration detection device of the present invention with the control gain increased to some extent. There is. In a motor control device, it is generally used in a state where the control gain is increased in order to improve the follow-up performance of the control response.

  Next, a processing procedure for determining the vibration state in the vibration state determination unit 11 in FIG. 1 will be described with reference to the flowchart of FIG. However, the following description is based on the assumption that the vibration state determination unit 11 works at a constant sampling period.

(Step 101)
The amplitude peak of the output vibration component signal 15 is searched, and the process proceeds to step 102.
Here, searching for the amplitude peak of the output vibration component signal 15 means, for example, that the vibration state determination unit 11 monitors the amplitude of the output signal for each sampling period, and the amplitude value of the output signal for several sampling periods. Is to detect and detect the upper limit peak value or the lower limit peak value. A specific search method will be described later with reference to the flowchart shown in FIG.

(Step 102)
At the time when the upper limit peak is detected, a value obtained by subtracting the immediately preceding lower limit peak value from the upper limit peak value is set as the value of Vpp. . However, in order to remove noise, even if the absolute value of Vpp is not 0, the value of Vpp is set to 0 if it is smaller than a certain small value (generally the detection resolution of the output signal 4).

  FIG. 3 is a diagram showing a difference signal Vpp20 between the upper limit peak and the lower limit peak calculated in step 102 based on the difference signal 10 in FIG. Since the vibration period is generally much shorter than the transient time, the transient component can be further removed and the vibration component extracted by taking the difference between the upper limit peak and the lower limit peak.

(Step 103)
During a certain time Tn, the square moving average value P of the difference signal Vpp20 is calculated, and the process proceeds to Step 104. However, Tn is given as a value that is several times the vibration period and smaller than the period of the command signal 1.

  FIG. 4 is a diagram showing the P signal 30 calculated in step 103 based on the difference signal Vpp20 in FIG. The P signal 30 represents the density of vibration and is closely related to the level of noise and mechanical breakdown.

(Step 104)
It is compared whether the P signal 30 is smaller than a predetermined threshold value Pm. If so, go to Step 107.

  Further, the value of the threshold value Pm is determined by experiment or simulation in consideration of noise and not breaking the machine.

(Step 105)
Based on the P signal 30 created in step 103, the vibration duration Tv is counted as follows. When the P signal 30 becomes larger than a certain small value P0 (previously given vibration minimum value), the vibration duration Tv starts to be counted, and when it becomes smaller, the vibration duration Tv is reset to zero. P0 is determined by experiment or simulation in consideration of the influence of noise.

(Step 106)
It is compared whether or not the vibration duration time Tv is smaller than a predetermined threshold value Tvm. If so, go to Step 107.

  In general, it can be said that after the external stimulus disappears, the longer the vibration duration Tv, the closer to complete oscillation. The threshold value Tvm is determined in consideration of the increase in control gain and safety. Tvm should be set small when the amount of increase in the control gain is large and absolutely no oscillation should occur.

(Step 107)
Judge as vibration and proceed to action to stabilize the control system. That is, for example, the control system is stabilized by an algorithm such as vibration suppression control.

  Thus, based on the difference signal 10 between the output signal 4 of the feedback control device and the estimated output signal 8 of the model unit 7 of the feedback control device, the amplitude peak is searched, and the difference signal Vpp20 between the upper limit peak and the lower limit peak. By calculating the vibration component, it is possible to accurately extract the vibration component without being affected by the model error (for example, deviation of the value of the moment of inertia of the motor or the load) and the transient characteristic. Further, by calculating an amount (P signal 30) representing the density of vibration from the difference signal Vpp20 between the upper limit peak and the lower limit peak to detect the vibration duration Tv and judging the vibration state based on the information, Since the vibration state can be determined before full oscillation occurs, it is possible to quickly detect an abnormality and prevent noise and damage to the mechanical system.

  FIG. 8 is a diagram for explaining a method of searching for the amplitude peak of the output vibration component signal 15. In the figure, the horizontal axis is the time axis, the vertical axis is the signal amplitude, and 15 is the output vibration component signal. FIG. 9 is a flowchart showing a processing procedure for searching for the amplitude peak of the output vibration component signal 15. Hereinafter, a method for searching for the amplitude peak of the output vibration component signal 15 will be described with reference to the flowchart shown in FIG.

(Step 301)
First, the number of times for setting a section for sampling the data of the output vibration component signal 15 is set to r (r = 0, 1, 2, 3,...), And r = 0 is set to an initial value, Proceed to 302.

(Step 302)
An interval of 2p + 1 data points A = [x _{1 + r} , x _{2 + r} ,..., X _{2p + 1 + r} ] is set based on the number of times r, and the process proceeds to step 303.
Here, p is a preset positive integer (p = 1, 2, 3, 4,...), And the frequency of the vibration component in which the peak is correctly detected decreases as p increases.

(Step 303)
The maximum value and the minimum value of the output vibration component signal 15 in the section A = [x _{1 + r} , x _{2 + r} ,..., X _{2p + 1 + r} ] set in step 302 are calculated. move on.

(Step 304)
It is checked whether the signal amplitude at the midpoint x _{p + 1 + r} of the section set in step 302 is the upper limit peak value. The signal amplitude of the midpoint x _{p + 1 + r} is compared with the maximum value calculated in step 303. If the signal amplitude of the midpoint x _{p + 1 + r} is the maximum value of the section, the process proceeds to step 305. On the other hand, if the signal amplitude at the midpoint x _{p + 1 + r} is not the maximum value of the section, the process proceeds to step 306.

(Step 305)
The signal amplitude of the section value x _{p + 1 + r} (midpoint) set in step 302 is set as the upper limit peak value, and the process proceeds to step 307.

(Step 306)
It is checked whether the signal amplitude at the midpoint x _{p + 1 + r} of the section set in step 302 is the lower limit peak value. The signal amplitude of the midpoint x _{p + 1 + r} is compared with the minimum value calculated in step 303. If the signal amplitude of the midpoint x _{p + 1 + r} is the minimum value of the section, the process proceeds to step 308. On the other hand, if the signal amplitude at the midpoint x _{p + 1 + r} is not the minimum value of the section, the process proceeds to step 310.

(Step 307)
The time at which the upper limit peak appears (the time of the middle point x _{p + 1 + r} of the section set in step 302) is stored in the memory, and the process proceeds to step 310.

(Step 308)
The signal amplitude of the section value x _{p + 1 + r} (midpoint) set in step 302 is set as the lower limit peak value, and the process proceeds to step 309.

(Step 309)
The time when the lower limit peak appears (the time of the middle point x _{p + 1 + r} of the section set in step 302) is stored in the memory, and the process proceeds to step 310.

(Step 310)
Add 1 to the value r to update it as a new r (r = r + 1), and go to Step 302.

  As described above, steps 302 to 310 are repeated until the upper and lower limit peak values of the output vibration component signal 15 are detected.

FIG. 10 is a block diagram showing a configuration of a feedback system vibration detection apparatus according to Embodiment 2 of the present invention. In the figure, components having the same action as in FIG.
The difference between the second embodiment and the configuration of the vibration detection device of the feedback system in the first embodiment is that there is no configuration of the model unit 13 and the low-pass filter 14 of the feedback control device in FIG. This is not performed based on the difference signal between the output signal 4 of the above and the estimated output signal 8 of the model unit 13 of the feedback control device.
That is, in the second embodiment, a new vibration state determination unit 51 that can detect vibration based on the output signal 4 of the feedback control device 12 even when there is no model unit 13 of the feedback control device is provided in the configuration. It is a thing.

  Here, when this feedback system vibration detection device is applied to a motor control device, when the feedback control device 12 is a position control system, the command signal 1 is a position command, the controller 2 is a position control or speed controller, The output signal 4 is a position detection signal. When the feedback control device 12 is a speed control system, the command signal 1 is a speed command, the controller 2 is a speed controller, and the output signal 4 is a speed detection signal or a position difference signal.

  FIG. 11 is a waveform diagram showing the output signal 4 in FIG. In the figure, the horizontal axis is the time axis, the vertical axis is the signal amplitude, and 4 is the output signal. In the case where the feedback system vibration detection device shown in FIG. 10 is applied to the motor control device, this waveform diagram shows feedback when the control gain is increased and the gain of the controller 2 is increased so that the motor drives the load. The control device 12 vibrates and the vibration state appears in the output signal 4.

  Next, a processing procedure for determining the vibration state in the vibration state determination unit 51 in FIG. 10 will be described with reference to the flowchart of FIG. However, the following description is based on the assumption that the vibration state determination unit 51 operates at a constant sampling period.

(Step 201)
The amplitude peak Vp of the output signal 4 is searched and the time tp at which the peak value appears is stored in a memory (not shown) or the like, and the process proceeds to step 102.
Here, the search for the amplitude peak Vp of the output signal 4 means that, for example, the vibration state determination unit 51 monitors the amplitude of the output signal 4 for each sampling period, and the amplitude value of the output signal 4 for several sampling periods. Is to detect and detect the upper limit peak value or the lower limit peak value.
The time tp at which the peak value appears is the time when the upper limit peak value or the lower limit peak value is detected by the above-described search.

  FIG. 13 is a diagram showing the peak (Vp) (including the upper and lower limit peaks) of the output signal 4 calculated in step 201 based on the output signal 4 in FIG. In the figure, t1, t2, and tp indicate the time at which the peak appears.

(Step 202)
At the time when the upper limit peak is detected, a value obtained by subtracting the immediately preceding lower limit peak value from the upper limit peak value is set as the value of Vpp, and at the same time, a time difference Tpp corresponding to the lower limit peak is calculated from the time corresponding to the upper limit peak. At the time when the upper limit peak is not detected, the value of Vpp is set to 0 and the process proceeds to step 203. However, in order to remove noise, even if the absolute value of Vpp is not 0, the value of Vpp is set to 0 if it is smaller than a certain small value (generally the detection resolution of the output signal 4).
FIG. 14 is a diagram showing the difference signal Vpp from the upper limit peak to the lower limit peak calculated in step 202 based on the output signal 4 in FIG. Since the vibration period is generally much shorter than the transient time, the transient component can be further removed and the vibration component extracted by taking the difference between the upper limit peak and the lower limit peak.

(Step 203)
A value obtained by dividing the difference signal Vpp value between the upper limit peak and the lower limit peak calculated in step 202, which is not zero, by the time interval Tpp between the corresponding upper and lower limit peaks is calculated as the vibration intensity App, and the process proceeds to step 104.
The vibration intensity App defined here is the amplitude value per unit time of the difference signal Vpp value (FIG. 14) from the upper limit peak to the lower limit peak based on the vibration state waveform of the output signal 4 (FIG. 11). It is expressed as strength.
FIG. 15 is a diagram showing the vibration intensity signal App calculated in step 203 based on the difference signal Vpp in FIG.

(Step 204)
The vibration intensity signal App obtained in step 203 is compared with a predetermined threshold value Appm. If the vibration intensity signal App is smaller than the threshold value Appm, the process proceeds to step 205. On the other hand, if larger, the process proceeds to step 209.
Here, for example, in the case where the threshold value Appm given in advance is applied to the motor control device of the vibration detection device of the feedback system in FIG. 10, the load machine driven by the motor control device is not damaged or the noise is not large. Alternatively, it may be determined to a value within the positioning accuracy allowed at the time of positioning.

(Step 205)
The moving average value Av of the vibration intensity signal App is calculated, and the process proceeds to step 206. The value of the fixed time Ta required for the moving average of the signal is several times the maximum period of the vibration to be detected.
Here, the moving average value Av of the vibration intensity signal App indicates the level of noise and mechanical breakdown, and emphasizes high frequency components.
FIG. 16 is a diagram showing the vibration moving average intensity signal Av calculated in step 205 based on the vibration intensity signal App in FIG.

(Step 206)
The vibration moving average intensity signal Av obtained in step 205 is compared with a predetermined threshold value Avm. If the vibration moving average intensity signal Av is smaller than the threshold value Avm, the process proceeds to step 107. On the other hand, if larger, the process proceeds to step 209.
Here, the threshold value Avm given in advance is, for example, a value that does not damage the load machine or the like driven by the motor control device in the case of application to the motor control device of the vibration detection device of the feedback system in FIG. Alternatively, it may be determined to a value within the positioning accuracy allowed at the time of positioning.

(Step 207)
Based on the vibration moving average intensity signal Av obtained in step 205, the vibration duration Tv is counted as follows. When the vibration moving average intensity signal Av becomes larger than a certain small value P0 (previously given vibration minimum value), the vibration duration Tv starts to be counted, and when it becomes smaller, the vibration duration Tv is reset to zero. P0 is determined by experiment or simulation in consideration of the influence of noise. Here, the vibration duration Tv is a time when the vibration moving average intensity signal Av is not continuously zero.
FIG. 17 is a diagram showing the vibration duration Tv calculated in step 207 based on the vibration moving average intensity signal Av in FIG.

(Step 208)
The vibration duration time Tv is compared with a predetermined threshold value Tvm. If the vibration duration time Tv is smaller than the threshold value Tvm, the process proceeds to step 201. On the other hand, if larger, the process proceeds to step 209.
Generally, since it can be said that the longer the vibration duration Tv is, the closer to complete oscillation after the external stimulus is eliminated, the predetermined threshold value Tvm is determined in consideration of the increase in control gain and safety, and the control is performed. When the gain increase range is large and absolutely no oscillation should occur, the threshold value Tvm given in advance should be set small.

(Step 209)
Judge as vibration and proceed to action to stabilize the control system. That is, for example, the control system is stabilized by an algorithm such as vibration suppression control.

  As described above, since the vibration state determination unit 51 that executes the processing procedure of steps 201 to 208 is provided, vibration detection is performed based on the output signal 4 of the feedback control device 12 even when the model unit 13 of the feedback control device is not provided. It can be done. Further, as in the first embodiment, since the vibration state can be determined before the complete oscillation is reached, it is possible to quickly detect an abnormality and prevent noise and mechanical system damage.

FIG. 18 is a block diagram showing a configuration of a feedback system vibration detecting apparatus according to Embodiment 3 of the present invention. In the figure, components having the same action as in FIG.
The difference between the third embodiment and the configuration of the vibration detection device of the feedback system in the second embodiment is that the vibration state determination unit 51 in FIG. 10 is replaced with a new vibration state determination unit 52.
The vibration state determination unit 52 is obtained by replacing the peak detection process of the output signal 4 of the feedback control device 12 in the vibration state determination unit 51 with the following process.

  FIG. 19 is a waveform diagram showing the output signal 4 in FIG. In the figure, the horizontal axis is the time axis, the vertical axis is the signal amplitude, and 4 is the output signal. 40 is a section in which peak detection is performed by one peak detection process, and 41 is an entire section in which peak detection is performed.

  FIG. 20 is a flowchart showing a processing procedure for peak detection in the feedback vibration detection apparatus. The method of the present invention will be described step by step with reference to this figure.

  A method for detecting the upper and lower peak values of the output signal 4 performed in the entire peak detection section 41 will be described.

(Step 401)
First, in order to set the entire peak detection section 41 for detecting the peak of the output signal 4, i = 1 is set as an initial value, and the process proceeds to step 402.

(Step 402)
Data sampling is performed in the set peak detection entire section 41. The sampled data is stored in X = [x (i−2p (N)), x (i−2p (N) +1),..., X (i)], and the process proceeds to step 403.
Here, the set P (P = [p (1), p (2),..., P (N)]) is a set that is set in advance to set an interval for peak detection, and each element of the set P Is a positive integer. Further, p (1) <p (2) <... <P (N), N is the number of elements of the set P.

  The frequency fM of the vibration component that can accurately detect the peak has the relationship of the equation (1) between the number of data points RM in the detection section and the data sampling period Ts. Furthermore, the number of data points required for accurate peak detection differs depending on the vibration frequency. If the necessary data points of the vibration frequencies fM and fm are RM and Rm, respectively, and fM> fm, RM and Rm are ( Since there is a relationship of the expression (2), it is effective to determine the number of sampling data points in the vibration detection section based on the frequency component included in the output signal 4.

(Step 403)
Using the data sampled in step 402, the peak detection process in the entire peak detection section 41 is performed, and the process proceeds to step 404.
The peak detection process in the entire peak detection section 41 will be described later using the flowchart shown in FIG.

(Step 404)
The entire peak detection section 41 is updated (i = i + 1), and the process proceeds to step 402.

  As described above, steps 402 to 404 are repeated to detect the upper and lower peak values of the output signal 4.

  Next, the peak detection process in the entire peak detection section 41 performed in step 403 will be described using the flowchart shown in FIG.

(Step 501)
In step 501, in order to set the peak detection section 40 to the first section, j = 1 is initialized, and the process proceeds to step 502.

(Step 502)
Based on the values i and j, the data X stored in step 402 is converted into L = [x (ip (N) −p (j)), x (ip (N) −p (j) +1). ,…, X (ip (N))] and R = [x (ip (N)), x (ip (N) +1),…, x (ip (N) + p (j)) ] To the right section, and proceed to step 503.
The center point x (ip (N)) of the entire peak detection section 41 is divided so as to be included in both the left section L and the right section R.

(Step 503)
A determination process is performed to determine whether the center point x (ip (N)) of the peak detection section 40, which is a common element in the left and right sections divided in step 502, is a peak, and the process proceeds to step 504.
In addition, the peak determination process implemented here is later mentioned using the flowchart shown in FIG.

(Step 504)
If it is determined that the center point x (ip (N)) has a peak, the process proceeds to step 507, and if it is not determined to be a peak, the process proceeds to step 505.

(Step 505)
Whether or not the peak detection processing has been performed over the entire peak detection section 41 is determined by comparing the values j and N. If the peak detection processing has been performed N times, the process proceeds to step 507. Otherwise, the process proceeds to step 506.

(Step 506)
The value j is updated (j = j + 1), and the process proceeds to Step 502.

(Step 507)
The peak detection process ends.
As described above, step 501 to step 506 are repeated to perform peak detection of the output signal 4 over the entire peak detection section 41.

  Next, the peak determination process performed in step 503 will be described using the flowchart shown in FIG.

(Step 601)
Left section L = [x (ip (N) −p (j)), x (ip (N) −p (j) +1),..., X (ip (N))] divided in step 502 Calculate the maximum and minimum values in the right section R = [x (ip (N)), x (ip (N) +1), ..., x (ip (N) + p (j))] Proceed to step 602.

(Step 602)
Using the value calculated in step 601, the difference between the maximum value and the minimum value in the left section L and the threshold value H are compared, and the difference between the maximum value and the minimum value in the right section R is compared with the threshold value H. If both the differences are greater than H, go to step 603, otherwise go to step 607.
Note that the value of the threshold value H is the maximum value of noise included in the output signal 4, and noise having a magnitude less than the threshold value H can be removed from the output signal 4 by the processing in this step.

(Step 603)
In step 502, it is checked whether or not the center point x (ip (N)) of the peak detection section 40 included in both the left and right sections of L and R is the upper limit peak.
First, the L section maximum value calculated in step 601 is compared with x (ip (N)), the R section maximum value is compared with x (ip (N)), and the maximum of the left and right sections of L and R is compared. If the value is x (ip (N)), the process proceeds to step 604; otherwise, the process proceeds to step 605.

(Step 604)
The common element x (ip (N)) in the left and right sections is set as the upper limit peak, and the process proceeds to step 608.

(Step 605)
In step 502, it is checked whether or not the center point x (ip (N)) of the peak detection section 40 that is included in both the left and right sections of L and R is the lower limit peak.
First, the L section minimum value calculated in step 601 is compared with x (ip (N)), the R section minimum value is compared with x (ip (N)), and the L and R left and right section minimums are compared. If the value is x (ip (N)), the process proceeds to step 606; otherwise, the process proceeds to step 607.

(Step 606)
The common element x (ip (N)) in the left and right sections is set as the lower limit peak, and the process proceeds to step 608.

(Step 607)
The common element x (ip (N)) in the left and right sections is not a peak, and the process proceeds to step 608.

(Step 608)
Here, the peak determination routine ends.
As described above, it is determined whether the center point x (ip (N)) of the peak detection section 40 is the upper limit peak, the lower limit peak, or not the peak.

In this way, the entire section for peak detection is divided into N sections, the maximum value and the minimum value in one divided section are obtained, and it is determined whether the maximum value and the minimum value are peaks in the section. Until the peak is detected, the detection interval is changed and the process is repeated over N intervals, so that the peak value of the output signal can be correctly detected.
In addition, if the difference between the maximum value and the minimum value of the vibration detection section of the output signal is not larger than the threshold value, the peak is not set, so that the vibration peak can be accurately detected even for a signal with superimposed noise. .

DESCRIPTION OF SYMBOLS 1 Command signal 2 Controller 3 Control object 4 Output signal 5 Output feedback signal 6 Controller model part 7 Control object model part 8 Estimated output signal 9 Estimated output feedback signal 10 Difference signal 11 between output signal and estimated output signal, 51, 52 Vibration state determination unit 12 Feedback control device 13 Model unit 14 of feedback control device Low-pass filter 15 Output vibration component signal 20 Difference signal (Vpp) between upper limit peak and lower limit peak
30 Signal P
40 Peak detection section 41 Peak detection entire sections S101 to S107, S201 to S209, S301 to S310, S401 to S404, S501 to S507, S601 to S608 Processing 101 Speed control 102 Motor and load 103 Overall observer 111 Motor and load of the observer Inertia part 112 Observer speed control gain 113 Observer speed control integration 114 Integration 201 Speed controller 202 Current controller 203 Motor 204 Machine 205 Encoder 206 Differentiator 207 Oscillation detector 208 Memory 209 Subtractor

Claims (3)

A vibration detection device for a control device that controls a motor connected to a load,
The control device includes a controller that controls the motor by feedback-controlling the control target;
The vibration detection device includes:
Obtained by dividing the upper and lower limit peak values , which are values obtained by subtracting the immediately preceding lower limit peak value from the upper limit peak value at the detection time of the upper limit peak value of the output signal of the control device , by the time interval between the corresponding upper and lower limit peaks A vibration detection apparatus comprising: a vibration state determination unit that determines occurrence of vibration based on a value . The vibration state determination unit
The vibration intensity value obtained by dividing the upper and lower limit peak value by the time interval between the corresponding upper and lower limit peaks , and
Vibration moving average intensity value calculated by moving average the vibration intensity value;
The vibration duration that is a time width in which the vibration moving average intensity value exceeds a preset vibration intensity minimum value;
A comparison with preset threshold values corresponding to each of the vibration intensity value, the vibration moving average intensity value and the vibration duration;
The vibration detection device according to claim 1, wherein occurrence of vibration of the output signal of the control device is determined based on the control signal. A motor control device comprising the vibration detection device according to claim 1 or 2 and driving the motor.