WO2023032175A1 - Frequency characteristic prediction device and frequency characteristic prediction method - Google Patents

Abstract

The present invention reduces the number of rounds of measurement and shortens the measurement time, by predicting frequency characteristics on the basis of measured frequency characteristics.　This frequency characteristic prediction device comprises: a motor control unit that moves a shaft of a machine; a moving instruction generation unit that outputs a moving instruction for changing the position of the shaft from a first position to a second position; a frequency characteristic measurement unit that measures the frequency characteristics of the machine at the first and second positions: a state switching unit that switches the state of the motor control unit at the first position; and a frequency characteristic prediction unit that predicts the frequency characteristics of the machine at the second position. The frequency characteristic measurement unit measures a plurality of first frequency characteristics of a plurality of states to be switched at the first position, and measures a second frequency characteristic of at least one of the states at the second position. The frequency characteristic prediction unit predicts third frequency characteristics of a state other than at least one of the states at the second position by using the first frequency characteristics and second frequency characteristic.

Description

Frequency characteristic prediction device and frequency characteristic prediction method

The present invention relates to a frequency characteristic prediction device and frequency characteristic prediction method, and more particularly to a frequency characteristic prediction device and frequency characteristic prediction method for predicting the frequency characteristics of machine tools or industrial machines.

A motor control device that moves the axis of a machine tool or industrial machine to a desired position is required to adjust the appropriate gain, filter characteristics, etc. according to the attitude (axis position) of the machine tool or industrial machine.

Patent Literature 1 describes a servo control device that automatically and appropriately adjusts control gains so that control characteristics are maintained even when a load or machine configuration is changed.
Specifically, in Patent Document 1, a servo control device includes a speed command generation unit, a torque command generation unit, a speed detection unit, a speed control loop, a speed control gain, a sine wave disturbance input unit, It includes an actual frequency characteristic calculator, a reference characteristic changer, a reference frequency characteristic calculator, and a control gain adjuster. The sinusoidal disturbance input section changes frequency sequentially. The reference frequency characteristic calculator sequentially calculates the reference frequency characteristic corresponding to the feature designated by the reference characteristic changing unit for each frequency. The actual frequency characteristic calculator sequentially calculates the actual frequency characteristic of the control system for each frequency. The reference frequency characteristic calculation unit stores a characteristic expression of the reference characteristic changing unit when the reference frequency characteristic and the actual frequency characteristic are most similar.

Patent Literature 2 describes a servomotor control device that shortens the machining time of a machined object by a machine tool and does not deteriorate the machining accuracy immediately after the feed axis of the machine tool switches from rapid feed operation to cutting feed operation. there is
Specifically, in Patent Document 2, a calculation coefficient setting unit sets a calculation coefficient for creating at least one of feedforward control information and feedback control information to a first value set for cutting feed operation. It is described that it is set to a value between the calculation coefficient value and a second calculation coefficient value for fast-forward operation that is smaller than the first calculation coefficient value. Further, in Patent Document 2, the operation coefficient changing unit switches the operation command from the rapid traverse operation command to the cutting feed operation command at a second time point after the first time point, which is an arbitrary time point during the rapid traverse operation. Continuously changing the computational coefficient from a second value to a first computational coefficient value, if predicted at time 1, is described.

JP 2018-128734 A JP 2013-218552 A

In order to adjust the gain and filter characteristics of the motor control device, it is required to measure the frequency characteristics of the gain and phase of the motor control device at various posture measurement points.
In addition, if there are multiple states of the motor control device, such as cutting feed and rapid feed, by switching the cutting feed parameter and rapid feed parameter, if you try to measure the frequency characteristics in each state, each Measurements for the number of states times the number of measurement points are required, resulting in a long measurement time.
Therefore, there is a demand for a frequency characteristic prediction apparatus and a frequency characteristic prediction method that can predict frequency characteristics based on measured frequency characteristics, reduce the number of measurements, and shorten the measurement time.

(1) A first aspect of the present disclosure provides a motor control unit for moving an axis of a machine tool or industrial machine;
a movement command generation unit that outputs a movement command to the motor control unit for changing the position of the shaft from the first position to the second position;
a frequency characteristic measuring unit that measures the frequency characteristic of the machine tool or the industrial machine at the first position and the second position;
a state switching unit that switches the state of the motor control unit at the first position;
a frequency characteristic prediction unit that predicts the frequency characteristic of the machine tool or the industrial machine at the second position;
The frequency characteristic measuring unit measures, at the first position, a plurality of first frequency characteristics for a plurality of states switched by the state switching unit, and at the second position, among the plurality of states, measuring a second frequency characteristic for at least one of
The frequency characteristic prediction unit uses the plurality of first frequency characteristics and the second frequency characteristic to perform a first prediction regarding a state other than at least one of the plurality of states at the second position. 3 is a frequency characteristic prediction device for predicting the frequency characteristic of .

(2) A second aspect of the present disclosure moves a position of an axis of a machine tool or an industrial machine to a first position by a motor control unit based on a first movement command,
measuring a plurality of first frequency characteristics with respect to a plurality of states of the motor control unit at the first position;
moving the position of the axis from the first position to the second position by the motor control unit based on a second movement command;
At the second position, the frequency characteristic measuring unit measures a second frequency characteristic for at least one of the plurality of states;
a frequency characteristic prediction unit, using the plurality of first frequency characteristics and the second frequency characteristics, by the frequency characteristic prediction unit, at least one of the plurality of states at the second position other than is a frequency characteristic prediction method for predicting a third frequency characteristic for the state of

According to each aspect of the present disclosure, the frequency characteristics can be predicted based on the measured frequency characteristics, the number of measurements can be reduced, and the measurement time can be shortened.

1 is a block diagram showing a frequency characteristic prediction device according to a first embodiment of the present disclosure; FIG. It is a figure which shows the base used as the movable part which moves to the X-axis direction of a machine. FIG. 4 is a diagram showing the setting of parameters for rapid feed and cutting feed at a representative measurement point and two measurement points on a table that is a movable part of a machine. 4 is a flow chart showing the operation of the frequency characteristic prediction device according to the first embodiment of the present disclosure; FIG. 4 is a characteristic diagram showing gain characteristics in rapid feed and cutting feed measured at representative measurement points; FIG. 4 is a characteristic diagram showing phase characteristics in rapid feed and cutting feed measured at representative measurement points; FIG. 5 is a characteristic diagram showing the difference between the gain characteristics in rapid feed and the gain characteristics in cutting feed measured at representative measurement points. FIG. 5 is a characteristic diagram showing the difference between the phase characteristics in rapid feed and the phase characteristics in cutting feed measured at representative measurement points. FIG. 5 is a characteristic diagram showing fast-forward gain characteristics predicted at measurement points and fast-forward gain characteristics measured at measurement points; FIG. 5 is a characteristic diagram showing fast-forward phase characteristics predicted at measurement points and fast-forward phase characteristics measured at measurement points; FIG. 11 is a block diagram showing a frequency characteristic prediction device according to a second embodiment of the present disclosure; FIG. It is a block diagram which shows the parameter adjustment part which functions as a machine-learning apparatus. 1 is a block diagram showing a reference model; FIG. FIG. 4 is a characteristic diagram showing input/output gain characteristics of a motor control unit of a reference model and input/output gain characteristics of the motor control unit before learning and after learning; FIG. 4 is a block diagram showing an example in which a plurality of filters are directly connected to form a filter; FIG. 2 shows two platforms that are movable parts that move in the X-axis and Y-axis directions of the machine. FIG. 4 is a diagram showing the setting of parameters for rapid feed and cutting feed at one representative measurement point and eight measurement points; It is a figure which shows two bases used as the movable part which moves to X-axis direction and Y-axis direction of a machine, and the movable part which moves the main axis of a machine to Z-axis direction. FIG. 4 is a diagram showing the setting of parameters for rapid feed and cutting feed at two representative measurement points and nine measurement points; FIG. 3 shows a movable part that moves in the X-, Y- and Z-axis directions of the machine; FIG. 4 is a diagram showing the setting of parameters for rapid feed and cutting feed at one representative measurement point and 26 measurement points;

(First embodiment)
Hereinafter, embodiments of the present disclosure will be described in detail with reference to the drawings.
FIG. 1 is a block diagram showing a frequency characteristic prediction device according to the first embodiment of the present disclosure.
The frequency characteristic prediction device 10A includes a motor control section 100, a movement command generation section 200, a frequency characteristic measurement section 300, a storage section 400, a frequency characteristic prediction section 500, and a state switching section 600.
One or more of the movement command generation unit 200 , the frequency characteristic measurement unit 300 , the storage unit 400 , the frequency characteristic prediction unit 500 , and the state switching unit 600 may be provided inside the motor control unit 100 .

(Motor control unit 100)
The motor control section 100 includes a subtractor 110 , a speed control section 120 , a filter 130 , a current control section 140 and a motor 150 . The subtractor 110, the speed control section 120, the filter 130, the current control section 140, and the motor 150 constitute a closed speed feedback loop servo system. As the motor 150, a linear motor that performs linear motion, a motor that has a rotating shaft, or the like can be used. A controlled object 700 driven by the motor 150 is, for example, a moving part of a machine tool or an industrial machine. Motor 150 may be provided as part of a machine tool or industrial machine. The frequency characteristic prediction device 10A may be provided as part of a machine tool or industrial machine.

The subtractor 110 obtains the difference between the input speed command, which is the movement command, and the detected speed fed back, and outputs the difference to the speed control unit 120 as a speed deviation.

The speed control unit 120 performs PI control (Proportional-Integral Control), adds a value obtained by multiplying the speed deviation by an integral gain K1v and integrates it, and adds a value obtained by multiplying the speed deviation by a proportional gain K2v to obtain a torque command. Output to filter 130 . Note that the speed control unit 120 is not particularly limited to PI control, and may use other control such as PID control (Proportional-Integral-Differential Control).
Equation 1 (shown as Equation 1 below) represents the transfer function H _V (s) of the speed control section 120 .

A filter 130 is a filter that attenuates a specific frequency component, and for example, a notch filter, low-pass filter, or band-stop filter is used. A machine such as a machine tool having a mechanical section driven by the motor 150 has a resonance point, and the resonance may increase in the motor control section 100 . Resonance can be reduced by using a filter such as a notch filter. The output of filter 130 is output to current control section 140 as a torque command.
Equation 2 (shown as Equation 2 below) represents the transfer function H _F (s) of the notch filter as filter 130 .
Here, the coefficient δ in Equation 2 is the attenuation coefficient, the coefficient _ωc is the central angular frequency, and the coefficient τ is the fractional bandwidth. Assuming that the center frequency is fc and the bandwidth is fw, the coefficient ω _c is expressed by ω _c =2πfc and the coefficient τ is expressed by τ=fw/fc. Each coefficient becomes a coefficient of the filter.

Current control unit 140 generates a voltage command for driving motor 150 based on the torque command, and outputs the voltage command to motor 150 .
When the motor 150 is a linear motor, the position of the movable portion is detected by a linear scale (not shown) provided on the motor 150, the detected speed value is obtained by differentiating the detected position value, and the detected speed is obtained. The value is input to subtractor 110 as velocity feedback.
If the motor 150 has a rotating shaft, the rotation angle position is detected by a rotary encoder (not shown) provided on the motor 150, and the speed detection value is input to the subtractor 110 as speed feedback.
In the following description, it is assumed that the motor 150 has a rotating shaft and the speed detection value is detected by a rotary encoder (not shown).

The motor control unit 100 is configured as described above.
To adjust one or both of the integral gain K1v and the proportional gain K2v of the speed control unit 120 and/or each coefficient ω _c , τ, δ of the transfer function of the filter 130 of the motor control unit 100, the machine tool Or it is required to measure the frequency characteristics of industrial machines. The frequency characteristic of the machine tool or industrial machine can be found by measuring the frequency characteristic of the motor control section 100 . Hereinafter, one or both of the integral gain K1v and the proportional gain K2v of the speed control unit 120 and/or the coefficients ω _c , τ, δ of the transfer function of the filter 130 are referred to as parameters.

When adjusting parameters of the motor control unit 100, including changes in mechanical properties at different positions of the moving part of the machine, it is required to measure the frequency characteristics of the motor control unit 100 at each position. In the following description, when simply referred to as a frequency characteristic, the frequency characteristic means the frequency characteristic of the motor control unit 100. FIG.
Further, when setting the parameters of the motor control unit 100 according to the state of the motor control unit 100 such as rapid feed and cutting feed, the parameters of the motor control unit 100 are set for each state. It is required to measure the frequency characteristics for each set parameter.
Setting multiple parameters according to multiple states at all positions of the moving part of a machine tool or industrial machine (hereafter referred to as a machine) and measuring the frequency characteristics of each parameter increases the number of measurements and reduces the frequency characteristics. Longer measurement time.

In this embodiment, in order to shorten the frequency characteristic measurement time, the frequency characteristic prediction device 10A measures a plurality of frequency characteristics with a plurality of parameter settings only at representative measurement points, and at measurement points other than the representative measurement points, Frequency characteristics at least one of the plurality of parameter settings are measured, and frequency characteristics at parameter settings other than at least one of the plurality of parameter settings are predicted from the measured frequency characteristics. For this purpose, the frequency characteristic prediction device 10A includes a movement command generation section 200, a frequency characteristic measurement section 300, a storage section 400, a frequency characteristic prediction section 500, and a state switching section 600. The representative measurement point becomes the first position, and the measurement points other than the representative measurement point become the second position.
In this embodiment, the parameters of the motor control unit 100 are preset by the user.

In the following description, the motor control unit 100 moves the table 810, which is the movable part of the machine, in the X-axis direction by the motor 150 as shown in FIG. A case in which the state switching unit 600 (to be described later) switches the parameters of fast feed and cutting feed in A1 and A3 to measure the frequency characteristics will be described as an example. Although the position of the representative measurement point A2 is not particularly limited, it is preferably set in the center of the movable range of the table.
FIG. 2 is a diagram showing a table that is a movable part that moves in the X-axis direction of the machine. Fig. 3 is a diagram showing the settings of rapid feed and cutting feed parameters at one representative measurement point (representative measurement point A2) and two measurement points (measurement points A1 and A3) on the table, which is the movable part of the machine. is. In FIG. 3, a representative measurement point A2 for measuring frequency characteristics with two parameter settings of rapid feed and cutting feed is indicated by a black circle, and a measurement point A1 for measuring frequency characteristics with parameter settings for cutting feed. A3 is circled. The double-headed arrows at representative measurement point A2 and measurement points A1 and A3 shown in FIG. Points A1 and A3 indicate that the X-axis frequency characteristics are measured.

(Movement command generating unit 200 and state switching unit 600)
The movement command generation unit 200 outputs a fast-forward speed command to the subtractor 110 and the state switching unit 600 to move the table 810 to the position of the representative measurement point A2. The motor control unit 100 is controlled based on the fast-forward speed command, and the platform 810 reaches the representative measurement point A2. The state switching unit 600 sets the parameters of the motor control unit 100 for fast forward based on the speed command for fast forward. After that, the movement command generation unit 200 outputs the sine wave signal as the speed command Vcmd to the subtractor 110 of the motor control unit 100 and the frequency characteristic measurement unit 300 while changing the frequency. After that, at the position of the representative measurement point A2, when the movement command generation unit 200 changes the speed command for rapid feed to the speed command for cutting feed, the state switching unit 600 changes the parameters of the motor control unit 100 based on the speed command for cutting feed. for cutting feed. The movement command generation unit 200 outputs a sine wave signal as a speed command Vcmd to the subtractor 110 and the frequency characteristic measurement unit 300 while changing the frequency.

At the position of the representative measurement point A2, the motor control unit 100 first operates with a sine wave signal in which the frequency changes with the parameters for rapid feed as the speed command Vcmd. A sine wave signal is used as the speed command Vcmd. Velocity detection values Vfd1 and Vfd2 obtained by operating the motor control unit 100 with the fast feed parameter and the cutting parameter are input to the subtractor 110 and the frequency characteristic measurement unit 300 as velocity feedback.

Next, the movement command generation unit 200 outputs a fast-forward speed command to the subtractor 110, the motor control unit 100 is controlled based on the fast-forward speed command, and the table 810 moves from the representative measurement point A2 to the measurement point A1 or A3. (change the position of the X-axis from the representative measurement point A2 to the measurement point A1 or A3). The state switching unit 600 changes the parameter of the motor control unit 100 from cutting to rapid traverse based on the rapid traverse speed command.
When the movement command generation unit 200 changes the speed command for rapid feed to the speed command for cutting feed when the table 810 reaches the measurement point A1 or A3, the state switching unit 600 switches the motor according to the speed command for cutting feed. The parameters of the control unit 100 are changed for cutting feed.
The motor control unit 100 operates at the position of the measurement point A1 or A3 using a sine wave signal whose frequency changes as a speed command Vcmd, and the speed detection value obtained by operating the motor control unit 100 with parameters for cutting. Vfd3 is input to subtractor 110 and frequency characteristic measuring section 300 as velocity feedback.
It should be noted that the setting of different parameters for rapid traverse and cutting feed is one or both of the integral gain K1v and the proportional gain K2v of the velocity control unit 120, and/or the coefficients ω _c and τ of the transfer function of the filter 130. , δ.

(Frequency characteristic measurement unit 300 and storage unit 400)
In the following description, the frequency characteristic measurement unit 300 measures the frequency characteristics at the representative measurement point A2 when setting parameters for rapid feed and when setting parameters for cutting feed, which are switched by the state switching unit 600, and measures the frequency characteristics at the measurement point A1. Next, the operation of measuring the frequency characteristics when setting parameters for cutting feed and predicting the frequency characteristics when setting parameters for rapid feed at measurement point A1 will be described.

The frequency characteristic measuring unit 300 uses the speed command Vcmd of the sine wave signal and the speed detection value Vfd1 when setting the fast-forwarding parameters at the representative measurement point A2 to obtain the speed command Vcmd and the speed command Vcmd as input signals at each frequency. By obtaining the amplitude ratio (input/output gain) and the phase delay of the speed detection value Vfd1 as an output signal, the frequency characteristic f1 is measured and stored in the storage unit 400 .
In addition, the frequency characteristic measuring unit 300 uses the speed command Vcmd and the speed detection value Vfd2 when setting parameters for cutting feed at the representative measurement point A2, and uses the speed command Vcmd as an input signal and the output speed command Vcmd at each frequency. By obtaining the amplitude ratio (input/output gain) and phase delay of the speed detection value Vfd2 as a signal, the frequency characteristic f2 is measured and stored in the storage unit 400 .
In this way, the storage unit 400 stores the frequency characteristic f1 at the representative measuring point A2 when setting the parameters for rapid feed and the frequency characteristic f2 at the representative measuring point A2 when setting the parameters for cutting feed. remembered. The frequency characteristic f1 consists of a gain characteristic L _2F and a phase characteristic ∠G _2F . The frequency characteristic f2 consists of a gain characteristic L _2C and a phase characteristic ∠G _2C . The frequency characteristic f1 and the frequency characteristic f2 are a plurality of first frequency characteristics.

Furthermore, the frequency characteristic measuring unit 300 uses the speed command Vcmd as an input signal and the speed detection value as an output signal using the speed command Vcmd and the speed detection value Vfd3 when setting parameters for cutting feed at the measurement point A1. By obtaining the amplitude ratio (input/output gain) and phase delay of the value Vfd3, the frequency characteristic f3 is measured and output to the frequency characteristic prediction section 500. FIG. The frequency characteristic f3 consists of a gain characteristic L _1C and a phase characteristic ∠G _1C . The frequency characteristic f3 becomes the second frequency characteristic.

The frequency characteristic measuring section 300 may change at least one of the amplitude of the speed command Vcmd serving as an input signal, the number of vibrations, and the vibration method.
The movement command generation unit 200 generates a signal such as a sine wave signal while changing the frequency, and vibrates the controlled object while changing the frequency.
The frequency characteristic measurement unit 300 can change at least one of the amplitude of the input signal (excitation input), the number of times of excitation, and the excitation method.
Here, the excitation input is the amplitude of the input signal, the number of excitations is the number of cycles for exciting the machine at the same frequency, and the excitation method is "normal mode" and "high precision mode".

The high-precision mode is a mode in which the high-frequency band excitation time is constant and the input phase is shifted while sweeping multiple times. Using this mode, the measurement accuracy in the high-frequency region of 1 kHz or higher can be improved.
If the amplitude of the input signal is small, the mechanical response cannot be taken correctly due to the influence of the frictional force, so the mechanical characteristics cannot be correctly indicated.
If the number of excitations is insufficient and the response of the machine is not a steady response but a transient response, the machine characteristics cannot be shown correctly.
If the input time is short in the high frequency range, the response of the machine cannot be taken correctly, so the machine characteristics cannot be shown correctly.
The frequency characteristic measurement unit 300 changes at least one of the amplitude of the velocity command Vcmd, which is an input signal, the number of vibrations, and the vibration method, taking the above situation into account.

(Frequency characteristic prediction unit)
The frequency characteristic prediction unit 500 obtains from the storage unit 400 the fast feed frequency characteristic f1 (gain characteristic _L2F and phase characteristic _∠G2F ) and the cutting feed frequency characteristic f2 (gain characteristic _L2C and phase Read out the characteristic ∠G _2C ).
Further, the frequency characteristic prediction unit 500 acquires the cutting feed frequency characteristic f3 (gain characteristic L _1C and phase characteristic ∠G _1C ) at the measurement point A1 from the frequency characteristic measurement unit 300 . The frequency characteristic f3 measured by the frequency characteristic measuring section 300 may be stored in the storage section 400 and the frequency characteristic prediction section 500 may read the frequency characteristic f3 from the storage section 400 .
Then, the frequency characteristic prediction unit 500 uses the rapid feed frequency characteristic f1 and the cutting feed frequency characteristic f2 at the representative measurement point A2, and the cutting feed frequency characteristic f3 at the measurement point A1, to determine the measurement point A1 , the predicted fast-forward frequency characteristic f4 (gain characteristic L _1F and phase characteristic ∠G _1F ) is calculated. The predicted frequency characteristic f4 becomes the third frequency characteristic.

Specifically, the fast-forward gain characteristic L _1F and the phase characteristic ∠G _1F , which are the fast-forward predicted frequency characteristic f4, are calculated using Equation 3 (shown as Equation 3 below).

Note that the calculation of the gain characteristic L _1F performed by the frequency characteristic prediction unit 500 is L _1F = (L _2F - L _2C ) + L _1C , L _1F = (L _1C - _{L 2C} ) + L _2F , and L _1F = (L _2F +L _1C )−L _2C .
Further, the calculation of the phase characteristic ∠G _1F performed by the frequency characteristic prediction unit 500 is ∠G _1F = (∠G _2F - ∠G _2C ) + ∠G _1C , ∠G _1F = (∠G _1C - ∠G _2C ) + ∠G _2F and ∠G _1F =(∠G _2F +∠G _1C )-∠G _2C .

The reason why the frequency characteristics of the measurement points A1 other than the representative measurement point A2 can be predicted using the above-mentioned formula 3 is as follows.
Regarding the frequency characteristics of the velocity loop (velocity feedback loop), the presence or absence of the influence of the position dependence of the moving part of the machine, and the presence or absence of the influence of the difference in the state of rapid feed, cutting feed, etc. are determined by the speed control unit 120 and the filter 130 , the current controller 140 and the controlled plant consisting of the motor 150 and the controlled object 700 are shown in Table 1.

As shown in Table 1, only the control plant has an effect on the frequency characteristics of the speed loop, and the speed control unit 120, the filter 130, and the current control unit 140 have no effect on the frequency characteristics of the speed loop. do not have. Therefore, the difference between the frequency characteristics of the two states at the same position, for example, the difference between the frequency characteristics in rapid traverse and the frequency characteristics in cutting feed, does not depend on the position.
Therefore, by adding the difference between the frequency characteristics of rapid traverse and the frequency characteristics of cutting feed at representative measuring points and the frequency characteristics of cutting feed at measuring points other than representative measuring points, It is possible to predict the fast-forward frequency characteristics at the measurement point.
The method of predicting the fast-forward frequency characteristics at measurement points other than the representative measurement points is not particularly limited to the method using Equation 3, and other methods may be used.

The operation of estimating the frequency characteristics when setting the fast-forward parameters at the measurement point A1 has been described above. can.

The functional blocks included in the frequency characteristic prediction device 10A have been described above.
In order to implement these functional blocks, the frequency characteristic prediction device 10A includes an arithmetic processing device such as a CPU (Central Processing Unit). The frequency characteristic prediction device 10A also includes an auxiliary storage device such as a HDD (Hard Disk Drive) that stores various control programs such as application software and an OS (Operating System), and an arithmetic processing device that executes the program. It also has a main memory such as a random access memory (RAM) for storing temporarily needed data.

Then, in the frequency characteristic prediction device 10A, the arithmetic processing unit reads the application software or the OS from the auxiliary storage device, and develops the read application software or OS in the main storage device, while performing arithmetic processing based on the application software or the OS. do Also, based on the result of this calculation, various hardware included in each device is controlled. This implements the functional blocks of the present embodiment. In other words, this embodiment can be realized by cooperation of hardware and software.

Next, the operation of the frequency characteristic prediction device 10A will be explained using a flowchart. FIG. 4 is a flow chart showing the operation of the frequency characteristic prediction device.

In step S11, the motor control unit 100 is controlled based on the fast-forward movement command from the movement command generation unit 200, and the table 810, which is the movable unit, moves to the representative measurement point A2.

In step S12, the state switching unit 600 sets the parameters of the motor control unit 100 as fast-forward parameters based on the fast-forward movement command. Step S12 may be performed simultaneously with step S11, or may be performed before step S11.

In step S13, the movement command generation unit 200 outputs the speed command Vcmd, which is a sine wave signal whose frequency changes, to the motor control unit 100. By obtaining the amplitude ratio (input/output gain) and the phase delay between the command Vcmd and the speed detection value Vfd1 as the output signal, the frequency characteristic f1 is measured and stored in the storage unit 400 .
do.

In step S14, when the movement command generation unit 200 changes the speed command for rapid feed to the speed command for cutting feed, the state switching unit 600 changes the parameters of the motor control unit 100 for cutting feed based on the speed command for cutting feed. do.

In step S15, the movement command generation unit 200 outputs a speed command Vcmd, which is a sinusoidal signal whose frequency changes, to the motor control unit 100. By obtaining the amplitude ratio (input/output gain) and the phase delay between the command Vcmd and the speed detection value Vfd2 as the output signal, the frequency characteristic f2 is measured and stored in the storage unit 400 .
In this manner, the storage unit 400 stores the frequency characteristic f1 when setting the parameters for rapid feed and the frequency characteristic f2 when setting the parameters for cutting feed at the representative measurement point A2.

In step S16, the motor control unit 100 is controlled based on the fast-forward movement command from the movement command generation unit 200, and the table 810, which is the movable unit, moves to the measurement point A1. The state switching unit 600 changes the parameters of the motor control unit 100 for fast forward based on the speed command for fast forward.

In step S17, when the movement command generation unit 200 changes the speed command for rapid feed to the speed command for cutting feed, the state switching unit 600 changes the parameters of the motor control unit 100 for cutting feed based on the speed command for cutting feed. do. When the table 810, which is the movable part, moves from the representative measurement point A2 to the measurement point A1, if it moves by cutting feed, in step S16, the state switching unit 600 changes the parameter of the motor control unit 100 to fast feed. and step S17 becomes unnecessary.

In step S18, the movement command generation unit 200 outputs the speed command Vcmd, which is a sine wave signal whose frequency changes, to the motor control unit 100 as a movement command, and the frequency characteristic measurement unit 300 generates an input signal Vcmd for each frequency. By obtaining the amplitude ratio (input/output gain) and the phase delay between the speed command Vcmd as the output signal and the speed detection value Vfd2 as the output signal, the frequency characteristic f3 is measured and output to the frequency characteristic prediction unit 500 .

In step S19, the frequency characteristic prediction unit 500 determines the rapid feed frequency characteristic f1 at the representative measurement point A2, the cutting feed frequency characteristic f2 at the representative measurement point A2, and the cutting feed frequency at the measurement point A1. Using the characteristic f3, a predicted fast-forward frequency characteristic f4 at the measurement point A1 is calculated. The fast-forward predicted frequency characteristic f4 can be calculated using Equation 3 as already described.

In step S20, it is determined whether or not the prediction of the frequency characteristics has been completed at all the measurement points. The processing from step S16 to step S20 is performed until the prediction is completed. When the prediction of the frequency characteristics is completed at all measurement points, the process is terminated.

The operation of estimating the frequency characteristics when setting the fast-forward parameters at the measurement point A1 has been described above. can.

In this embodiment, the frequency characteristics are measured only at the representative measurement point A2 when setting the parameters for fast-forwarding, and the frequency characteristics when setting the parameters for fast-forwarding are not measured at the measurement points A1 and A3. Measurement time can be shortened.

An example in which frequency characteristics are predicted using the frequency characteristics prediction device of the present embodiment in a cutting machine will be described below. The cutting machine is a cutting machine shown in FIG. 18 which will be described later.
Move the table, which is the movable part of the cutting machine, in the X-axis direction, set the representative measurement point at the center of the movable range in the X-axis direction, and the frequency characteristic prediction device predicts rapid traverse and cutting feed at the representative measurement point. Then, the difference between the frequency characteristics in rapid traverse and the frequency characteristics in cutting feed was obtained.

FIG. 5 is a characteristic diagram showing gain characteristics in rapid traverse and cutting feed measured at representative measurement points, and FIG. 6 is a characteristic diagram showing phase characteristics in rapid traverse and cutting feed measured at representative measurement points. . FIG. 7 is a characteristic diagram showing the difference between the gain characteristics in rapid traverse and the gain characteristics in cutting feed measured at the representative measurement points, and FIG. , and is a characteristic diagram showing the difference from the phase characteristics in cutting feed.
Next, the frequency characteristics prediction device measures the frequency characteristics in cutting feed at a measurement point away from the representative measurement point, and calculates the difference between the frequency characteristics in rapid traverse and the frequency characteristics in cutting feed. In addition, we predicted the fast-forward frequency characteristics at a measurement point distant from the representative measurement point. For comparison, we actually measured the fast-forward frequency characteristics at a measurement point away from the representative measurement point.
FIG. 9 is a characteristic diagram showing the fast-forward gain characteristics predicted at the measurement points and the fast-forward gain characteristics measured at the measurement points. FIG. FIG. 10 is a characteristic diagram showing phase characteristics of fast-forwarding.
As shown in FIGS. 9 and 10, the fast-forward frequency characteristics predicted at the measurement points almost overlap with the fast-forward frequency characteristics actually measured at the measurement points. The prediction was found to be valid.

(Second embodiment)
In the first embodiment, an example of estimating the frequency characteristics during fast forward at measurement points other than the representative measurement points has been described. In the present embodiment, a frequency characteristic prediction device having a parameter adjustment unit that adjusts the parameters of the motor control unit 100 based on the frequency characteristics measured by the frequency characteristic measurement unit 300 or the frequency characteristics predicted by the frequency characteristic prediction unit 500. explain.

FIG. 11 is a block diagram showing a frequency characteristic prediction device according to the second embodiment of the present disclosure. In FIG. 11, the same constituent members as those shown in FIG.
As shown in FIG. 11, frequency characteristic prediction device 10B has a configuration in which parameter adjustment section 800 is added to frequency characteristic prediction device 10A shown in FIG.
In the following description, as in the description of the first embodiment, the motor control unit 100 moves the table 810, which is the movable unit, in the X-axis direction by the motor 150 as shown in FIG. A case where the state switching unit 600, which will be described later, switches the parameters of rapid feed and cutting feed to measure the frequency characteristics at the representative measurement point A2 and the measurement points A1 and A3 shown in FIG.

The parameter adjustment unit 800 adjusts the parameters of the motor control unit 100 during rapid feed and cutting feed based on the frequency characteristics of rapid feed and cutting feed at the representative measurement point A2 measured by the frequency characteristic measurement unit 300.
Further, the parameter adjustment unit 800 adjusts the parameters of the motor control unit 100 during cutting feed based on the frequency characteristic of the cutting feed at the measurement point A1 measured by the frequency characteristic measurement unit 300 .
Furthermore, the parameter adjustment unit 800 adjusts the parameters of the motor control unit 100 during fast-forward based on the fast-forward frequency characteristic at the measurement point A1 predicted by the frequency characteristic prediction unit 500 .

The method of adjusting the parameters of the motor control unit 100 based on the frequency characteristics measured by the frequency characteristics measurement unit 300 and the frequency characteristics predicted by the frequency characteristics prediction unit 500 is not particularly limited. An example using reinforcement learning will be described. Reinforcement learning used in this embodiment is described, for example, in Japanese Patent Application Laid-Open No. 2020-177257. Although machine learning can use reinforcement learning, it is not particularly limited to reinforcement learning, and for example, supervised learning may be used.

Hereinafter, parameter adjustment section 800 that functions as a machine learning device will be referred to as parameter adjustment section 800A.
The parameter adjustment unit 800A acquires the frequency characteristics measured by the frequency characteristics measurement unit 300 or the frequency characteristics predicted by the frequency characteristics prediction unit 500, and the acquired frequency characteristics are the same as the target frequency characteristics or within a certain range. Then, the optimum values of the parameters of the motor control unit 100 are machine-learned (hereinafter, "machine learning" is referred to as "learning"). Then, the parameter adjustment unit 800A sets the parameters of the motor control unit 100, that is, the integral gain K1v, the proportional gain K2v, and the coefficients ω _c , τ, and δ of the transfer function of the filter 130 to optimum values.
Although the parameter adjustment by the parameter adjustment section 800A is performed before shipment, re-learning may be performed after shipment.

<Parameter adjustment unit 800A>
The parameter adjustment unit 800A sets the frequency characteristic (input/output gain and phase delay) measured by the frequency characteristic measurement unit 300 or the frequency characteristic (input/output gain and phase delay) predicted by the frequency characteristic prediction unit 500 as a state S, and sets the state Q-learning is performed in which adjustment of the parameter value related to S is action A. As is well known to those skilled in the art, Q-learning aims to select the action A with the highest value Q(S, A) from among possible actions A in a certain state S as the optimum action. do.

Parameter adjustment section 800A observes state information S including frequency characteristics measured by frequency characteristic measurement section 300 or frequency characteristics predicted by frequency characteristic prediction section 500, and determines action A. FIG. The parameter adjustment unit 800A receives a reward each time action A is performed. Rewards will be discussed later.
In Q-learning, the parameter adjustment unit 800A, for example, searches for the optimal action A that maximizes the total future reward by trial and error. By doing so, the parameter adjusting section 800A can select the optimum action A (that is, the optimum servo parameter value) for the state S.

In parameter adjusting section 800A, based on the frequency characteristics measured by frequency characteristic measuring section 300, the parameters of motor control section 100 are adjusted. is different from adjusting the parameters of Each operation will be described below.

(When adjusting the parameters of the motor control part based on the measured frequency characteristics)
FIG. 12 is a block diagram showing the configuration of parameter adjusting section 800A.
In order to perform the above-described reinforcement learning, as shown in FIG. A section 805 is provided.

The state information acquisition unit 801 operates the motor control unit 100 using the adjusted parameters, and acquires the frequency characteristics (the gain characteristics of the input/output gains and the phase characteristics indicating the phase delay) measured by the frequency characteristics measurement unit 300. and output to the learning unit 802 . State information acquisition section 801 acquires the frequency characteristic of motor control section 100 with parameters before adjustment from frequency characteristic measurement section 300 and outputs it to learning section 802 when Q learning is first started. The frequency characteristic acquired from the frequency characteristic measurement unit 300 becomes the state information S. FIG. Note that the parameters are shown as coefficients of the filter 130 in FIG.

It should be noted that, at the time when Q-learning is first started, the initial value parameters are generated in advance by the user. If the machine tool is adjusted in advance by the operator, the initial values of the parameters may be adjusted values.

A learning unit 802 is a part that learns the value Q(S, A) when a certain action A is selected under a certain state S. The learning unit 802 includes a reward output unit 8021, a value function update unit 8022, and an action information generation unit 8023.

The reward output unit 8021 is a part that calculates a reward when action A is selected under a certain state S. FIG.
The reward output unit 8021 compares the input/output gain gs for each frequency when the initial value parameter is adjusted with the input/output gain value gb for each frequency of the preset reference model. The reward output unit 8021 gives a negative reward when the input/output gain gs is greater than the input/output gain value gb of the reference model. On the other hand, if the input/output gain gs is equal to or less than the input/output gain value gb of the reference model, the reward output unit 8021 outputs a positive A reward is given, giving a negative reward when the phase lag increases and a zero reward when the phase lag does not change.

First, the operation of the reward output unit 8021 to give a negative reward when the input/output gain gs is greater than the input/output gain value gb of the reference model will be described with reference to FIGS. 13 and 14. FIG.
The reward output unit 8021 stores a reference model of input/output gains. The reference model is a model of a motor controller that has ideal characteristics without resonance. The reference model can be calculated from the inertia Ja, torque constant _Kt , proportional gain _Kp , integral gain KI _, and differential gain _KD of the model shown in FIG. 13, for example. Inertia Ja is the sum of motor inertia and mechanical inertia.
FIG. 14 is a characteristic diagram showing the frequency characteristics of the input/output gain of the motor control unit of the reference model and the frequency characteristics of the input/output gain of the motor control unit 100 before learning and after learning. As shown in the characteristic diagram of FIG. 14, the reference model has an area A, which is a frequency area in which an ideal input/output gain is obtained above a certain input/output gain, for example, -20 dB or above, and a frequency area below the certain input/output gain. and a region B which is a frequency region where In region A of FIG. 14, the ideal input/output gain of the reference model is indicated by curve MC ₁ (thick line). In region B of FIG. 14, the ideal virtual input/output gain of the reference model is indicated by curve MC ₁₁ (thick dashed line), and the input/output gain of the reference model is indicated by straight line MC ₁₂ (thick line) as a constant value. In areas A and B of FIG. 14, curves RC ₁ and RC ₂ indicate the curves of input/output gains with respect to the motor control unit before and after learning, respectively.

In region A, the reward output unit 8021 gives a first negative reward when the input/output gain curve _RC1 before learning exceeds the ideal input/output gain curve _MC1 of the reference model.
In the region B exceeding the frequency where the input/output gain becomes sufficiently small, even if the input/output gain curve _RC1 before learning exceeds the ideal virtual input/output gain curve _MC11 of the reference model, the stability is not affected. become smaller. Therefore, in region B, as described above, the input/output gain of the reference model uses a straight line _MC12 of a constant input/output gain (eg, -20 dB) instead of the ideal gain characteristic curve _MC11 . However, if the pre-learning input/output gain curve _RC1 exceeds the constant input/output gain straight line _MC12 , there is a possibility of instability, so a first negative value is given as a reward.

Next, when the input/output gain gs is equal to or less than the input/output gain value gb of the reference model, the operation of the reward output unit 8021 to determine the reward based on the phase lag will be described.
In the following description, D(S) is the phase delay that is the state variable related to the state information S, is denoted by D(S'). At the time when the Q learning is first started, the phase delay is not obtained. The following reward is determined with the phase delay of the control unit 100 as D(S).

The method by which the reward output unit 8021 determines the reward based on the phase lag includes, for example, the following method.
When the state S changes to state S', the reward can be determined depending on whether the frequency at which the phase delay is 180 degrees increases, decreases, or remains the same. Here, the case where the phase delay is 180 degrees is taken up, but it is not particularly limited to 180 degrees, and other values may be used.
For example, when the state S changes to the state S', the phase delay increases if the curve changes so that the frequency at which the phase delay is 180 degrees becomes smaller. On the other hand, if the curve changes so that the frequency at which the phase delay is 180 degrees increases when the state S changes to the state S', the phase delay decreases.

Therefore, when the state S changes to state S′, when the frequency at which the phase delay is 180 degrees becomes small, the phase delay D(S)<phase delay D(S′) is defined, and the reward output unit 8021 sets the reward value to the second negative value. Note that the absolute value of the second negative value is made smaller than the first negative value.
On the other hand, when the state S changes to state S′, when the frequency at which the phase delay is 180 degrees increases, the phase delay D(S) is defined as >phase delay D(S′), and the reward output A unit 8021 sets the reward value to a positive value.
Further, when the state S changes to the state S′, when the frequency at which the phase delay is 180 degrees does not change, the phase delay D(S) is defined as the phase delay D(S′), and the reward output unit 8021 sets the reward value to a value of zero.

The method of determining the reward based on the phase lag is not limited to the above method. may use a method that rewards positive values and rewards zero when they are the same.

The reward output unit 8021 has been described above.

The value function updating unit 8022 performs Q-learning based on the state S, the action A, the state S′ when the action A is applied to the state S, and the reward obtained as described above. The value function Q stored in the value function storage unit 804 is updated.
The value function Q may be updated by online learning, batch learning, or mini-batch learning.
Online learning is a learning method in which, by applying a certain action A to the current state S, the value function Q is updated immediately each time the state S transitions to a new state S'. In batch learning, learning data is collected by applying a certain action A to the current state S, and by repeating the transition of the state S to a new state S′. This is a learning method in which the value function Q is updated using learning data. Furthermore, mini-batch learning is a learning method intermediate between online learning and batch learning, in which the value function Q is updated every time learning data is accumulated to some extent.

The action information generation unit 8023 selects action A for the current state S in the process of Q learning. The action information generation unit 8023 generates action information A in order to perform an operation (corresponding to action A in Q-learning) to adjust the value of the servo parameter in the process of Q-learning. Output to the action information output unit 803 .
More specifically, the action information generation unit 8023 calculates, for example, the integral gain K1v and the proportional gain K2v of the speed control unit 120 in the parameters included in the action A with respect to the adjusted parameters included in the state S, and each coefficient ω _c , τ, δ of the transfer function of filter 130 may be incrementally added or subtracted.

All of the integral gain K1v and proportional gain K2v of the speed control unit 120 and the coefficients ω _c , τ, and δ of the filter 130, which are parameters, may be modified. good. When adjusting the coefficients ω _c , τ, δ of the filter 130, for example, it is easy to find the center frequency fc that causes resonance and to specify the center frequency fc. Therefore, the behavior information generation unit 8023 temporarily fixes the center frequency fc, corrects the bandwidth fw and the attenuation coefficient δ, that is, fixes the coefficient ω _c (=2πfc), and modifies the coefficient τ (=fw/fc). and the action information A may be generated and output to the action information output unit 803 in order to perform the operation of correcting the attenuation coefficient δ.

Further, the action information generation unit 8023 may use a greedy method for selecting the action A' with the highest value Q(S, A) among the values of the action A currently estimated, or a random action with a certain small probability ε. A' is selected, and otherwise, a known method such as the ε-greedy method of selecting the action A' with the highest value Q(S, A) may be used to select the action A'.

The action information output unit 803 is a part that transmits the action information A output from the learning unit 802 to the motor control unit 100 . As described above, based on this behavior information, in the current state S, that is, the current setting in the motor control unit 100, the integral gain K1v and the proportional gain K2v of the speed control unit 120 and/or each Adjusting the coefficients ω _c , τ, δ transitions to the next state S′ (that is, the adjusted integral gain K1v and proportional gain K2v of the speed control unit 120 and/or each coefficient of the filter 130).

The value function storage unit 804 is a storage device that stores the value function Q. The value function Q may be stored as a table (hereinafter referred to as an action value table) for each state S and action A, for example. Value function Q stored in value function storage unit 804 is updated by value function update unit 8022 . Also, the value function Q stored in the value function storage unit 804 may be shared with the parameter adjustment unit 800A of another machine tool. If the value function Q is shared by the parameter adjustment units 800A of a plurality of machine tools, it becomes possible to perform reinforcement learning in a distributed manner in the parameter adjustment units 800A of the respective machine tools, thereby increasing the efficiency of reinforcement learning. can be improved.

Based on the value function Q updated by the value function updating unit 8022 performing Q-learning, the optimized action information output unit 805 selects the action that maximizes the value Q(S, A) from the speed control unit 120 and the filter 130 . Behavior information A (hereinafter referred to as “optimization behavior information”) is generated.
More specifically, the optimized behavior information output unit 805 acquires the value function Q stored in the value function storage unit 804. FIG. This value function Q is updated by the value function updating unit 8022 performing Q learning as described above. Then, the optimized behavior information output unit 805 generates behavior information based on the value function Q, and outputs the generated behavior information to the speed control unit 120 and/or the filter 130 of the motor control unit 100 . This optimization behavior information includes information for correcting the integral gain K1v and proportional gain K2v of the speed control unit 120 of the motor control unit 100 and/or each coefficient ω _c , τ, δ of the transfer function of the filter 130. .

The speed control unit 120 corrects the integral gain K1v and the proportional gain K2v based on this action information, and the filter 130 corrects each coefficient ω _c , τ, δ of the transfer function based on this action information.

The parameter adjustment unit 800A can optimize the parameters and simplify the adjustment of the parameters through the operations described above.

(When adjusting the parameters of the motor control part based on the predicted frequency characteristics)
The configuration of parameter adjustment section 800A is the same as the configuration shown in FIG.

By changing various parameters of the motor control unit 100, the frequency characteristic measuring unit 300 measures the frequency characteristics during rapid feed at the representative measurement point A2, and measures the frequency characteristics during cutting feed at the representative measurement points A2 and A1. Measure. Using these frequency characteristics, the frequency characteristic prediction section 500 predicts the predicted fast-forward frequency characteristics at the measurement point A1. The parameters of the motor control section 100 when the predicted frequency characteristics during fast-forward are obtained are the same as the parameters of the motor control section 100 when setting the fast-forward at the representative measurement point A2.

The frequency characteristic prediction unit 500 associates the parameter of the motor control unit 100 at the time of fast-forward setting at the representative measurement point A2 with the predicted frequency characteristic at the time of fast-forward at the measurement point A1, and performs various parameters for each parameter. Stored in the storage unit in the frequency characteristic prediction unit 500 .

When performing the above-described reinforcement learning, the state information acquisition unit 801 of the parameter adjustment unit 800A, from the storage unit in the frequency characteristic prediction unit 500, selects a certain parameter at the representative measurement point and Get the frequency characteristics (gain and phase).
The behavior information output unit 803 designates another parameter stored in the storage unit within the frequency characteristic prediction unit 500 .

The state information acquisition unit 801 acquires another specified parameter and the frequency characteristics (gain and phase) of the fast-forward mode predicted by the other parameter from the storage unit in the frequency characteristics prediction unit 500 .
The learning operation of parameter adjustment section 800A other than the above operation is the same as the operation of parameter adjustment section 800A already described.
By performing machine learning in this manner, optimum parameters corresponding to optimum frequency characteristics can be obtained.

When the amount of calculation is large for the parameter adjustment unit 800A that performs machine learning, for example, a personal computer is equipped with a GPU (Graphics Processing Units), and a technique called GPGPU (General-Purpose computing on Graphics Processing Units) is used to perform arithmetic processing on the GPU. If you use it for , it will be possible to perform high-speed processing. Furthermore, in order to perform faster processing, multiple computers equipped with such GPUs are used to construct a computer cluster, and the multiple computers included in this computer cluster perform parallel processing. may

The above-described embodiments are preferred embodiments of the present invention, but the scope of the present invention is not limited to the above-described embodiments, and various modifications are made without departing from the scope of the present invention. can be implemented in

In each of the above-described embodiments, the case where one filter is provided has been described, but the filter 130 may be configured by connecting in series a plurality of filters corresponding to different frequency bands. FIG. 15 is a block diagram showing an example of configuring a filter by directly connecting a plurality of filters. In FIG. 15, when there are m resonance points (where m is a natural number of 2 or more), filter 130 is configured by connecting m filters 130-1 to 130-m in series.

In addition, as multiple states in which parameter switching occurs, an example of switching parameters between rapid feed and cutting feed has been described. The parameter may be switched, and the parameter may be switched according to the driving conditions of the machine tool, such as emphasis on accuracy and emphasis on low heat generation.

Examples of the parameters of each state of the plurality of states are the gain of the speed control unit 120 and the coefficient of the filter 130, but other parameters include the position loop gain, the gain of the current control unit 140, the PWM period, and the like. .

Further, in each of the above-described embodiments, an example in which the table, which is the movable portion of the machine, moves one-dimensionally has been described. The present invention can also be applied to the case where the movable portion of moves three-dimensionally.

<Example in which the movable part of the machine moves two-dimensionally (X-axis direction and Y-axis direction)>
FIG. 16 shows two stages 810 and 820 which are movable parts moving in the X-axis direction and the Y-axis direction of the machine. FIG. 17 is a diagram showing the setting of parameters for rapid feed and cutting feed at one representative measuring point A22 and eight measuring points A11 to A13, A21, A23 and A31 to A33. In FIG. 17, a representative measurement point A22 for measuring frequency characteristics with two parameter settings of rapid feed and cutting feed is indicated by a black circle, and a measurement point A11 for measuring frequency characteristics with parameter settings for cutting feed. ˜A13, A21, A23 and A31-A33 are circled.

The frequency characteristic prediction device 10A or frequency characteristic prediction device 10B is provided for each of the X-axis and Y-axis. The frequency characteristic measuring unit 300 of the frequency characteristic prediction device 10A or the frequency characteristic prediction device 10B provided for the X-axis and the Y-axis, respectively, at the representative measurement point A22, at the time of parameter setting for rapid feed and parameter setting for cutting feed Frequency characteristics are measured at measuring points A11 to A13, A21, A23 and A31 to A33 when setting parameters for cutting feed. Then, the frequency characteristic prediction section 500 predicts the frequency characteristic at the time of setting parameters for fast-forwarding at measurement points A11 to A13, A21, A23 and A31 to A33. The double-headed arrows at representative measurement point A22, measurement points A11 to A13, A21, A23 and A31 to A33 shown in FIG. , and the frequency characteristics are measured at a representative measurement point A22, measurement points A11 to A13, A21, A23 and A31 to A33.

<Two tables that are the moving parts of the machine move two-dimensionally (X-axis direction and Y-axis direction), and the moving part that moves the main axis of the machine is independent of the two tables and moves one-dimensionally (Z-axis direction).> Example when moving to>
FIG. 18 is a diagram showing two stages 810 and 820, which are movable parts that move in the X-axis and Y-axis directions of the machine, and a movable part 830 that moves the main shaft of the machine in the Z-axis direction. FIG. 19 is a diagram showing the setting of parameters for rapid feed and cutting feed at two representative measurement points A22 and B33 and nine measurement points A11 to A13, A21, A23, A31 to A33 and C33. In FIG. 19, representative measurement points A22 and B33 for measuring frequency characteristics with two parameter settings of rapid feed and cutting feed are indicated by black circles. Points A11-A13, A21, A23, A31-A33 and C33 are indicated by circles. The movable part 830, which moves the main axis in the Z-axis direction, moves in one dimension (Z-axis direction) independently of the two stages 810 and 820. Measure the frequency characteristics. In FIG. 19, only representative measurement point A22 and measurement point A33 show double-headed arrows for the X and Y axes, and for measurement points A11 to A13, A21, A23 and A31 to A32, double-headed arrows are omitted. The X-axis and Y-axis frequency characteristics are measured at measurement points A11 to A33. In FIG. 19, only the measurement points A33, B33 and C33 are shown with double-headed arrows for the Z axis, indicating that the frequency characteristics are measured for the Z axis.

The frequency characteristic prediction device 10A or frequency characteristic prediction device 10B is provided for each of the X, Y and Z axes. The frequency characteristic measuring unit 300 of the frequency characteristic prediction device 10A or the frequency characteristic prediction device 10B provided for the X-axis, Y-axis, and Z-axis, respectively, measures at representative measurement points A22 and B33 when setting parameters for rapid feed and for cutting feed. At measurement points A11 to A13, A21, A23, A31 to A33 and C33, the frequency characteristics at the time of parameter setting for cutting feed are measured. Then, the frequency characteristic prediction section predicts the frequency characteristic at the time of setting parameters for fast-forwarding at measurement points A11 to A13, A21, A23, A31 to A33 and C33.

<Example of a case where the movable part that moves the main shaft of the machine moves in three dimensions (X-axis direction, Y-axis direction, and Z-axis direction)>
FIG. 20 is a diagram showing movable parts for moving the main shaft of the machine in the X-axis, Y-axis and Z-axis directions. FIG. 21 shows the setting of parameters for rapid feed and cutting feed at one representative measuring point B22 and 26 measuring points A11 to A33, B11 to B13, B21, B23, B31 to B33 and C11 to C33. It is a diagram. In FIG. 21, a representative measurement point B22 for measuring frequency characteristics with two parameter settings of rapid feed and cutting feed is indicated by a black circle, and a measurement point A11 for measuring frequency characteristics with parameter settings for cutting feed. ˜A33, B11-B13, B21, B23, B31-B33 and C11-C33 are indicated by circles.
In FIG. 21, only the representative measurement point A22 shows double-headed arrows for the X, Y, and Z axes, and for the measurement points A11 to A33, B11 to B13, B21, B23, B31 to B33, and C11 to C33. , the double-headed arrows for the X-, Y-, and Z-axes are omitted. The frequency characteristics of the X-axis, Y-axis and Z-axis are measured at all measurement points A11 to C33.
The frequency characteristic prediction device 10A or frequency characteristic prediction device 10B is provided for each of the X, Y and Z axes. The frequency characteristic measuring unit 300 of the frequency characteristic prediction device 10A or the frequency characteristic prediction device 10B provided for each of the X-axis, Y-axis, and Z-axis measures the parameters for fast feed and for cutting feed at the representative measurement point B22. The frequency characteristics during setting are measured at measurement points A11 to A33, B11 to B13, B21, B23, B31 to B33, and C11 to C33 during parameter setting for cutting feed. Then, the frequency characteristic predicting section predicts the frequency characteristic at the time of fast-forward parameter setting at measurement points A11 to A33, B11 to B13, B21, B23, B31 to B33, and C11 to C33.

Also, the embodiments described above can be realized by hardware, software, or a combination thereof. Here, "implemented by software" means implemented by a computer reading and executing a program. When configured by hardware, part or all of each embodiment, for example, integrated circuits such as LSI (Large Scale Integrated circuit), ASIC (Application Specific Integrated Circuit), gate array, FPGA (Field Programmable Gate Array) ( IC).

In addition, when part or all of the above-described embodiments are configured by a combination of software and hardware, a hard disk, ROM, or the like storing a program describing all or part of the operation of the machine learning unit shown in the flowchart In a computer composed of a storage unit, a DRAM that stores data necessary for calculation, a CPU, and a bus that connects each unit, information necessary for calculation is stored in the DRAM, and the program is executed by the CPU. be able to.

Programs can be stored and supplied to computers using various types of computer readable media. Computer readable media includes various types of tangible storage media. Computer-readable media include, for example, magnetic recording media (e.g., hard disk drives), magneto-optical recording media (e.g., magneto-optical discs), CD-ROMs (Read Only Memory), CD-Rs, CD-R/Ws, semiconductor memories (eg, mask ROM, PROM (programmable ROM), EPROM (erasable PROM), flash ROM, or RAM (random access memory)).

The frequency characteristic prediction device and frequency characteristic prediction method according to the present disclosure can take various embodiments having the following configurations, including the embodiments described above.

(1) a motor control unit (for example, motor control unit 100) for moving the axis of the machine tool or industrial machine;
a movement command generation unit (for example, movement command generation unit 200) that outputs a movement command for changing the position of the axis from the first position to the second position to the motor control unit;
a frequency characteristic measuring unit (for example, a frequency characteristic measuring unit 300) that measures the frequency characteristics of the machine tool or the industrial machine at the first position and the second position;
a state switching unit (for example, a state switching unit 600) that switches the state of the motor control unit at the first position;
a frequency characteristic prediction unit (for example, a frequency characteristic prediction unit 500) that predicts the frequency characteristics of the machine tool or the industrial machine at the second position;
The frequency characteristic measuring unit measures, at the first position, a plurality of first frequency characteristics for a plurality of states switched by the state switching unit, and at the second position, among the plurality of states, measuring a second frequency characteristic for at least one of
The frequency characteristic prediction unit uses the plurality of first frequency characteristics and the second frequency characteristic to perform a first prediction regarding a state other than at least one of the plurality of states at the second position. A frequency characteristic prediction device for predicting frequency characteristics of No. 3.
According to this frequency prediction device, the frequency characteristics can be predicted based on the measured frequency characteristics, the number of measurements can be reduced, and the measurement time can be shortened.

(2) The frequency characteristic prediction device according to (1) above, comprising at least a storage unit (for example, storage unit 400) that stores the plurality of first frequency characteristics.

(3) The frequency characteristic prediction device according to (1) or (2) above, wherein at least one value of the gain, filter coefficient, and PWM period of the motor control unit differs among the plurality of states.

(4) The frequency characteristic prediction device according to any one of (1) to (3) above, wherein the first position is set at the center of the movement range of the axis.

(5) the movement command generation unit generates a signal whose frequency changes, inputs the signal to the motor control unit;
The frequency characteristic measuring unit uses the signal and the output signal of the motor control unit to obtain an amplitude ratio and a phase delay between the signal and the output signal for each frequency defined by the signal. The frequency characteristic prediction device according to any one of (1) to (4) above, which measures frequency characteristics.

(6) The frequency characteristic prediction device according to (5) above, wherein the frequency characteristic measurement unit can change at least one of the amplitude of the signal, the number of vibrations, and the vibration method.

(7) A parameter adjuster (for example, parameter The frequency characteristic prediction device according to any one of (1) to (6) above, including an adjustment unit 800).

(8) The frequency characteristic prediction device according to (7) above, wherein the parameters of the motor control section include at least one of a gain of a speed control section, a coefficient of a filter, a gain of a current control section, and a PWM period.

(9) The frequency characteristic prediction device according to (7) or (8) above, wherein the parameter adjustment unit determines the parameters of the motor control unit using reinforcement learning.

(10) moving the position of the axis of the machine tool or the industrial machine to the first position by the motor control unit (for example, the motor control unit 100) based on the first movement command;
At the first position, a plurality of first frequency characteristics for a plurality of states of the motor control unit are measured by a frequency characteristic measuring unit (e.g., frequency characteristic measuring unit 300);
moving the position of the axis from the first position to the second position by the motor control unit based on a second movement command;
At the second position, the frequency characteristic measuring unit measures a second frequency characteristic for at least one of the plurality of states;
Using the plurality of first frequency characteristics and the second frequency characteristics, a frequency characteristic prediction unit (for example, frequency characteristic prediction unit 500) predicts the plurality of frequencies at the second position. A frequency characteristic prediction method for predicting a third frequency characteristic for a state other than at least one of the states of .
According to this frequency prediction method, the frequency characteristics can be predicted based on the measured frequency characteristics, the number of measurements can be reduced, and the measurement time can be shortened.

10A, 10B frequency characteristic prediction device 100 motor control section 200 movement command generation section 300 frequency characteristic measurement section 400 storage section 500 frequency characteristic prediction section 600 state switching section 700 controlled object 800 parameter adjustment section

Claims (10)

1. a motor control for moving an axis of a machine tool or industrial machine;
   a movement command generation unit that outputs a movement command to the motor control unit for changing the position of the shaft from the first position to the second position;
   a frequency characteristic measuring unit that measures the frequency characteristic of the machine tool or the industrial machine at the first position and the second position;
   a state switching unit that switches the state of the motor control unit at the first position;
   a frequency characteristic prediction unit that predicts the frequency characteristic of the machine tool or the industrial machine at the second position;
   The frequency characteristic measuring unit measures, at the first position, a plurality of first frequency characteristics for a plurality of states switched by the state switching unit, and at the second position, among the plurality of states, measuring a second frequency characteristic for at least one of
   The frequency characteristic prediction unit uses the plurality of first frequency characteristics and the second frequency characteristic to perform a first prediction regarding a state other than at least one of the plurality of states at the second position. A frequency characteristic prediction device for predicting frequency characteristics of No. 3.
2. 2. The frequency characteristic prediction device according to claim 1, comprising a storage unit that stores at least the plurality of first frequency characteristics.
3. The frequency characteristic prediction device according to claim 1 or 2, wherein at least one value of the gain, filter coefficient, and PWM period of the motor control unit differs among the plurality of states.
4. The frequency characteristic prediction device according to any one of claims 1 to 3, wherein the first position is set at the center of the movement range of the axis.
5. The movement command generation unit generates a signal whose frequency changes, inputs the signal to the motor control unit,
   The frequency characteristic measuring unit uses the signal and the output signal of the motor control unit to obtain an amplitude ratio and a phase delay between the signal and the output signal for each frequency defined by the signal. 5. The frequency characteristic prediction device according to any one of claims 1 to 4, which measures frequency characteristics.
6. 6. The frequency characteristic prediction device according to claim 5, wherein the frequency characteristic measurement unit can change at least one of the amplitude of the signal, the number of vibrations, and the vibration method.
7. 2. A parameter adjustment unit that determines a parameter of the motor control unit based on the plurality of first frequency characteristics or the second frequency characteristics, or based on the third frequency characteristic. 7. The frequency characteristic prediction device according to any one of 6.
8. 8. The frequency characteristic prediction device according to claim 7, wherein the parameters of the motor control section include at least one of a speed control section gain, a filter coefficient, a current control section gain, and a PWM period.
9. The frequency characteristic prediction device according to claim 7 or 8, wherein the parameter adjustment unit determines parameters of the motor control unit using reinforcement learning.
10. moving the position of the axis of the machine tool or the industrial machine to the first position by the motor control unit based on the first movement command;
    measuring a plurality of first frequency characteristics with respect to a plurality of states of the motor control unit at the first position;
    moving the position of the axis from the first position to the second position by the motor control unit based on a second movement command;
    At the second position, the frequency characteristic measuring unit measures a second frequency characteristic for at least one of the plurality of states;
    a frequency characteristic prediction unit, using the plurality of first frequency characteristics and the second frequency characteristics, by the frequency characteristic prediction unit, at least one of the plurality of states at the second position other than A frequency characteristic prediction method for predicting a third frequency characteristic for the state of

Images