JP2006333594A - Mechanical characteristic modeling apparatus and method, motor controller, and machine control system - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To consider friction so as to cause no errors in machine modeling. <P>SOLUTION: This modeling apparatus includes a frequency characteristic calculation apparatus 6, a machine modeling estimation apparatus 10, a frequency characteristic peak detector 11, and a damping estimated value analyzer 12. A mechanical property model 7 includes a rigid body load model 8, a friction model 9, and a vibration system model 17 having a resonance model 18, an antiresonance model, and a damping model 20 to determine a machine model from frequency characteristics by means of a method of least squares or curve fitting. <P>COPYRIGHT: (C)2007,JPO&INPIT

Description

  The present invention relates to a motor control device used in, for example, a positioning device such as a semiconductor manufacturing device or a machine tool or an industrial robot, and in particular, a machine model estimation of a motor control device that grasps a control target for optimal servo adjustment. The present invention relates to an apparatus and a mechanical property modeling method.

  A conventional machine model estimation device of an electric motor control device is an inexpensive electric motor that can easily estimate a machine model without using an expensive measuring device and without an operator having advanced specialized knowledge and experience. In order to provide a machine model estimation device for a control device, in an electric motor control device comprising: an electric motor that drives a load machine; a rotation detector that detects the rotation angle of the electric motor; and a servo control device that controls the electric motor. The device automatically calculates the shape of the protrusions having the resonance frequency and the anti-resonance frequency from the frequency characteristics obtained from the operation command signal and the rotation detector signal, and is calculated by the frequency characteristic formulas of the two inertia model and the rigid body model. For each measured frequency characteristic, an error from the measured frequency characteristic is calculated, and the calculated and measured values in the frequency characteristics of the two-inertia model and the rigid body model are calculated. And so as to automatically model the characteristics of the machine by comparing the error, respectively (e.g., see Patent Document 1).

  Furthermore, in order to reduce inertia identification errors with a mechanism with large Cron friction, etc., the conventional motor control device reduces each torque in the steady state with a forward / reverse command with a low speed command and the same size. The constant disturbance torque Td is calculated from the commands Tref1 and Tref2 using the equation (200).

  Td = (Tref1 + Tref2) / 2 Formula (200)

  A Coulomb friction Tc is calculated from the value obtained by subtracting the calculated constant disturbance torque Td from each torque command Tref3 and Tref4 in a steady state of a certain speed command Vref and an α-times speed command αVref, from Equation (201),

  Tc = {α (Tref3-Td)-(Tref4-Td)} / (α-1) Expression (201)

  Viscosity friction Dc is calculated from equation (202) from each torque command Tref3, Tref4 and velocity Vfb3, Vfb4 in a steady state of a certain velocity command Vref and a velocity command αVref that is α times that,

  Dc = (Tref3-Tref4) / (Vfb3-Vfb4) Formula (202)

  In some cases, the static friction torque Tg is calculated from a torque command obtained by adding these to the model and a torque command until the motor speed starts from 0 (for example, see Patent Document 2).

(First conventional example)
In the first conventional example, in order to determine the mechanical model from the measured frequency characteristic value, a rigid body model and a two-inertia model are prepared in advance as mechanical characteristic modeling, and the difference between the model and the measured frequency characteristic value is determined. Whether the model is a rigid body model or a two-inertia model is determined.
FIG. 12 is an overall configuration diagram of an electric motor control device including a machine model estimation device according to a first conventional example. In FIG. 12, 101 is an arithmetic unit, 102 is a servo control device, 103 is a rotation detector, 104 is an electric motor, 105 is a transmission mechanism, 106 is a movable part, 107 is a non-movable part, 108 is an operation command signal, and 109 is a rotation. Detector signal, 110 is a control signal, 119 is an input device, 120 is a frequency characteristic equation, and 121 is an output device. The measured frequency characteristic is compared with the frequency characteristic formula 120.
FIG. 13 is a diagram showing an example of the unsuitability for curve fitting of the rigid model formula of the first conventional example.
FIG. 13 compares the measured frequency characteristic value with the rigid body model, and is an example in which the measured frequency characteristic value does not match the rigid body model at the location where the resonance frequency exists.

  As described above, the machine model estimation device of the conventional motor control device of the first example measures the frequency characteristic value and compares the frequency characteristic value with the rigid body model or the two-inertia model to make the machine model.

(Conventional second example)
FIG. 14 is a block diagram for explaining the basic concept of the second conventional example.
If the speed Vfb of the speed control unit and the speed Vfb 'of the model speed control unit are the same and neither Vfb nor Vfb' is zero, the torque command integral value STref, inertia J of the speed control unit and the model speed control unit The relationship of the formula (203) is established between the torque command integral value STref ′ and the inertia J ′.

J / J '= STref / STref' (203)

Inertia J is immediately obtained from equation (204).

J = (STref / STref ') * J' (204)

However, in general, there are viscous friction Dc, constant disturbance torque Td, cron friction Tc, and static friction torque Tg, and the influences thereof are removed.
(A) For the viscous friction Dc, the integral value of the velocity Vfb in the predetermined section [a, b] may be zero.
(B) For the constant disturbance torque Td, the average value of the inertia J1 obtained from a certain speed command Vref1 and the inertia J2 obtained from Vref2 obtained by inverting the sign of the speed command Vref1 do it.
(C) For the Cron friction Tc, the section [a, b] and the speed command Vref are set so that the forward rotation time and the reverse rotation time of the motor in the section [a, b] are equal. You only have to set it.
(D) For the static friction torque Tg, time integration of the torque command is performed except for the case where the speed Vfb is expressed by the equation (205) in the section [a, b].

X1 = Vfb = X2 Formula (205)

However, X1 ≠ 0, X2 ≠ 0
The above is a method for obtaining an inertia which is hardly affected by constant disturbance torque, Coulomb friction, viscous friction and static friction torque.

  As described above, the conventional motor control device estimates the inertia (load inertia moment) by estimating the viscous friction, the constant disturbance, the clone friction, and the static friction to remove the influence.

JP 2003-79174 A (FIGS. 1 and 12) JP-A-11-46489 (FIG. 2)

  Since the conventional machine model estimation device of an electric motor control device discriminates between a rigid body and a vibration system (two-inertia system) and cannot consider friction, if the friction is large, the estimation accuracy of the rigid body component is high. There was a problem of lowering. Further, when the friction is very large, there is a problem that the rigid body component cannot be estimated because the characteristic of the rigid system is lost in the low frequency region of the gain.

  Further, in order to estimate the inertia (moment of inertia), the conventional motor control device estimates the friction that is an estimation error in advance and removes the influence of the friction, so that the resonance of the machine to be controlled is reduced. Is not considered at all, and there is a problem that the model of the vibration system is not understood at all.

  The present invention has been made in view of such problems, and estimates the load moment of inertia and load mass in consideration of the friction characteristics, and considers the resonance and damping characteristics of the vibration system, and the mechanical characteristics model. An object of the present invention is to provide a mechanical characteristic modeling device and a mechanical characteristic modeling method for an electric motor control apparatus that can estimate the motor.

In order to solve the above problem, the present invention is configured as follows.
According to the first aspect of the present invention, the motor control device including command means for creating an operation command for the motor that drives the machine drives the machine, the detection means detects the operation amount of the machine, and the detection In the mechanical characteristic modeling apparatus for modeling the characteristics of the machine, the frequency characteristic calculating apparatus for obtaining the frequency characteristic of the machine by calculating the operation amount and the operation command are provided. A quantity of rigid body load model, a mechanical characteristic model including a friction model in which friction is quantified, and a mechanical model estimation means for calculating the mechanical characteristic model including the rigid body load model and the friction model from the frequency characteristic, It is what you have.

  The invention according to claim 2 is characterized in that the mechanical model estimating means directly calculates the rigid body load model and the friction model from the frequency characteristic.

  Further, in the invention according to claim 3, the mechanical model estimation means is adapted to curve-fit the frequency characteristic with the mechanical characteristic model including the rigid body load model and the friction model, and the rigid body load model and the The mechanical property model including a friction model is calculated.

  According to a fourth aspect of the present invention, the mechanical model estimation means converges the amount of change of the mechanical characteristic model including the rigid body load model and the friction model that are fitted to a curve so as to approach the frequency characteristic by iterative calculation. It is characterized by making it.

  The invention according to claim 5 is characterized in that the detection means detects the position, speed or acceleration of the electric motor, or the position, speed or acceleration of the machine.

  The invention described in claim 6 is characterized in that the output of the command means is a swept sine wave.

  The invention according to claim 7 is characterized in that the output of the command means is an M-sequence signal.

  The invention according to claim 8 is characterized in that the output of the command means is a random wave.

  The invention according to claim 9 is characterized in that the electric motor is a rotary motor, and the rigid load model which is a load amount of the electric motor and the machine is an inertia moment.

  The invention described in claim 10 is characterized in that the electric motor is a translational motor, and the rigid body load model which is a load amount of the electric motor and the machine is a mass.

  The invention according to claim 11 is characterized in that the mechanical model estimating means has frequency characteristic peak detecting means for automatically calculating a protrusion shape comprising a resonance frequency and an anti-resonance frequency of the frequency characteristic. Is.

  The invention according to claim 12 is characterized in that the mechanical model estimation means includes attenuation estimated value analysis means for estimating attenuation from the resonance frequency or anti-resonance frequency detected by the frequency characteristic peak detection means. Is.

  In the invention according to claim 13, the mechanical characteristic model includes a vibration system model including a resonance model and a damping model input in advance to the mechanical model estimation unit, and the mechanical model estimation unit includes the frequency A vibration system model including the resonance model and the damping model is calculated from the characteristics.

  Further, in the invention described in claim 14, from the plurality of frequency characteristics measured by using the plurality of detection means, the machine model estimation means, means the frequency characteristic peak detection means, and the attenuation estimated value analysis means include: The load model and the friction model, the resonance frequency and the anti-resonance frequency, and the attenuation are estimated, respectively.

  The invention according to claim 15 is characterized in that the mechanical model estimating means further includes a vibration mode estimating means for estimating a vibration mode from the plurality of frequency characteristics.

  The invention described in claim 16 includes at least two or more of the electric motor control device, the electric motor, and the detection unit, and the frequency characteristic calculation device measures at least two or more of the frequency characteristics. It is a feature.

  The invention described in claim 17 is characterized by comprising output means, input means, or storage means.

  The invention according to claim 18 is characterized in that the mechanical model estimating means calculates the mechanical characteristic model by arbitrarily specifying at least two points of the frequency characteristic and limiting a frequency range. It is.

  The invention according to claim 19 is an electric motor comprising a machine, an electric motor that drives the machine, command means for creating a control command for the motor, and a controller that receives the control command and drives the motor. In the mechanical characteristic modeling method of the electric motor control system comprising the control device, the command means creates an operation command of the electric motor, and measures the frequency characteristic of the machine driven by the operation command; And a step of deriving a mechanical characteristic model including a rigid body load model and a friction model by calculating a mechanical characteristic model from the frequency characteristic.

  The invention according to claim 20 is characterized in that the mechanical characteristic model includes a vibration system model including a resonance model and a damping model.

  The invention according to claim 21 has a step of calculating a vibration system model from the frequency characteristic.

  Further, in the invention according to claim 22, the step of calculating the vibration system model automatically detects a protrusion shape including a resonance frequency and an anti-resonance frequency of the frequency characteristic to detect a peak of the frequency characteristic, It has the step which estimates a resonant frequency or an antiresonance frequency, It is characterized by the above-mentioned.

  The invention according to claim 23 is characterized in that the step of calculating the vibration system model includes a step of estimating attenuation from a resonance frequency or an anti-resonance frequency of the frequency characteristic.

  The invention described in claim 24 is characterized in that the step of calculating the vibration system model includes a step of calculating a vibration mode using a plurality of the frequency characteristics.

  According to a twenty-fifth aspect of the present invention, the mechanical property modeling apparatus according to the first aspect is provided.

  The invention described in claim 26 is controlled by the control device described in claim 25, and the characteristic is modeled.

  According to a twenty-seventh aspect of the present invention, there is provided an electric motor control device including the mechanical characteristic modeling device according to the first aspect, an electric motor controlled by the electric motor control device, and a machine driven by the electric motor. It is characterized by.

  According to the first aspect of the present invention, the frequency characteristic of the machine can be measured, and a mechanical characteristic model including a rigid load model and a friction model can be obtained from the measured frequency characteristic.

  According to the invention described in claim 2, a mechanical characteristic model including a rigid load model and a friction model can be directly obtained from the measured frequency characteristic.

  According to the third aspect of the present invention, the mechanical characteristic model including the rigid body load model and the friction model can be obtained while comparing with the measured frequency characteristic.

  According to the invention described in claim 4, the mechanical characteristic model including the rigid body load model and the friction model is compared with the measured frequency characteristic, and the calculation is repeatedly performed so as to converge so that the difference is reduced. Can do.

  Further, according to the invention described in claim 5, since the detecting means can detect the position, speed or acceleration of the electric motor, or the position, speed or acceleration of the machine, various places and types of the detecting means can be selected. Thus, the frequency characteristic can be measured.

  Further, according to the invention described in claim 6, since the swept sine wave can be selected, signals corresponding to various situations can be used when measuring the frequency characteristics.

  In addition, according to the seventh aspect of the invention, since the M-sequence signal can be selected, a signal corresponding to various situations can be used when measuring the frequency characteristic.

  According to the invention described in claim 8, since a random wave can be selected, signals corresponding to various situations can be used when measuring frequency characteristics.

  According to the ninth aspect of the present invention, a rotary motor can be used for the electric motor.

  According to the invention described in claim 10, a translational linear motor can be used for the electric motor.

  According to the eleventh aspect of the present invention, the resonance frequency or anti-resonance frequency can be detected using the measured frequency characteristic.

  According to the invention of claim 12, attenuation can be estimated from the resonance frequency or the anti-resonance frequency using the measured frequency characteristic.

  According to the invention described in claim 13, the resonance frequency and attenuation obtained using the measured frequency characteristics can be quantified as a resonance model and an attenuation model.

  According to the fourteenth aspect of the present invention, since a plurality of frequency characteristics can be measured, resonance that cannot be detected by one frequency characteristic can be grasped. Further, since the friction and the rigid load are also obtained as an average of a plurality of frequency characteristics, the reliability can be quantified with increased reliability such as a rigid load model, a friction model, a resonance model, and a damping model.

  According to the fifteenth aspect of the present invention, since the vibration mode can be obtained, it is possible to grasp how the entire machine vibrates and to grasp the characteristics of the machine in detail.

  Further, according to the invention described in claim 16, in a configuration including a plurality of electric motors, frequency characteristics including interference between shafts can be measured.

  According to the seventeenth aspect of the present invention, the output of each function can be grasped, each function can be instructed, and the output of each function and the instruction input to the function can be stored.

  According to the invention described in claim 18, since the range for obtaining the mechanical characteristic model can be selected from the measured frequency characteristic, the work can be executed efficiently and accurately.

  According to the nineteenth aspect of the invention, the frequency characteristic can be measured, and a mechanical characteristic model including a rigid load model and a friction model can be obtained from the measured frequency characteristic.

  According to the inventions of claims 20 and 21, a mechanical characteristic model including a rigid load model and a friction model can be obtained from the measured frequency characteristics.

  According to the twenty-second aspect of the present invention, the resonance frequency or anti-resonance frequency can be detected using the measured frequency characteristic.

  According to the invention of claim 23, the attenuation can be estimated from the resonance frequency or the anti-resonance frequency using the measured frequency characteristic.

  According to the invention described in claim 24, since the vibration mode can be obtained, it is possible to grasp how the entire machine vibrates and to grasp the characteristics of the machine in detail.

  According to the invention described in claims 25, 26, and 27, the frequency characteristic of the machine can be measured, and a mechanical characteristic model including a rigid load model and a friction model can be obtained from the measured frequency characteristic.

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

FIG. 1 is a flowchart of a mechanical characteristic modeling method for an electric motor control apparatus according to a first embodiment of the present invention.
STEP 10 is a step of measuring frequency characteristics, STEP 20 is a step of calculating a mechanical characteristic model from the frequency characteristics, and STEP 30 is a step of calculating a vibration system model from the frequency characteristics.
The step 30 for calculating the vibration system model from the frequency characteristics further includes two steps, STEP 31 is a step for estimating the resonance frequency or anti-resonance frequency, and STEP 32 is a step for estimating the attenuation from the resonance frequency or anti-resonance frequency. is there.

FIG. 2 is an overall configuration diagram of the mechanical characteristic modeling device for the motor control device according to the first embodiment of the present invention.
FIG. 3 is a block diagram of the mechanical characteristic modeling device for the motor control device according to the first embodiment of the present invention.
In the figure, 1 is an electric motor (linear motor), 2 is a detection means, 3 is a controller, 4 is a command means, 5 is a machine, 6 is a frequency characteristic calculation device, 7 is a mechanical characteristic model, 8 is a rigid load model, 9 Is a friction model, 10 is a machine model estimation means, 11 is a frequency characteristic peak detection means, 12 is an attenuation estimated value analysis means, 14 is an output means, 15 is an input means, 16 is a storage means, 17 is a vibration system model, 18 is A resonance model, 19 is an anti-resonance model, and 20 is a damping model.

  The part where the present invention is different from Patent Document 1 is a part provided with a friction model 9.

  The present invention differs from Patent Document 2 in that the frequency characteristic calculation device 6, the machine model estimation means 10, the frequency characteristic peak detection means 11, the estimated attenuation value analysis means 12, the input means 15, the storage means 16, and the vibration system model 17. As shown, the resonance model 18, the antiresonance model 19, and the damping model 20 are provided.

The mechanical property modeling method proceeds in three steps as shown in FIG.
In step STEP 10 for measuring the frequency characteristics, the output of the detection means 2 is received, the feedback loop formed by the controller 3 is cut, and as an open loop, the command means 4 sends the sweep sine wave C to the controller 3 as a thrust command T. Then, the electric motor 1 (linear motor) is driven.
When the electric motor 1 is driven, the machine 5 is shaken, and the detection means 2 detects the response r of the amount of movement.
The thrust command T or the swept sine wave C and the response r are given to the frequency characteristic calculator 6. The response r may be given to the frequency characteristic calculation device 6 via the controller 3.
The frequency characteristic calculation device 6 performs frequency analysis on the thrust command T or the swept sine wave C and the response r by, for example, FFT (Fast Fourier Transform).
The thrust command T or the swept sine wave C and the response r are divided into a plurality of times and given to the frequency characteristic calculation device 6, frequency analysis is performed, and averaged auto power spectrum and cross spectrum are calculated.
From the auto power spectrum Gxx of the swept sine wave C, which is an averaged thrust command T, and the averaged response r, and the cross spectrum Gyx of the swept sine wave C, which is the thrust command T, the frequency is expressed by Equation (1). obtaining a characteristic _{H M.}

H _M = Gyx / Gxx Formula (1)

Note that the averaged auto power spectrum and cross spectrum may be obtained by dividing the sweep sine wave C as the thrust command T and the response r into multiple times as described above, or the sweep sine as the thrust command T. The wave C and the response r may be measured for a long time, the frequency may be divided into a plurality of parts, and the averaged auto power spectrum and cross spectrum may be obtained.
In step STEP 10 for measuring frequency characteristics, the frequency characteristics are measured as described above.

FIG. 4 is a diagram of frequency response characteristics having friction showing the first embodiment of the present invention.
In step STEP10, frequency response characteristics as shown in FIG. 4 are measured.
In step STEP 20 for calculating the mechanical characteristic model from the frequency characteristic, a rigid body load model and a friction model are obtained from the influence of mass and friction appearing in the low frequency region from the frequency characteristic obtained in step STEP 10.
FIG. 4 has one anti-resonance and resonance, and the gain in the low frequency region changes due to friction.
When there is no friction, if the frequency on the horizontal axis is logarithmically expressed, the gain falls linearly from the upper left to the lower right. This inclination indicates the mass, that is, the rigid load. When this part is extracted, Equation (2) Become.

Here, H _R: frequency characteristic of rigid load velocity response, J: mass (rigid load), s: shows a Laplace operator.

FIG. 5 is a diagram showing a rigid system and a two-inertia system according to the first embodiment of the present invention.
Equation (1) can be expressed by a model as shown in FIG.
FIG. 4 is a graph of a two-inertia system, which can be expressed by a model as shown in FIG. 5B, and is expressed by equation (3) when there is no friction.

Here, H _F : frequency characteristics of a two-inertia system of speed response, J ₁ : mass on the motor side, J ₂ : mass on the machine side, C: damping, K: rigidity are shown.
Further, the sum of the mass on the electric motor side and the mass on the machine side is the total mass, which is expressed by Equation (4).

J = J ₁ + J ₂ formula (4)

  The resonance frequency fr is expressed by equation (5), and the anti-resonance frequency fa is expressed by equation (6).

That is, even in the case of the frequency characteristic of the two inertia system, if it is limited to the low frequency region, it indicates a rigid load and the two coincide. (It is shown in FIG. 14 of the first conventional example.)
For this reason, the rigid load, that is, the mass can be obtained by approximating the slope of the gain graph in the low frequency region to the equation (1).
However, if friction exists as shown in FIG. 4, the graph is not a straight line, and the accuracy of the obtained mass may be reduced if it is left as it is.
Therefore, a method will be described in which the mechanical model estimation means 10 obtains a rigid load, that is, a mass in consideration of the presence of friction.

The friction includes coulomb friction, viscous friction, and static friction, but coulomb friction and static friction that appear as a constant drift have little influence on the frequency characteristics, and thus viscous friction may be considered.
FIG. 6 is a diagram showing a rigid system model having friction according to the first embodiment of the present invention.
Since a thrust proportional to the speed is generated, FIG. 6 is transformed into the expression (7) by modifying the expression (2).

Here, H ′ _{R is} a frequency characteristic of a rigid body load with a speed response considering friction, D: a (viscous) friction characteristic, and M ′ is a reciprocal of a frequency characteristic of a rigid body load with a speed response considering friction.

  Even when the frequency characteristic including friction in FIG. 4 is measured, the mass J can be obtained in consideration of the viscous friction characteristic D by using the equation (7).

The reciprocal of the measured frequency characteristic H, as shown in equation (8), if M, may be a difference ε between M and H 'inverse of _R M' to a minimum. That is, by solving the equation (9) using the least square method, the mass and the friction are obtained as in the equations (10) and (11).

Here, n represents the number of frequency characteristic data (number of lines in the frequency domain) used in the least square method.

The number of frequency characteristic data (number of lines in the frequency domain) n may be selected from two arbitrary points such as AA ′ and A ″ -A ′ in FIG. 4 to estimate the mass J and the friction characteristic D. .
Further, an anti-resonance described later may be estimated, and the mass J and the friction characteristic D may be estimated in a frequency region lower than the anti-resonance.

In the above example, the least square method is used. However, the mass J and the friction characteristic D may be estimated by fitting the equation (7) to the measured frequency characteristic.
The initial value may be the result of the above least square method, or the initial value is determined in advance, and the mass J and the friction are reduced so that the difference between the measured frequency characteristic and the result of curve fitting according to the equation (7) becomes small. The characteristic D may be estimated.

The frequency characteristic H _M measured, the equation (7) so as to approach the difference ΔH results H _R of curve fitting by, since changing ΔP parameters P comprising a mass J and friction characteristics D,
Results and the difference ΔH of the partial differential of the parameter of the result H _R of Formula (7) by curve fitting P (J, D), the change amount ΔP of the parameters P, a relationship of Equation (12).

If the said Formula (12) is written in detail, it will become Formula (13).

Here, X ^{R represents} a real component of the complex value X, and X ^I represents an imaginary component of the complex value X.

This, for example, solving the Gauss-Newton method, the change amount ΔP of the parameters P, the formula (7) Results from H _R approaches the frequency characteristic H _M measured values of curve fitting, that is, change in mass J The amount ΔP and the change amount ΔD of the friction characteristic D are known.
The variation [Delta] P, the mass at [Delta] D J, if modified by curve fitting the friction characteristics D, a value that matches the frequency characteristic H _M measured.
It is also possible to determine the amount of change ΔP of the parameter P with a weight W to the difference ΔH results H _R curve fitting by the formula consisting of the frequency characteristic H _M and current value measured (7). In this case, in consideration of the equation (14), ΔH ′ is used in place of ΔH in the equation (13).

  [ΔH ′] = [ΔH] · [W] Equation (14)

However, as mentioned above, there are cases where the parameters even lead to a shift amount ΔP of P, the result H _R curve fitting by the formula (7) does not approach the frequency characteristic H _M measured. The variation [Delta] P, but the rate of change of the result H _R of the curve fit high, there are cases where ΔH is increased. In this case, the change amount ΔP of the parameter P is converged by iterative calculation. The parameter is sequentially updated as in Expression (15), and the converged change amount ΔP is derived.

P _n = P _n-1 + ΔP Formula (15)

  When ΔH increases from the previous value, the parameter P is sequentially updated as in Expression (16) so that ΔH decreases.

P _n = P _n-1 + e · ΔP Formula (16)

Where 0 <e <1

As described above, the machine model estimation means 10 estimates the mass J and the friction characteristic D from the measured frequency characteristics.
That is, since the mass J and the friction characteristic D are estimated, the mechanical model estimation means 10 can model the rigid body load model 8 and the friction model 9 which are the mechanical characteristic model 7.

  In step STEP30 for calculating the vibration system model from the frequency characteristics, there are further two steps. From the frequency characteristics measured in step STEP31 for estimating the resonance frequency or the antiresonance frequency, the resonance / antiresonance frequency is obtained from the frequency characteristic peak detecting means 11. Step 32 where the attenuation is estimated from the resonance frequency or the anti-resonance frequency is calculated by the attenuation estimated value analysis means 12.

  In step STEP 31 for estimating the resonance frequency or anti-resonance frequency, if there is mechanical resonance in the measured frequency characteristics, the gain (amplitude) appears uneven, the convex side is resonant, and the concave side is anti-resonant. If the frequency characteristic peak detecting means 11 detects the peak of the unevenness of the gain (amplitude), the resonance / antiresonance frequency can be estimated.

Next, in step STEP 32 for estimating attenuation from the resonance frequency or anti-resonance frequency,
If the resonance / antiresonance frequency can be detected, the attenuation is a convex or concave angle, so that the attenuation estimated value analysis means 12 can estimate the attenuation from the resonance / antiresonance peak and its peripheral values.

Since the resonance / anti-resonance frequency is estimated in step STEP 31 for estimating the resonance frequency or anti-resonance frequency, and the attenuation can be estimated in step STEP 32 for estimating the attenuation from the resonance frequency or anti-resonance frequency, the resonance model 18 which is the vibration system model 17. The anti-resonance model 19 and the attenuation model 20 can be obtained.
As estimated from the low frequency region AA ′ and A ″ -A ′ in FIG. 4 of the frequency characteristic, the mass J and the friction characteristic D are estimated, as shown in BB ′ in FIG. May be selected to estimate the resonance / anti-resonance frequency and estimate the attenuation.

  As described above, step STEP 10 for measuring the frequency characteristics, step STEP 20 for calculating the mechanical characteristic model from the frequency characteristics, step STEP 31 for estimating the resonance frequency or anti-resonance frequency, and the attenuation estimation from the resonance frequency or anti-resonance frequency. By carrying out step STEP 30 of calculating the vibration system model from the frequency characteristic consisting of step STEP 32, the frequency characteristic calculation device 6 measures the frequency characteristic, and the mechanical model estimation means 10 is a rigid body load model 8 that is the mechanical characteristic model 7. The friction model 9 can be obtained, and the frequency characteristic peak detection means 11 and the attenuation estimated value analysis means 12 can obtain the resonance model 18, the antiresonance model 19, and the attenuation model 20 which are the vibration system models 17.

FIG. 7 is a flowchart of the mechanical characteristic modeling method for the motor control device according to the second embodiment of the present invention.
STEP 10 is a step of measuring frequency characteristics, STEP 20 is a step of calculating a mechanical characteristic model from the frequency characteristics, and STEP 30 is a step of calculating a vibration system model from the frequency characteristics.
Step 31 of calculating vibration system model from frequency characteristics STEP 31 of STEP 30 is a step of estimating resonance frequency or anti-resonance frequency, STEP 32 is a step of estimating attenuation from resonance frequency or anti-resonance frequency, and STEP 33 is a step of calculating vibration mode. .

FIG. 8 is an overall configuration diagram of a mechanical characteristic modeling device for an electric motor control device according to a second embodiment of the present invention.
FIG. 9 is a block diagram of a mechanical characteristic modeling device for an electric motor control device according to a third embodiment of the present invention.
In the figure, 1 is an electric motor, 2 is a detection means, 3 is a controller, 4 is a command means, 5 is a machine, 6 is a frequency characteristic calculation device, 7 is a mechanical characteristic model, 8 is a rigid load model, 9 is a friction model, 10 is a machine model estimation means, 11 is a frequency characteristic peak detection means, 12 is an attenuation estimated value analysis means, 13 is a vibration mode estimation means, 14 is an output means, 15 is an input means, 16 is a storage means, 17 is a vibration system model , 18 is a resonance model, 19 is an anti-resonance model, 20 is a damping model, and 21 is a vibration mode model.
The detection means 2 includes 2 a for detecting the operation amount of the electric motor 1, 2 b for detecting the operation amount of the movable part of the machine 5, and accelerometers 2 c to 2 m attached to the machine 5.
The accelerometers 2d and 2f detect the acceleration in the movable direction of the movable part of the machine 5, and the acceleration 2e is attached to the movable part of the machine 5, but is attached in a direction perpendicular to the movable direction, and the accelerometers 2c and 2g. Reference numerals 2h and 2m are attached to a non-movable support portion of the machine 5.

This embodiment is different from the first embodiment in that the electric motor 1 is a rotary motor, and the detection means 2 detects a motor operation amount associated with the rotary motor, and the movable part of the machine 5. This is the point that it has a fully closed configuration with a linear encoder that detects the amount of movement, and the point that a plurality of accelerometers are attached to the machine 5.
Since the electric motor 1 is a rotary motor, the load J and the rigid load model shown in the first embodiment are the moment of inertia J.
Moreover, the point which included step STEP33 which calculates vibration mode in step STEP30 which calculates a vibration system model from a frequency characteristic differs from 1st Example.

  The part where the present invention differs from Patent Document 1 is a part provided with a friction model 9, a plurality of detection means 2, a vibration mode estimation means 13, and a friction model 9.

  The present invention differs from Patent Document 2 in that the frequency characteristic calculation device 6, the machine model estimation means 10, the frequency characteristic peak detection means 11, the attenuation estimated value analysis means 12, the vibration mode estimation means 13, the input means 15 and the storage means 16 The vibration system model 17 includes a resonance model 18, an anti-resonance model 19, a damping model 20, and a vibration mode model 21.

As shown in FIG. 7, the mechanical property modeling method proceeds in three steps as in the first embodiment.
Step STEP10 for measuring the frequency characteristics is made open-loop as in the first embodiment, and the command means 4 gives the sweep sine wave C as a torque command T to the controller 3 to drive the motor 1 and its operation amount. The detecting means 2 detects the response r of the amount shaken by the machine 5.
Unlike the first embodiment, the detection means 2a, 2b and the accelerometer detection means 2d, 2f, 2e, 2c, 2g, 2h, 2m also detect the response ri.
Nine types of frequency characteristics can be measured by giving the sweep characteristic sine wave C, which is the torque command T, and the response ri to the frequency characteristic calculator 6. The responses ra and rb may be given to the frequency characteristic calculation device 6 via the controller 3.

Each frequency characteristic can be calculated in the same manner as in the first embodiment.
The averaged cross spectrum may be obtained from the nine types of responses ri and the thrust command T or the swept sine wave C, and the nine types of frequency response characteristics HM _{, i} may be obtained from the above equation (1).
If the accelerometer detection means 2d, 2f, 2e, 2c, 2g, 2h, and 2m are detachable, a frequency characteristic is further added by changing the measurement point for step STEP33 for calculating the vibration mode. May be obtained.

In step STEP20 for calculating the mechanical characteristic model from the frequency characteristic, the rigid body load model and the friction model are calculated from the influence of the moment of inertia and friction appearing in the low frequency region from the frequency characteristic obtained in step STEP10, as in the first embodiment. obtain.
However, the frequency characteristic H _Me obtained from the acceleration 2e attached to the direction perpendicular to the moving direction of the movable part of the machine 5 and the accelerometers 2c, 2g, 2h, 2m attached to the non-movable support part of the machine 5. , H _Mc , H _Mg , H _Mh , and H _Mm are not affected by the rigid body component and are not used for estimation of the rigid body load model and the friction model.
The machine model estimating means 10 estimates the load inertia moment J and the friction characteristic D using the detecting means 2a and 2b and the accelerometers 2d and 2f for detecting the acceleration in the moving direction of the moving part of the machine 5.

The detection means 2a detects the operation amount of the speed of the electric motor 1, the detection means 2b detects the operation amount of the position and displacement of the movable part of the machine 5, and the detection means 2d and 2f of the accelerometer 5 is detected, the unit system of the equation (7) is different.
The machine model estimation means 11 differentiates or integrates the measured frequency characteristics H _Mb , H _Md , H _Mf to make the same unit system as in the first embodiment, and the moment of inertia J and the friction characteristics as in the first embodiment. D may be estimated, or it may be constructed by changing the unit system of the above equation, and the moment of inertia J and the friction characteristic D may be estimated by performing the calculation as in the first embodiment.

  If the expression (7) is the frequency characteristic of the acceleration response, the expression (17) is obtained. If the frequency response of the position / displacement response is obtained, the expression (18) is obtained.

Here, H _R ^(a) ′: frequency characteristic of acceleration response in consideration of friction, H _R ^(d) ′: frequency response of position / displacement response in consideration of friction.

  Even when the unit systems are different, the moment of inertia J and the friction characteristic D are estimated by the least square method or curve fitting as in the first embodiment.

Since the inertia moment J and the friction characteristic D are estimated from a plurality of frequency characteristics, nine types of inertia moment J and the friction characteristic D are obtained here.
These can be averaged to obtain the final moment of inertia J and friction characteristic D.

  As described above, in step STEP 20 for calculating the mechanical characteristic model from the frequency characteristic, the mechanical model estimation means 11 obtains the rigid body load model 8 and the friction model 9 which are the mechanical characteristic model 7.

  Next, the process proceeds to step STEP 30 for calculating the vibration system model from the frequency characteristics, step STEP 31 for estimating the resonance frequency or anti-resonance frequency, step STEP 32 for estimating attenuation from the resonance frequency or anti-resonance frequency, and step STEP 33 for calculating the vibration mode. To implement.

In step STEP 31 for estimating the resonance frequency or anti-resonance frequency, the nine frequency response characteristics H _Ma, H _Mb, H _Mc , H _Md, H _Me , H _Mf, H _Mg , obtained in step STEP 10 for measuring the frequency characteristics, Similar to the first embodiment, the frequency characteristic peak detecting means 11 can estimate the resonance / anti-resonance frequency using H _{Mh and} H _Mm .
Since various points of the electric motor 1 and the machine 5 are measured using the plurality of detection means 2, it is possible to grasp resonance that cannot be detected in the first embodiment.
As described above, the frequency characteristic peak detecting means 11 can calculate the resonance model 18 and the anti-resonance model 19 that are the vibration system model 17.

In step STEP 32 for estimating attenuation from the resonance frequency or anti-resonance frequency, the attenuation of the resonance frequency or anti-resonance frequency obtained in step STEP 31 for estimating the resonance frequency or anti-resonance frequency is attenuated in the same manner as in the first embodiment. The estimated value analysis means 12 calculates.
In this embodiment, since a plurality of frequency characteristics are measured, the same resonance frequency exists in a plurality of frequency characteristics, and a plurality of attenuations can be obtained. However, individual attenuations may be used, and averaged attenuation may be obtained. You may ask.
As described above, the attenuation estimated value analysis means 12 calculates the attenuation model 20 that is the vibration system model 17.

In step STEP33 for calculating the vibration mode, a plurality of frequency characteristics obtained in step STEP10 for measuring the frequency characteristics are estimated in step STEP31 for estimating the resonance frequency or antiresonance frequency, and the step for estimating attenuation from the resonance frequency or antiresonance frequency. Based on STEP 32, the vibration mode estimation means 13 calculates a vibration mode.
The vibration mode is a behavior in which the machine vibrates. It is obtained from the gain (amplitude) and phase of the frequency characteristic at a certain resonance frequency, and the movement of each point measured by the machine 5 including the electric motor 1 is monitored to determine the mechanical rigidity. Can be judged qualitatively.
Thus, in step STEP33 for calculating the vibration mode, the vibration mode estimation means 13 calculates the vibration mode model 21 that is the vibration system model 17.

  As described above, in this embodiment, in step STEP 10 for measuring frequency characteristics, a plurality of frequency characteristics are measured, and in step STEP 20 for calculating a mechanical characteristic model from the frequency characteristics, the frequency characteristics obtained by detecting the rigid body component are obtained. The mechanical model estimating means 11 can calculate the rigid body load model 8 and the friction model 9 which are the mechanical characteristic model 7, and the process proceeds to step 30 where the vibration system model is calculated from the frequency characteristics, and the resonance frequency or anti-resonance frequency is estimated. In step STEP 31, the frequency characteristic peak detecting means 11 calculates anti-resonance / resonance, and estimates attenuation from the resonance frequency or anti-resonance frequency. In step STEP 32, the attenuation estimated value analyzing means 12 determines the resonance or anti-resonance attenuation. In step STEP33 for calculating the vibration mode, vibration mode estimation is performed. By means 13 calculates the vibration mode, the resonant model 18 is a vibration system model 17, the anti-resonance model 19, decay model 20, it can calculate the vibration mode model 21.

FIG. 10 is an overall configuration diagram of a mechanical characteristic modeling device for an electric motor control device according to a third embodiment of the present invention.
FIG. 11 is a block diagram of an electric motor control apparatus showing a third embodiment of the present invention.
In the figure, 1 is an electric motor, 2 is a detection means, 3 is a controller, 4 is a command means, 5 is a machine, 6 is a frequency characteristic calculation device, 7 is a mechanical characteristic model, 8 is a rigid load model, 9 is a friction model, 10 is a machine model estimation means, 11 is a frequency characteristic peak detection means, 12 is an attenuation estimated value analysis means, 14 is an output means, 15 is an input means, 16 is a storage means, 17 is a vibration system model, 18 is a resonance model, 19 Is an anti-resonance model and 20 is a damping model.

This embodiment is different from the first embodiment and the second embodiment in that it has a two-axis configuration having two electric motors 1 and a controller 3.
Further, the second embodiment is different from the first embodiment in that the motor 1 is not a linear motor but a rotary motor for both axes. The procedure of the mechanical property modeling method is the same as that in FIG. 1 of the first embodiment.
The second embodiment differs from the second embodiment in that the electric motor 1 is a rotary motor, except that the X axis has a semi-closed configuration in which the detecting means 2 detects the operation amount of the electric motor 1b. Step Y is the same in that the Y-axis detection means 2 has a fully closed configuration, but there is no other accelerometer, no vibration mode estimation means 13, and a step STEP33 for calculating a vibration mode for obtaining the vibration system model 17. The difference is that it is not included in the procedure of the mechanical property modeling method.
Further, since the multi-axis configuration is _{adopted, the} frequency characteristic H _{Mb, 1} due to the response of the detection means 2b when the motor 1a is operated, and the frequency characteristic H _Ma due to the response of the detection means 2a and 2c when the motor 1b is operated. _{, 2} and H _{Mc, 2} are different, but the same as the second embodiment is that a plurality of frequency characteristics are obtained.

Since the present invention is different from Patent Document 1 in that the friction model 9 and the multi-axis configuration are used, the frequency characteristic H _{Mb, 1} depending on the response of the detection means 2b when the motor 1a is operated, and the motor 1b. The frequency characteristics H _{Ma, 2} and H _{Mc, 2} are obtained by the response of the detecting means 2a and 2c when the is operated.

The present invention differs from Patent Document 2 in that the frequency characteristic calculation device 6, the machine model estimation means 10, the frequency characteristic peak detection means 11, the estimated attenuation value analysis means 12, the input means 15, the storage means 16, and the vibration system model 17. Since the resonance model 18, the anti-resonance model 19 and the damping model 20 are provided and have a multi-axis configuration, the frequency characteristics H _{Mb, 1 depending on} the response of the detection means 2b when the motor 1a is operated, and the motor 1b are operated. The frequency characteristics H _{Ma, 2} and H _{Mc, 2} depending on the response of the detection means 2a and 2c at that time are obtained.

  In the same manner as in the first embodiment, step STEP 10 for measuring the frequency characteristic, step STEP 20 for calculating the mechanical characteristic model from the frequency characteristic, and step STEP 30 for calculating the vibration system model from the frequency characteristic are performed.

In step STEP 10 for measuring frequency characteristics, frequency characteristics of a multi-axis configuration are obtained.
Basically the same as in the first and second embodiments, but the frequency characteristic H _{Mb, 1} according to the response of the detection means 2b when the motor 1a is operated, and the detection means when the motor 1b is operated There is a feature that frequency characteristics H _{Ma, 2} and H _{Mc, 2} can be obtained by responses of 2a and 2c.

The procedure for measuring the frequency characteristics is as follows.
When the two axes rotate the feedback loop and open the loop, the command means 4 generates a swept sine wave C, which is given to the controller 3a as a torque command, and operates with the machine 5 to which the electric motor 1a is added. The controller 3b is not given a torque command and is left open looped.
The detecting means 2a and 2c detect the operation amounts ra and rc of the operated electric motor 1a and machine 5.
On the other hand, the detection means 2b also obtains a response rb.
Accordingly, the frequency characteristic operation unit 6, ra torque command T1 and the detection means 2 of swept sine C command means 4 detects, rb, from rc, the frequency characteristic _{H M a, H M c,} H M b, Get one.
Next, when the operating shaft is switched, the command means 4 generates a swept sine wave C and gives it to the controller 3b as a torque command, so that it operates together with the machine 5 to which the electric motor 1b is added. The detecting means 2b detects the operation amount rb of the operated electric motor 1b.
Since the motor 1b also operates on the other shaft, responses ra and rc are also obtained by the detection means 2a and 2c.
Accordingly, the frequency characteristic operation unit 6, ra torque command T2 and the detection means 2 of swept sine C command means 4 detects, rb, from rc, the frequency characteristic _{H M a, 2, H M} c, 2, Obtain H _M b.

As described above, at step STEP 10, the frequency characteristic operation unit 6 is six frequency characteristic _{H M a, H M c,} H M b, 1 and _{H M a, 2, H M} c, 2, H M b Can be measured.
In the above, the motors 1a and 1b are alternately operated and one of them is stopped at the time of measurement. However, if the two types of commands C of the command means 4 are uncorrelated with each other, the motors 1a and 1b are simultaneously operated. The frequency characteristics H _Ma , H _M c, H _M b, 1 and H _M a, 2, H _M c, 2, H _M b may be measured.

When the motors 1a and 1b are operated simultaneously, the motor 1a is operated with the random wave C1 of the command means 4, the motor 1b is operated with the M-sequence wave C2 of the command means 4, and C1 and C2 are not correlated with each other. The three detection means 2a, 2b, and 2c obtain responses ra, rb, and rc, and the frequency characteristic calculation device 6 performs frequency analysis using, for example, FFT (Fast Fourier Transform) to obtain a spectrum. The spectrum is calculated as an auto power spectrum and a cross spectrum. Auto power spectrum and cross spectrum are averaged. From these spectra obtained frequency characteristic H _M. Frequency characteristic _{H M} is related simulated manner expressions (19).

The left side of the command signals C1, C2 and square matrix of, by multiplying the inverse matrix, since the matrix of the frequency characteristics can be solved, simulatively H from equation (20) shown _{M a, H M c, H} M b, It can be seen that 1 and H _M a, 2, H _M c, 2, and H _M b are obtained.

Where A ^* ; complex conjugate of A
[A] ^-1 ; Indicates the inverse matrix of [A].

In other words, the auto power spectrum and cross-spectrum obtained by averaging, the matrix of the frequency characteristic H _M can be solved by a formula (21).

here,
G'ra, c1; averaged cross spectrum of response ra and command signal C1
G'rb, c1; averaged cross spectrum of response rb and command signal C1
G'rc, c1; averaged cross spectrum of response rc and command signal C1
G'ra, c2; averaged cross spectrum of response ra and command signal C2
G'rb, c2: averaged cross spectrum of response rb and command signal C1
G'rc, c2: averaged cross spectrum of response rc and command signal C1
G'c1, c1; averaged auto power spectrum of command signal C1
G'c1, c2; averaged cross spectrum of command signal C1 and command signal C2
G'c2, c1; averaged cross spectrum of command signal C2 and command signal C1
G′c2, c2; shows an averaged auto power spectrum of the command signal C2.

Thus, the frequency characteristic calculation device 6 can measure six frequency characteristics by any method.

Step 20 for calculating the mechanical characteristic model from the frequency characteristic can be executed in the same manner as in the first and second embodiments.
The point of using a plurality of frequency characteristics is the same as in the second embodiment, and the mechanical model estimation means 10 calculates the moment of inertia and the friction characteristics, whereby the rigid load model 8 and the friction model 9 which are the mechanical characteristics model 7 are calculated. Is demanded.
Since the present embodiment has a two-axis configuration, two types of rigid load models 8 and friction models 9 are obtained.
On the X-axis side, the moment of inertia and the friction characteristic are calculated using the frequency characteristic _HMb obtained from the response obtained by the detection means 2b. Further, frequency characteristics _{HMb, 1} may be used.
On the Y-axis side, the moment of inertia and the friction characteristic are calculated using the frequency characteristics H _{Ma and} H _Mc obtained from the responses obtained by the detection means 2a and 2c. Further, frequency characteristics H _{Ma, 2 and} H _{Mc, 2} may be used.

  Next, the process proceeds to step STEP 30 for calculating a vibration system model from the frequency characteristics, step 31 for estimating the resonance frequency or anti-resonance frequency, and step 32 for estimating attenuation from the resonance frequency or anti-resonance frequency, as in the first embodiment. To implement.

Step STEP31 for estimating the resonance frequency or the anti-resonance frequency can be executed in the same manner as in the first and second embodiments.
The use of a plurality of frequency characteristics is the same as in the second embodiment, and the frequency characteristic peak detecting means 11 can calculate the resonance model 18 and the anti-resonance model 19 which are the vibration system model 17.

Step 32 for estimating the attenuation from the resonance frequency or the anti-resonance frequency can also be executed in the same manner as in the first and second embodiments.
The use of a plurality of frequency characteristics is the same as in the second embodiment, and the attenuation estimated value analysis means 12 calculates the attenuation model 20 that is the vibration system model 17.

  Thus, in step STEP 30 for calculating the vibration system model from the frequency characteristics, the resonance model 18, the anti-resonance model 19, and the attenuation model 20 that are the vibration system model 17 are calculated as in the first and second embodiments. To do.

  As described above, in this embodiment, in step STEP 10 for measuring frequency characteristics, a plurality of frequency characteristics of a multi-axis configuration are measured, and a rigid body component is detected in step STEP 20 for calculating a mechanical characteristic model from the frequency characteristics. The machine model estimation means 11 can calculate the rigid body load model 8 and the friction model 9 which are the mechanical characteristic model 7 using the frequency characteristics thus obtained, and the process proceeds to STEP 30 where a vibration system model is calculated from the frequency characteristics, and the resonance frequency or anti-resonance is obtained. In step STEP31 for estimating the frequency, the frequency characteristic peak detecting means 11 calculates the antiresonance / resonance, and in step STEP32 for estimating the attenuation from the resonance frequency or the antiresonant frequency, the attenuation estimated value analyzing means 12 The anti-resonance attenuation is calculated, and the resonance model 18 which is the vibration system model 17, the anti-resonance model is calculated. 19, we can calculate the attenuation model 20.

Since the machine model can be estimated by measuring the frequency characteristics, it can be applied to the use of grasping information for optimally operating the electric motor.
In addition, since the machine model can be estimated, it can be applied to a model to be controlled used for simulation of a control system using a numerical model as a preliminary study before actually operating the motor optimally.

The flowchart of the mechanical characteristic modeling method of the motor control apparatus which shows 1st Example of this invention. 1 is an overall configuration diagram of a mechanical characteristic modeling device for an electric motor control device according to a first embodiment of the present invention. 1 is a block diagram of an apparatus for modeling mechanical characteristics of an electric motor control apparatus according to a first embodiment of the present invention. The figure of the frequency response characteristic which has friction which shows 1st Example of this invention The figure which shows the rigid system and 2 inertia system which show 1st Example of this invention The figure which shows the rigid system model which has friction which shows 1st Example of this invention Flowchart of a method for modeling mechanical characteristics of an electric motor control device according to a second embodiment of the present invention Overall configuration diagram of a mechanical characteristic modeling device of an electric motor control device showing a second embodiment of the present invention The block diagram of the mechanical-characteristic modeling apparatus of the motor control apparatus which shows 2nd Example of this invention. Overall configuration diagram of a mechanical characteristic modeling device of an electric motor control device showing a third embodiment of the present invention The block diagram of the motor control apparatus which shows 3rd Example of this invention. Whole block diagram of an electric motor control device provided with a machine model estimation device showing a first conventional example The figure which showed an example unsuitable for the curve fitting of the rigid body model type | formula which shows the conventional 1st example The block diagram which showed the details of the basic idea which shows the conventional 2nd example

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Electric motor 2 Detection means 3 Controller 4 Command means 5 Machine 6 Frequency characteristic calculation apparatus 7 Mechanical characteristic model 8 Rigid body load model 9 Friction model 10 Mechanical model estimation means 11 Frequency characteristic peak detection means 12 Attenuation estimated value analysis means 13 Vibration mode estimation Means 14 Output means 15 Input means 16 Storage means 17 Vibration system model 18 Resonance model 19 Anti-resonance model 20 Damping model 21 Vibration mode model 101 Arithmetic unit
102 Servo control device 103 Rotation detector 104 Electric motor 105 Transmission mechanism 106 Movable portion 107 Non-movable portion 108 Operation command signal 109 Rotation detector signal 110 Control signal 119 Input device 120 Frequency characteristic equation 121 Output device 201 Speed command generator 202 Speed control Unit 203 model speed control unit 204 identification unit 205 adjustment unit

Claims (27)

An electric motor control device comprising command means for creating an operation command for an electric motor that drives the machine drives the machine, and a detection means detects an operation amount of the machine, and the detected operation amount and the operation command In the mechanical characteristic modeling device for modeling the characteristic of the machine, comprising a frequency characteristic calculating device that obtains the frequency characteristic of the machine by calculating
A rigid load model of the load of the motor and the machine;
A mechanical property model including a friction model that quantifies friction;
And a mechanical model estimating unit for calculating the mechanical characteristic model including the rigid body load model and the friction model from the frequency characteristic. 2. The mechanical characteristic modeling device for an electric motor control device according to claim 1, wherein the mechanical model estimation unit directly calculates the rigid load model and the friction model from the frequency characteristic. The mechanical model estimation means curve-fits the frequency characteristic with the mechanical characteristic model including the rigid body load model and the friction model, and calculates the mechanical characteristic model including the rigid body load model and the friction model. The machine characteristic modeling device for an electric motor control device according to claim 1. The said mechanical model estimation means converges the variation | change_quantity of the said mechanical characteristic model containing the said rigid body load model and the said friction model which curve-fitted so that the said frequency characteristic might be approximated by iterative calculation. Machine characteristic modeling device for electric motor control device. 2. The machine characteristic modeling device for an electric motor control device according to claim 1, wherein the detecting means detects the position, speed or acceleration of the electric motor, or the position, speed or acceleration of the machine. The mechanical characteristic modeling device for an electric motor control device according to claim 1, wherein the output of the command means is a swept sine wave. The machine characteristic modeling device for an electric motor control device according to claim 1, wherein the output of the command means is an M-sequence signal. The mechanical characteristic modeling device for an electric motor control device according to claim 1, wherein the output of the command means is a random wave. The electric motor is a rotary motor,
2. The mechanical characteristic modeling device for an electric motor controller according to claim 1, wherein the rigid load model that is a load amount of the electric motor and the machine is an inertia moment. The electric motor is a translational motor,
The mechanical characteristic modeling device for an electric motor control device according to claim 1, wherein the rigid body load model that is a load amount of the electric motor and the machine is a mass. 2. The machine of the motor control device according to claim 1, wherein the machine model estimation means includes frequency characteristic peak detection means for automatically calculating a protrusion shape including a resonance frequency and an anti-resonance frequency of the frequency characteristic. Characteristic modeling device. 12. The machine of the motor control device according to claim 11, wherein the machine model estimation means includes attenuation estimated value analysis means for estimating attenuation from a resonance frequency or anti-resonance frequency detected by the frequency characteristic peak detection means. Characteristic modeling device. The mechanical characteristic model includes a vibration system model including a resonance model and a damping model that are input in advance to the mechanical model estimation unit,
2. The mechanical characteristic modeling apparatus for an electric motor control apparatus according to claim 1, wherein the mechanical model estimation unit calculates a vibration system model including the resonance model and the damping model from the frequency characteristic. From a plurality of the frequency characteristics measured using a plurality of the detection means,
The machine model estimating means, the frequency characteristic peak detecting means, and the attenuation estimated value analyzing means estimate the load model and the friction model, resonance frequency and antiresonance frequency, and attenuation, respectively. The mechanical characteristic modeling device for the motor control device according to claim 1. 15. The mechanical characteristic modeling device for an electric motor control device according to claim 14, wherein the mechanical model estimation unit further includes a vibration mode estimation unit that estimates a vibration mode from the plurality of frequency characteristics. At least two or more of the motor control device, the motor, and the detection means,
The mechanical characteristic modeling device for an electric motor control device according to claim 1, wherein the frequency characteristic calculation device measures at least two of the frequency characteristics. The mechanical characteristic modeling device for an electric motor control device according to claim 1, further comprising output means, input means, or storage means. 2. The mechanical characteristic of the motor control device according to claim 1, wherein the mechanical model estimation unit calculates the mechanical characteristic model by arbitrarily specifying at least two points of the frequency characteristic and limiting a frequency range. Modeling device. A machine of an electric motor control system comprising: a machine; an electric motor that drives the machine; command means that creates a control command for the electric motor; and a controller that receives the control command and drives the electric motor. In the characteristic modeling method,
The command means creates an operation command for the motor and measures the frequency characteristics of the machine driven by the operation command;
A method for modeling a mechanical characteristic of a motor control device, comprising: calculating a mechanical characteristic model from the frequency characteristic to derive a mechanical characteristic model including a rigid body load model and a friction model. The method for modeling a mechanical characteristic of an electric motor control device according to claim 19, wherein the mechanical characteristic model includes a vibration system model including a resonance model and a damping model. The method for modeling a mechanical characteristic of an electric motor control device according to claim 19, further comprising a step of calculating a vibration system model from the frequency characteristic. The step of calculating the vibration system model includes a step of automatically calculating a protrusion shape including a resonance frequency and an antiresonance frequency of the frequency characteristic to detect a peak of the frequency characteristic and estimating a resonance frequency or an antiresonance frequency. The method for modeling a mechanical characteristic of an electric motor control device according to claim 21, comprising: The method for modeling a mechanical characteristic of an electric motor control device according to claim 21, wherein the step of calculating the vibration system model includes a step of estimating attenuation from a resonance frequency or an anti-resonance frequency of the frequency characteristic. The method for modeling a mechanical characteristic of an electric motor control device according to claim 21, wherein the step of calculating the vibration system model includes a step of calculating a vibration mode using a plurality of the frequency characteristics. An electric motor control device comprising the mechanical property modeling device according to claim 1. A machine controlled by the control device of claim 25 and characterized in characteristics. An electric motor control device comprising the mechanical property modeling device according to claim 1;
An electric motor controlled by the electric motor control device;
A machine driven by the motor;
A machine control system comprising: