JP2005285030A - Servo control apparatus - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a servo control apparatus that achieves high responsiveness for controlled objects requiring high responsiveness, keeps vibration from occurring even with controlled objects of low rigidity, and reduces the electric power consumed by a motor for controlling objects requiring no high responsiveness. <P>SOLUTION: The servo control apparatus includes a predictive controller 81 or 82 and an on/off switch 70 and outputs a control input u(i) and a signal u'(i) alternately as a control input, the signal u'(i) being a deviation e(i-K) multiplied by a proportional gain K<SB>p</SB>. <P>COPYRIGHT: (C)2006,JPO&NCIPI

Description

  The present invention relates to a servo control device that drives a machine tool, a semiconductor manufacturing apparatus, a mounting machine, and the like that require a high command tracking system.

As a first conventional technique, an example of a servo control device that provides a control input to a control target so that the output of the control target matches a target value using predictive control control has been proposed (see, for example, Patent Document 1). ).
FIG. 11 is a block diagram showing the configuration of a prediction controller used in the servo control apparatus of the first prior art. In FIG. 11, the prediction controller 90 includes memories 62 to 65, an arithmetic unit 66, a subtracter 67, an accumulator 68, and a difference unit 69. The predictive controller 90 then sets the target command increment value of the M sampling future at the current time i · Ts (hereinafter, referred to as time i for convenience, Ts: sampling period) so that the output of the controlled object matches the target command. Δr (i + M) and K (K ≧ 0) sampling Incremental output value of the control target in the past (hereinafter referred to as output increment value) y (i−K) is input, and control input u (i) is the control target. To give.
The memory 62 stores target command increment values for a plurality of samplings. Memory 63 stores constants _v m for control, E, and _{p n,} and _{g n.} The memory 64 stores output increment values for a plurality of samplings. The memory 65 stores control inputs for a plurality of samplings.
The subtractor 67 obtains the difference between the target command increment value Δr (i−K) in the past of K sampling and the output increment value Δy (i−K) obtained from the difference unit 69. The accumulator 68 obtains the deviation e (i−K) by accumulating the output of the subtractor 67.
The computing unit 66 sets the control input u (i) so that the future deviation predicted value obtained using the transfer function model from the control input to the output to be controlled, the deviation, and the evaluation function related to the control input are minimized. ,

Asking.
FIG. 12 is a block diagram showing the configuration of the second conventional servo control apparatus to which the prediction controller 90 is applied. In FIG. 12, the control input from the prediction controller 90 is given to the motor controller 3 together with the output of the controlled object. Then, the motor 2 is driven by the control of the motor controller 3, and the control target output output from the motor 2 is sent to the prediction controller 90.
In this way, since the control input is determined so that the future deviation prediction value is minimized, a servo control device with good tracking accuracy is realized.

In addition, as a second prior art, a feedforward signal creation command filter that does not deteriorate the prediction accuracy even when feedforward control is performed, and a compensation signal calculator that further improves the tracking accuracy when the speed or acceleration changes are provided. A servo control device with high tracking accuracy has been proposed (see, for example, Patent Document 2).
FIG. 13 is a block diagram showing an example of the configuration of a compensation signal calculator used in a conventional servo control device. In FIG. 13, the compensation signal calculator 101 includes an inverse transfer function unit 11, a multiplier 12, and a subtracter 13. The inverse transfer function unit 11 has an inverse transfer function characteristic of an arbitrary low-pass filter. Then, the compensation signal calculator 101 inputs the command increment value so that the deviation between the target command at the time of acceleration / deceleration and the output of the controlled object becomes small, and outputs a compensation signal. In the compensation signal computing unit 101, the command increment value is input to the inverse transfer function unit 11, the output and the target command increment value are subtracted by the subtractor 13, the output is multiplied by the gain K2 by the multiplier 12, and the output To the control object 1 as a compensation signal.
FIG. 14 is a block diagram illustrating a configuration of a servo control device including a compensation signal calculator 101 and a prediction controller. The target command increment value is given to the prediction controller 90 and the compensation signal calculator 101. Further, the compensation calculation signal output from the compensation signal calculator 101 and the control input from the prediction controller 90 are given to the motor controller 3 together with the output to be controlled. Then, the motor 2 is driven by the control of the motor controller 3, and the control target output output from the motor 2 is sent to the prediction controller 90.

According to this configuration, since the control input is determined so that the predicted deviation value in the future is minimized, a servo control device with a good tracking system is realized, and the change in speed and acceleration is large, so the controller alone follows. Even when the compensation is not possible, the compensation signal computing unit compensates with the feed-forward compensation signal for the controlled object, so that the controlled object 1 has high follow-up accuracy without causing overshoot or continuous vibration with respect to a command whose speed and acceleration vary greatly. Can be controlled.
Japanese Patent No. 3175877 JP 2002-62906 A

However, in both the first prior art and the second prior art, when predictive control is used, there is a problem that vibration is likely to occur in a control object with low rigidity because the response is fast and the followability is high. Further, when predictive control is used, there is a problem in that the power consumption of the motor increases because high followability is pursued.
The present invention has been made in view of such problems, and realizes high responsiveness to a controlled object requiring high responsiveness, and vibration is generated even for a controlled object having low rigidity. An object of the present invention is to provide a servo control device that suppresses power consumption of a motor for a control object that is difficult and does not require high responsiveness.

In order to solve the above problems, the present invention is configured as follows.
The invention according to claim 1 is a servo control device that outputs a control input to a control target so that the output of the control target matches a target command,
A target command increment value that is an increment between sampling cycles of the target command is input, and a control input is sent to the control target so that the target command that is an integrated value of the target command increment value matches the output of the control target. With a predictive controller,
The prediction controller, at a current time i, a target command increment value Δr (i + M), which is an increment between sampling periods of a target command of M sampling future,
K (K ≧ 0) is a predictive controller that inputs an output increment value Δy (i−K) of a control target before sampling and outputs the control input to the control target;
A changeover switch (70) for switching the control input;
Means for storing a target command increment value, a predictive control constant, a past output increment value, and a past control input;
Means for obtaining a deviation e (i−K) from the target command increment value and the output increment value to be controlled;
Evaluation function, formula (a)

However, W _m, the coefficient applied to the α deviation, c and c _d is a coefficient applied to the control input and its increment, e * ⁽ⁱ + m) is the dynamic characteristic model of the controlled object, the future target command, control inputs and the controlled object Means for determining the control input u (i) so as to minimize the estimated deviation value of m sampling future determined from the output;
When the changeover switch (70) is ON, the control input u (i) that minimizes the evaluation function and equation (a) is output as a control input. When the changeover switch (70) is OFF, the deviation e it is characterized in further comprising a means for outputting (i-K) to the signal u multiplied by a proportional gain K _p ': (i) as a control input.

The invention according to claim 2 is a servo control device that outputs a control input and a feedforward signal to a control target so that the output of the control target matches a target command.
Using a target command signal that is information on the target command in the future as an input, a future command increment value that is an increment between sampling periods of the target command signal from the current sampling time to a plurality of sampling futures, and the feedforward signal A feedforward signal creation command filter to be created;
A prediction controller that inputs the future command increment value and sends a control input to the control target so as to match the target command that is an integrated value of the target command increment value and the output of the control target;
A changeover switch (21) for enabling the feedforward signal,
The feedforward signal generation command filter receives the target command signal at the current sampling time, and calculates an increment between the sampling periods of the target command signal or a signal obtained by filtering the target command signal in the future command increment value. Output as
The changeover switch (21) is a switch that outputs the feedforward signal only when ON,
The predictive controller includes the feedforward signal;
At the current time i, a target command increment value Δr (i + M) that is an increment between sampling periods of the target command of M sampling future,
K (K ≧ 0) is a predictive controller that inputs an output increment value Δy (i−K) of a control target before sampling and outputs a control input to the control target, and a changeover switch (70) for switching the control input When,
Means for storing a target command increment value, a predictive control constant, a past output increment value, and a past control input; and a means for obtaining a deviation e (i−K) from the target command increment value and the output increment value of the controlled object. ,
Evaluation function, formula (a)

However, W _m, the coefficient applied to the α deviation, c and c _d is a coefficient applied to the control input and its increment, e * ⁽ⁱ + m) is the dynamic characteristic model of the controlled object, the future target command, control inputs and the controlled object Means for determining the control input u (i) so as to minimize the estimated deviation value of m sampling future determined from the output;
When the changeover switch (70) is ON, the evaluation function, the control input u (i) that minimizes the expression (a) is output as a control input, and when the changeover switch (70) is OFF, the deviation is characterized in further comprising a means for outputting e a (i-K) signal multiplied by a proportional gain K _p to u '(i) as a control input.

The invention according to claim 3 is a servo control device that outputs a control input and a compensation calculation signal to the control target so that the output of the control target matches the target command.
A target command increment value that is an increment between sampling cycles of the target command is input, and a control input is sent to the control target so that the target command that is an integrated value of the target command increment value matches the output of the control target. A predictive controller;
A compensation signal calculator for generating a compensation signal for compensating so that a deviation between the target command at the time of acceleration / deceleration and the output of the control target is reduced by using the target command increment value as input, and sending the compensation signal to the control target;
A changeover switch (22) for enabling the output of the compensation signal calculator,
The prediction controller, at a current time i, a target command increment value Δr (i + M), which is an increment between sampling periods of a target command of M sampling future,
K (K ≧ 0) is a predictive controller that inputs an output increment value Δy (i−K) of a control target before sampling and outputs a control input to the control target.
A changeover switch (70) for switching the control input;
Means for storing a target command increment value, a predictive control constant, a past output increment value, a past control input;
Means for obtaining a deviation e (i−K) from the target command increment value and the output increment value to be controlled;
Evaluation function, formula (a)

However, W _m, the coefficient applied to the α deviation, c and c _d is a coefficient applied to the control input and its increment, e * ⁽ⁱ + m) is the dynamic characteristic model of the controlled object, the future target command, control inputs and the controlled object Means for determining the control input u (i) so as to minimize the estimated deviation value of m sampling future determined from the output;
When the changeover switch (70) is ON, the evaluation function, the control input u (i) that minimizes the expression (a) is output as a control input, and when the changeover switch (70) is OFF, the deviation is characterized in further comprising a means for outputting e a (i-K) signal multiplied by a proportional gain K _p to u '(i) as a control input.

  According to a fourth aspect of the present invention, the compensation signal calculator receives the target command increment value, and includes an inverse transfer function unit having an inverse transfer function characteristic of an arbitrary low-pass filter, and the inverse transfer function unit. A subtracter that subtracts the target command increment value from an output, and a multiplier that multiplies the output of the subtractor.

  According to a fifth aspect of the present invention, the switch (70) for switching the control input, the switch (21) for enabling the feedforward signal, and the switch for enabling the compensation calculation signal. The switch (22) is characterized in that it is turned on when the control object is required to have high-accuracy followability.

  The invention according to claim 6 is the switch (70) for switching the control input, the switch (21) for enabling the feedforward signal, and the switch for enabling the compensation calculation signal. The switch (22) is characterized in that it is turned off when the target command increment value is changing.

In the invention according to claim 7, the predictive controller includes a filter function unit instead of the changeover switch (70).
The filter function unit includes the changeover switch (21) for enabling the feedforward signal or the changeover switch (22) for enabling the compensation calculation signal.
When switching from the OFF state to the ON state, at the current time i, the initial value is u ′ (i), the final value after T sampling is u (i + T), and the continuous control input u (i + t) ”(0 ≦ t ≦ T),
When switching from the ON state to the OFF state, at the current time i, the initial value is u (i), the final value after T sampling is u '(i + T), and the continuous control input u (i + t) "' (0≤ (t ≦ T) is output.

  In the invention according to claim 8, the filter function unit receives the control input u (i), inputs a weight function unit that outputs a weight, and the control input u ′ (i). And a weight function unit that outputs a weight and an adder that adds the outputs of the weight function unit.

A control switch (70) for switching the control input, a switch (21) for enabling the feedforward signal, or the switch (22) for enabling the compensation calculation signal, so that the control target has high-accuracy followability. When requested, the changeover switch (70), the changeover switch (21), and the changeover switch (22) are turned on to use the predictive control,
When the target command increment value is changing, the changeover switch (70), the changeover switch (21), and the changeover switch (22) are turned off and proportional control is used.
High responsiveness is achieved for controlled objects that require high responsiveness, and vibration is unlikely to occur even for controlled objects with low rigidity. A servo control device with reduced power consumption can be provided.

Hereinafter, embodiments of the present invention will be described with reference to the drawings.
Although various functions and means are built in the actual servo control device, only the functions and means related to the present invention will be described and described in the figure. Hereinafter, the same reference numerals are assigned to the same names as much as possible, and the duplicate description is omitted.

FIG. 1 is a block diagram showing a configuration of a servo control apparatus according to a first embodiment of the present invention. In FIG. 1, 1 is a control object, 2 is a motor, 3 is a motor controller, and 81 is a prediction controller.
The control target 1 includes a motor 2 and a motor controller 3.
The prediction controller 81 inputs the target command increment value and the motor position output signal, and outputs a control input to the motor controller 3. The motor controller 3 inputs the control input and the motor position output signal, and outputs a motor driving current to the motor 2. The motor 2 is driven by the motor driving current and outputs the motor position output signal to the motor controller 3 and the prediction controller 81.
The prediction controller 81 performs a prediction control calculation so that the motor position output signal matches the target command increment value, and calculates the control input. The motor controller 3 performs a control calculation so that the motor position output signal matches the control input, and calculates the motor drive current.

FIG. 2 shows a predictive controller of the servo controller according to the first embodiment of the present invention. In FIG. 2, 62 is a memory, 63 is a memory, 64 is a memory, 65 is a memory, 66 is an arithmetic unit, 67 is a subtractor, 68 is an accumulator, 69 is a difference unit, and 70 is a changeover switch.
The memory 62 receives the target command increment value Δr (i + K), stores the target command increment value Δr (i + M),..., Δr (i−K), and stores the target command increment value Δr (i−M). Output to the subtractor 67. Memory 63 stores constants _{v m, E, p n,} g n, the _{K p.} The memory 64 receives the actual measurement value Δy (i−K) of the output increment value and stores the actual measurement value Δy (i−K),..., Δy (i−K−Na + 1) of the output increment value. The memory 65 receives the control input u (i) and stores the control inputs u (i−K),..., U (i−K−Nb + 1).
The differencer 69 receives the motor position output signal y (i−K), calculates an actually measured value Δy (i−K) of the output increment value, and outputs it to the subtractor 67 and the memory 64.
The subtractor 67 inputs the target command increment value Δr (i−M) and the actually measured value Δy (i−K) of the output increment value, and outputs the output increment value from the target command increment value Δr (i−M). The deviation obtained by subtracting the actual measurement value Δy (i−K) is output to the integrator 68. The integrator 68 integrates the deviation and outputs an integrated value e (i−K) of the deviation to the calculator 66.
The computing unit 66 calculates the control input u (i) based on the values stored in the memory 62, the memory 63, the memory 64, and the memory 65 and the integrated value e (i−K) of the deviation, and at time i. calculates a target command increment value Δr (i-K) and the output increment measured value Δy (i-K) and the deviation e (i-K) to the proportional gain _{K p} multiplied by the control input u '(i) And output to the changeover switch 70.
The control input u (i) may be equivalent to a known input described in Patent Documents 1 and 2 and the like.

  The changeover switch 70 receives the control input u (i) and the control input u ′ (i), and turns ON when the control target 1 is required to have a high-accuracy follow-up property. 1), and when the target command increment value is changing, the control input u ′ (i) is output (to the controlled object 1). The signal for turning ON / OFF the changeover switch 70 may be given from a host controller (not shown in the figure), or not shown in the figure, but with a digital operator equipped in the servo control device. It may be input from a so-called operating device, and is not limited by the present invention.

In the example of FIG. 2, the transfer function model of the controlled object 1 is
_{^{Gp (z) = (b 1}} z -1 + ··· + b Nb z -Nb) / {(1-z -1) (1-a 1 z -1 - ··· -a Na z -Na)}
Is obtained in the discrete time system, the output incremental value model is expressed by Equation (2).

Here, Δ indicates an increment value between sampling periods.
At time i, an actual measurement value Δy (i−n) (n ≧ K) of the output increment value up to time i−K is obtained. Therefore, the subsequent output increment value is calculated using the actual measurement value.

The predicted output increment value Δy ^* (i + m) is given by equation (4).

Here, the coefficients A _mn and B _mn are as follows: A _{(−K + 1) n} = a _{(n−K + 1)} m = −K + 1, K ≦ n ≦ when the future control input is u (j) = 0 (j> i) Na + K-1 (5a)

B _{(−K + 1) n} = b _{(n−K + 1)} m = −K + 1, 0 ≦ n ≦ Nb + K−1 (6a)

Given in. However, it is assumed that a _n = 0 (n> Na) and b _n = 0 (n <1 and n> Nb).

If u (j) = u (i) (j> i), B _m0 in equation (6b) becomes equation (6b ′).

Therefore, the future deviation e ^* (i + m) is

Evaluation function, formula (8)

When the control input u (i) is determined so as to minimize, Equation (1) is obtained from ∂J / ∂u (i) = 0. However, _constants, v _m, p n, E, is _{g n} given by Equation (9).

  Here, if K = 0, the evaluation function, equation (8) is

Equation (10) is obtained.
The control input u (i) that minimizes the evaluation function, equation (10), is

It is obtained with.
Therefore, when the changeover switch is ON, u (i) is output as a control input.
The control input from the prediction controller 81 is given to the motor controller 3 together with the output of the control object 1. Then, the motor 2 is driven by the control of the motor controller 3.

Is the prior art differs from the present embodiment, the predictive controller 81 provided with a selector switch 70, and stored in the memory 63 constant _v m, _E, p n, a proportional gain _{K p} in addition to _{g n,} arithmetic the 'calculates a (i) and control input u (i), the control input u to select the changeover switch 70' proportional gain K _p obtained by multiplying the control input u to the deviation e (i-K) in vessel 66 ( i) or the control input u (i) is output. Thus, according to the first embodiment, the same servo control device can be used regardless of whether or not high followability is necessary and regardless of the rigidity of the control target.

FIG. 3 is a block diagram showing the configuration of the servo control apparatus according to the second embodiment of the present invention. In FIG. 3, 82 is a prediction controller, and 83 is a feedforward signal creation command filter.
The feedforward signal generation command filter 83 inputs a future target command at the current sampling time i, and M (M is a natural number) sampling increment value Δr (i),..., Δr (i + M) until the sampling future, The feedforward signals T _FF (i) and V _FF (i) are calculated, and the command increment values Δr (i),..., Δr (i + M) are sent to the prediction controller 82, and the feedforward signal T _FF ( i) and V _FF (i) are output to the prediction controller 82 and the motor controller 3. Here, a variable with Δ indicates an increment value during one sampling period.

FIG. 4 shows a feedforward signal creation command filter 83 of the servo control device according to the second embodiment of the present invention. In FIG. 4, 21 is a changeover switch, 51 is a filter, 52 is a memory, and 53 is an FF signal calculator.
The feedforward signal creation command filter 83 includes a filter 51, a memory 52, an FF signal calculator 53, and a changeover switch 21.
The filter 51 receives the future target command, and increments the sampling period in the M sampling future of the input signal or a signal obtained by filtering the input signal with an internal digital filter (not shown in the figure). The value is output to the memory 52 as a command increment value Δr (i + M).
The digital filter may be any filter that can be used for the purpose of filtering the target command, for example, an IIR filter having an infinite impulse response or a FIR filter having a finite length. In addition, a low-pass filter, a notch filter, or a signal that suppresses vibration of the control target output in consideration of the dynamic characteristics of the control target may be generated.
The memory 52 receives the command increment value Δr (i + M), sequentially stores the command increment value, and command increment values Δr (i), Δr (i + 1),... From the current time i to M sampling future. , Δr (i + M) are output to the FF signal calculator 53, and command increment values Δr (i + M),..., Δr (i + 1), Δr (i) are output (to the prediction controller 82).
The FF signal calculator 53 receives the command increment values Δr (i), Δr (i + 1),..., Δr (i + M), and based on the command increment values, the feedforward signal T _FF (i) and V _FF (i) is calculated and output to the changeover switch 21. Changeover switch 21, the feed enter the forward signal _T FF (i) and _V FF (i), when the changeover switch 21 is ON outputs the feed-forward signal _T FF (i) and _V FF (i) When the changeover switch 21 is OFF, 0 is output. An arithmetic expression for calculating the feedforward signal when the changeover switch 21 is ON is not particularly limited. For example, when a disturbance applied to the control target 1 is known, the calculation may be canceled by calculation.

V _FF (i) = Gain1 · Δr (i + m1)
T _FF (i) = Gain2 · {Δr (i + m2) −Δr (i + m2-1)}
It is good.
Here, Gain1 and Gain2 are constants, Δr (i + m1) is a command increment value in the future of m1 sampling, and m1 and m2 are integers of 0 ≦ m1 ≦ m2.

FIG. 5 shows a predictive controller of the servo controller according to the second embodiment of the present invention. In FIG. 5, 60 is a memory, 61 is a memory, and 82 is a prediction controller.
The prediction controller 82 includes memories 60 to 65, a calculator 66, a subtractor 67, an integrator 68, and a changeover switch 70.
The memory 60 receives the feed forward signal V _FF (i), and the past feed forward signals V _FF (i−K−Nd + 1), V _FF (i−K−Nd + 2),..., V _FF (i). Remember.
The memory 61 receives the feed forward signal T _FF (i) and stores the past feed forward signals T _FF (i−K−Nc + 1), T _FF (i−K−Nc + 2),..., T _FF (i). Remember. In the first embodiment, the memory 62 receives the target command increment value Δr (i + K) and stores the target command increment values Δr (i + M),..., Δr (i−K). Future command increment values Δr (i), Δr (i + 1),..., Δr (i + M) are input, and command value increment values Δr (i), Δr (i−1),. -K) is stored. In the first embodiment, the memory 63 has constants v _m (m = −K + 1, −K + 2,..., M), p _n (n = 0, 1,..., Na−1), E, g _n ( n = 0, 1,..., Nb + K−1) and K _p are stored, but in this embodiment, constants x _n (n = 0, 1,..., Nd + K−1), t _n (n = 0, 1, ..., Nc + K-1).
The arithmetic unit 66 stores the values stored in the memories 60 to 65, the integrated value e (i−K) of deviation, and the command increment value Δr (i) until the future of M (M is a natural number) sampling,. Δr (i + M) is inputted, the value stored in the memories 60 to 65, the integrated value e (i−K) of deviation, and the command increment value Δr (i) up to the sampling future of M (M is a natural number), ..., control input u (i) and control input u '(i) are calculated based on Δr (i + M).
The control input u (i) calculates a future deviation prediction value using a feedforward signal and a transfer function model from the control input to the output, and the evaluation function relating to the future deviation prediction value and the control input u (i) is minimized. And is calculated by the equation (103). The control input u ′ (i) is proportional to the deviation e (i−K) between the target command increment value Δr (i−K) and the actually measured output increment value Δy (i−K) at time i. calculated by multiplying the K _p.
As in the first embodiment, the changeover switch 70 receives the control input u (i) and the control input u ′ (i), and is turned on when the control object 1 is required to have high-accuracy followability. ) (To the control target 1), and when the target command increment value is changing, the control input u ′ (i) is output (to the control target 1). Further, the signal for turning ON / OFF the changeover switch 70 may be given from a host controller (not shown in the figure) as in the first embodiment, or not shown in the figure, but to the servo controller. It may be input from an operating device called a digital operator equipped, and is not limited by the present invention.
In addition, the changeover switch 21 is turned ON when the changeover switch 70 is turned ON, and turned OFF simultaneously with the changeover switch being turned OFF.

Equation (103) is derived.
Two feedforward signals V _FF (i), T _FF (i) to be controlled, and a discrete time transfer function model from the control input u (i) to the output y (i)
Y (z) = {(b ₁ z ^-1 + ... + b _Nb z ^-Nb ) U (z) + (d ₁ z ^-1 + ... + d _Nd z ^-Nd ) V _FF (z)
+ (c ₁ z ⁻¹ + ・ ・ ・ + c _Nc z ^−Nc ) T _FF (z)} / {(1−z ⁻¹ ) (1−a ₁ z ⁻¹ − ・ ・ ・ −a _Na z ^{− Na} )} (128)
Is obtained at time i, the output increment value after time (i−K) is

Predicted by the output increment value Δy ^* (i + M) is

It becomes.
Here, future control inputs and feedforward signals are expressed as u (j) = u (i), V _FF (j) = V _FF (i), T _FF (j) = T _FF (i) (j = i + 1, i + 2) ,..., A _mn , B _mn , D _mn , C _mn are
A _{(−K + 1) n} = a _{(n−K + 1)} m = −K + 1, K ≦ n ≦ Na + K−1 (131a)

B _{(-K + 1) n} = b _{(n-K + 1)} m = -K + 1, 0≤n≤Nb + K-1 (132a)

D _{(-K + 1) n} = d _{(n-K + 1)} m = -K + 1, 0≤n≤Nd + K-1 (133a)

C _{(-K + 1) n} = c _{(n-K + 1)} m = -K + 1, 0≤n≤Nc + K-1 (134a)

Given in. However, a _n = 0 (n> Na), b _n = 0 (n <1 and n> Nb), d _n = 0 (n <1 and n> Nd), c _n = 0 (n <1 and n > Nc).
Therefore, the future deviation predicted value e ^* (i + m) is

Evaluation function, formula (136)

If the control input u (i) is determined so that is minimized, Equation (103) can be obtained from ∂J / ∂u (i) = 0. However, each constant vm, E, pn, gn, xn, tn is

It is.
The control input from the prediction controller 82 and the feedforward signals T _FF (i) and V _FF (i) of the feedforward signal generation command filter are given to the motor controller 3 together with the output of the control target 1. Then, the motor 2 is driven by the control of the motor controller 3.

  The second embodiment is different from the first embodiment in that a feedforward signal creation command filter is provided, and the feedforward signal creation command filter outputs an FF signal to the prediction controller. Thus, the second embodiment can obtain higher followability than the first embodiment.

FIG. 6 is a block diagram showing the configuration of the servo control apparatus according to the third embodiment of the present invention. In FIG. 6, reference numeral 10 denotes a compensation signal calculator.
The compensation signal calculator 10 inputs a target command increment value, which is an input signal of the prediction controller 81, simultaneously with the prediction controller 81, and outputs a compensation calculation signal to the motor controller 3.
The servo control device includes a prediction controller 81 and a compensation signal calculator 10.
The prediction controller 81 is the same as the prediction controller 81 (FIG. 2) shown in the first embodiment.

FIG. 7 shows a compensation signal calculator 10 of the servo control device according to the third embodiment of the present invention. In FIG. 7, 11 is an inverse transfer function unit, 12 is a multiplier, 13 is a subtractor, and 22 is a changeover switch.
The compensation signal calculator 10 includes an inverse transfer function unit 11, a multiplier 12, a subtracter 13, and a changeover switch 22.
The reverse transfer function unit 11 receives the target command increment value and outputs a signal S1 to the subtractor 13. The subtractor 13 inputs the signal S1 and the target command increment value, and outputs a deviation ε obtained by subtracting the target command increment value from the signal S1 to the multiplier 12. The multiplier 12 inputs the deviation epsilon, by multiplying the gain K2, to calculate the signal epsilon _k, and outputs it to the selector switch 22. Changeover switch 22 inputs the signals epsilon _k, when the changeover switch 22 is ON, outputs the signal epsilon _k, 0 when the changeover switch 22 is OFF, the motor controller 3 as a compensation operation signal To do.
The changeover switch 22 is turned ON when the changeover switch 70 is turned ON, and turned OFF simultaneously with the changeover switch 70 being turned OFF.
The inverse transfer function unit 11 is an arithmetic unit having an inverse transfer function characteristic of an arbitrary low-pass filter set in advance, and calculates the signal S1 based on the target command increment value.

  The control input from the prediction controller 81 and the compensation calculation signal output from the compensation signal calculator 10 are given to the motor controller 3 together with the output of the control target 1. Then, the motor 2 is driven by the control of the motor controller 3.

  The third embodiment differs from the first and second embodiments in that a compensation signal calculator 10 is provided and a compensation calculation signal is output to the motor controller 3. Thus, the third embodiment can obtain higher followability than the first embodiment or the second embodiment.

FIG. 8 is a block diagram showing the configuration of the prediction controller 81a of the servo control apparatus according to the fourth embodiment of the present invention. In FIG. 8, 71 is a filter function unit, and 81a is a prediction controller.
The prediction controller 81a is configured such that the changeover switch 70 of the prediction controller 81 (FIG. 2) of the first embodiment is a filter function unit 71.
FIG. 9 is a block diagram showing the configuration of the prediction controller 82a of the servo control apparatus according to the fourth embodiment of the present invention. In FIG. 9, the prediction controller 82 a is obtained by replacing the changeover switch 70 of the prediction controller 82 (FIG. 5) of the third embodiment with a filter function unit 71.
FIG. 10 shows a filter function unit of the servo control apparatus according to the fourth embodiment of the present invention. In FIG. 10, 72 is a first weight function unit, 73 is a second weight function unit, and 74 is an adder.
The filter function unit 71 includes a first weight function unit 72, a second weight function unit 73, and an adder 74.

  When switching from predictive control to normal proportional control, if control input u (i) is input, j is the time to switch from predictive control to normal proportional control, i is the current time, and T is a preset time constant, The output of the weight function unit 72 is

It becomes.
When the control input u ′ (i) is input, j is the time when the predictive control is switched to the normal proportional control, i is the current time, and T is a preset time constant, the output of the second weight function unit 73 is

It becomes.
That is, the filter function unit 71
When (i−j) <0, u (i),
When 0 ≦ (i−j) ≦ T,
u ″ ′ (i−j) = {1− (i−j) / T} · u (i) + {(i−j) / T} · u ′ (i),
When (i−j)> T, u ′ (i),
Is output as a control input (to control object 1).

  In addition, when switching from normal proportional control to predictive control, when a control input u (i) is input, j is the time to switch from normal proportional control to predictive control, i is the current time, and T is a preset time constant. The output of the first weight function unit 72 is

It becomes.
When the control input u ′ (i) is input, the time when the normal proportional control is switched to the predictive control is j, the current time is i, and the preset time constant is T, the output of the second weight function unit 73 is

It becomes.
The outputs of the first weight function 72 and the second weight function 73 are added by an adder 74 and output as a control input.
That is, the filter function 71 is
When (i−j) <0, u ′ (i),
When 0 ≦ (i−j) ≦ T,
u ″ (i−j) = {(i−j) / T} · u (i) + {1− (i−j) / T} · u ′ (i),
When (i−j)> T, u (i),
Is output as a control input (to control object 1).

  The fourth embodiment is different from the first to third embodiments in that the control input u (i) and the control input u ′ (i) are not switched by the changeover switch, and the control input u (i) and the control input u are switched by the filter function unit. '(I) is a point to be synthesized. Thereby, in Example 4, control input does not become discontinuous, and proportional control and predictive control can be switched even when the motor 2 is rotating.

  INDUSTRIAL APPLICABILITY The present invention can be used for a positioning control servo control device for a semiconductor manufacturing apparatus and a servo control device for driving a machine tool or an industrial robot.

The block diagram of the servo control apparatus of 1st Example of this invention Prediction controller of servo controller of first embodiment of the present invention The block diagram of the servo control apparatus of 2nd Example of this invention Feedforward signal generation filter of servo control apparatus of second embodiment of the present invention Prediction controller of servo controller of second embodiment of the present invention The block diagram of the servo control apparatus of 3rd Example of this invention Compensation arithmetic unit of servo control apparatus of third embodiment of the present invention Prediction controller 81a of servo controller of fourth embodiment of the present invention Prediction controller 82a of servo controller of fourth embodiment of the present invention Filter function unit of servo control apparatus of fourth embodiment of the present invention 1 is a block diagram of a prediction controller of a servo control apparatus according to a first prior art. 1 is a block diagram of a servo control apparatus to which a first conventional prediction controller is applied. Block diagram of compensation signal calculator of servo controller of second prior art Block diagram of second prior art servo controller

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Control object 2 Motor 3 Motor controller 10 Compensation signal calculator 11 Inverse transfer function unit 12 Multiplier 13, 67 Subtractor 21, 22, 70 Changeover switch 51 Filter 52, 60, 61, 62, 63, 64, 65 Memory 53 FF signal computing unit 66 computing unit 68 accumulator 69 difference unit 71 filter function unit 72 first weight function unit 73 second weight function unit 74 adder 81, 81a, 82, 82a prediction controller 83 feedforward signal creation filter 90 Conventional prediction controller 101 Conventional compensation signal calculator

Claims (8)

A servo control device that outputs a control input to a control target so that the output of the control target matches a target command,
A target command increment value that is an increment between sampling cycles of the target command is input, and a control input is sent to the control target so that the target command that is an integrated value of the target command increment value matches the output of the control target. With a predictive controller,
The prediction controller, at a current time i, a target command increment value Δr (i + M), which is an increment between sampling periods of a target command of M sampling future,
K (K ≧ 0) is a predictive controller that inputs an output increment value Δy (i−K) of a control target before sampling and outputs the control input to the control target;
A changeover switch (70) for switching the control input;
Means for storing a target command increment value, a predictive control constant, a past output increment value, and a past control input;
Means for obtaining a deviation e (i−K) from the target command increment value and the output increment value to be controlled;
Evaluation function, formula (a)

However, W _m, the coefficient applied to the α deviation, c and c _d is a coefficient applied to the control input and its increment, e * ⁽ⁱ + m) is the dynamic characteristic model of the controlled object, the future target command, control inputs and the controlled object Means for determining the control input u (i) so as to minimize the estimated deviation value of m sampling future determined from the output;
When the changeover switch (70) is ON, the evaluation function, the control input u (i) that minimizes the expression (a) is output as a control input, and when the changeover switch (70) is OFF, the deviation servo controller, characterized in that it comprises a means for outputting e a (i-K) to the signal u multiplied by a proportional gain K _p '(i) as a control input. A servo control device that outputs a control input and a feedforward signal to a control target so that the output of the control target matches a target command,
Using a target command signal that is information on the target command in the future as an input, a future command increment value that is an increment between sampling periods of the target command signal from the current sampling time to a plurality of sampling futures, and the feedforward signal A feedforward signal creation command filter to be created;
A prediction controller that inputs the future command increment value and sends a control input to the control target so as to match the target command that is an integrated value of the target command increment value and the output of the control target;
A changeover switch (21) for enabling the feedforward signal,
The feedforward signal generation command filter receives the target command signal at the current sampling time, and calculates an increment between the sampling periods of the target command signal or a signal obtained by filtering the target command signal in the future command increment value. Output as
The changeover switch (21) is a switch that outputs the feedforward signal only when ON,
The predictive controller includes the feedforward signal;
At the current time i, a target command increment value Δr (i + M) that is an increment between sampling periods of the target command of M sampling future,
K (K ≧ 0) is a predictive controller that inputs an output increment value Δy (i−K) of a control target before sampling and outputs a control input to the control target, and a changeover switch (70) for switching the control input When,
Means for storing a target command increment value, a predictive control constant, a past output increment value, and a past control input; and a means for obtaining a deviation e (i−K) from the target command increment value and the output increment value of the controlled object. ,
Evaluation function, formula (a)

However, W _m, the coefficient applied to the α deviation, c and c _d is a coefficient applied to the control input and its increment, e * ⁽ⁱ + m) is the dynamic characteristic model of the controlled object, the future target command, control inputs and the controlled object Means for determining the control input u (i) so as to minimize the estimated deviation value of m sampling future determined from the output;
When the changeover switch (70) is ON, the evaluation function, the control input u (i) that minimizes the expression (a) is output as a control input, and when the changeover switch (70) is OFF, the deviation servo controller, characterized in that it comprises a means for outputting e a (i-K) to the signal u multiplied by a proportional gain K _p '(i) as a control input. A servo control device that outputs a control input and a compensation calculation signal to the control target so that the output of the control target matches the target command,
A target command increment value that is an increment between sampling cycles of the target command is input, and a control input is sent to the control target so that the target command that is an integrated value of the target command increment value matches the output of the control target. A predictive controller;
A compensation signal calculator for generating a compensation signal for compensating so that a deviation between the target command at the time of acceleration / deceleration and the output of the control target is reduced by using the target command increment value as input, and sending the compensation signal to the control target;
A changeover switch (22) for enabling the output of the compensation signal calculator,
The prediction controller, at a current time i, a target command increment value Δr (i + M), which is an increment between sampling periods of a target command of M sampling future,
K (K ≧ 0) is a predictive controller that inputs an output increment value Δy (i−K) of a control target before sampling and outputs a control input to the control target.
A changeover switch (70) for switching the control input;
Means for storing a target command increment value, a predictive control constant, a past output increment value, a past control input;
Means for obtaining a deviation e (i−K) from the target command increment value and the output increment value to be controlled;
Evaluation function, formula (a)

However, W _m, the coefficient applied to the α deviation, c and c _d is a coefficient applied to the control input and its increment, e * ⁽ⁱ + m) is the dynamic characteristic model of the controlled object, the future target command, control inputs and the controlled object Means for determining the control input u (i) so as to minimize the estimated deviation value of m sampling future determined from the output;
When the changeover switch (70) is ON, the evaluation function, the control input u (i) that minimizes the expression (a) is output as a control input, and when the changeover switch (70) is OFF, the deviation servo controller, characterized in that it comprises a means for outputting e a (i-K) to the signal u multiplied by a proportional gain K _p '(i) as a control input.   The compensation signal calculator inputs the target command increment value, and an inverse transfer function device having an inverse transfer function characteristic of an arbitrary low-pass filter, and a subtraction that subtracts the target command increment value from the output of the inverse transfer function device The servo control device according to claim 3, further comprising a multiplier and a multiplier that multiplies the output of the subtracter.   The changeover switch (70) for switching the control input, the changeover switch (21) for enabling the feedforward signal, and the changeover switch (22) for enabling the compensation calculation signal are: 4. The servo control device according to claim 1, wherein the servo control device is turned on when high-accuracy followability is required.   The changeover switch (70) for switching the control input, the changeover switch (21) for enabling the feedforward signal, and the changeover switch (22) for enabling the compensation calculation signal are the target command increment. 4. The servo control device according to claim 1, wherein the servo control device is turned off when the value changes. The prediction controller includes a filter function unit instead of the changeover switch (70),
The filter function unit includes the changeover switch (21) for enabling the feedforward signal or the changeover switch (22) for enabling the compensation calculation signal.
When switching from the OFF state to the ON state, at the current time i, the initial value is u ′ (i), the final value after T sampling is u (i + T), and the continuous control input u (i + t) ”(0 ≦ t ≦ T),
When switching from the ON state to the OFF state, at the current time i, the initial value is u (i), the final value after T sampling is u '(i + T), and the continuous control input u (i + t) "' (0≤ 4. The servo control device according to claim 1, wherein t ≦ T) is output. The filter function unit inputs the control input u (i), outputs a weight function unit with a weight, and the weight function unit inputs the control input u ′ (i), and outputs the unit with a weight. And an adder for adding the outputs of the weight function unit.