JPH02307384A - Motor speed controller - Google Patents

Abstract

PURPOSE:To enable a stable speed control with good response even at the time of a low speed by estimating a motor speed in the section, where the output signal of a position transducer does not change, from the operated value of a load torque and the measured value of a motor-generated torque. CONSTITUTION:The output pulse from an encoder 5 is counted by a counter 9, the output of a clock generator 10 is counted by a counter 11, and the enumerated data of said counter 11 is latched in a register 12 for every output pulse of said encoder 5. A microcomputer 1 takes in the enumerated data of said counter 9 and the stored value of said register 12 for every speed control period to perform speed detection processing 101 for every encoder pulse. Said microcomputer 1 performs load torque operation processing 104 from the output of said speed detection processing 101 and from that of mean generated torque operation processing 102 for every encoder pulse. On the other hand, speed estimation processing 105 is performed from the operated value of a load torque and the output of momentary generated torque operation processing 103 for every control period. Thus, it is possible to perform a stable speed control with good response even at the time of a low speed.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to a speed control device for an electric motor using a digital position detector.

[Conventional technology]

In control devices such as robots and numerically controlled machine tools, only a digital position detector is used for the drive motor, and speed control is often performed using this as a minor loop control system of the position control system. In addition, for speed control of electric motors, many methods have been announced that use digital position detectors, which allow easier signal transmission and are less affected by drift than analog types, to detect speed and execute speed control. There is. Typically, an incremental encoder is used as a digital position detector for motor control.

In order to detect the speed from such a digital position detector, there are two methods, such as a method of obtaining the reciprocal of the time interval of the output signal from the position detector, or a method of calculating the number of pulses of the position detection signal at each predetermined period. Are known. In either case, in high-speed conditions, the time interval between the output signals from the position detector is short, and a sufficient number of position detection signals can be obtained at each predetermined cycle, allowing for highly accurate speed detection with little detection delay. value is obtained. But at low speed conditions.

Since the time interval at which the output signal of the position detector changes increases, the speed detection delay increases, and there are also cases where the output signal of the position detector does not change at every predetermined period, making it impossible to detect the speed.

This effect is particularly noticeable when the resolution of the position detector is low.

As a method to solve these problems,
As stated in Publication No. 30984, a model expressing the drive characteristics is constructed, and a period in which the position detection value does not change from the speed detection value and the motor drive torque detection value calculated every time an output signal from the position detector is obtained. A method of estimating the instantaneous speed at and controlling it is known. Here, the speed is estimated by separating the drive torque detection value into an acceleration torque component and a load torque component, dividing the former by the moment of inertia value, and integrating the result. For this reason, it is necessary to estimate the load torque component, but in the conventional technology, the load torque is estimated by proportional integration of the deviation between the speed detection value from the position detector and the speed estimate value. As a result, the load torque can be estimated asymptotically with a response time constant determined by the reciprocal of the proportional-integral compensation bandwidth. Speed detection delay is reduced by estimating the instantaneous speed from this load torque estimate and the drive torque detector.

[Problem to be solved by the invention]

At this time, in the conventional technique described above, in order to estimate the load torque with good response, it is necessary to set a large proportional-integral compensation constant. However, actual drive systems exhibit resonance characteristics, nonlinear friction characteristics, and backlash due to backlash. For this reason, in the method of estimating load torque from the deviation between the estimated speed value and the speed detection BS value calculated from the position detector, the actual drive characteristics and speed The load torque estimation system may become unstable due to differences with the estimation model. For this reason, the proportional-integral compensation constant for estimating the load torque cannot be set high enough, and when speed control is performed using the speed estimate,
There were problems such as increased speed fluctuations due to load disturbances.

An object of the present invention is to achieve control performance with small speed fluctuations even in low speed conditions by stably calculating load torque even in low speed conditions where the time interval between output signals from a 1-position detector is long.

[Means to solve the problem]

In order to achieve the above object, a means is provided to calculate the average acceleration from the time change of the detected speed value calculated for each output signal from the position detector, and the calculated value of this acceleration and the measured value of the torque generated by the motor are provided. By calculating the average load torque from the above, the load torque can be stably determined even when the time interval between the output signals from the position detector becomes long. By estimating the motor speed in a section where the output signal of the position detector does not change from this load torque calculation value and the measured value of the motor generated torque, stable and responsive speed control is achieved even at low speeds.

In addition, in the method of estimating load torque by proportional integration of the deviation between the detected speed value and the estimated speed value, the proportional integral constant for estimating the load torque is By providing means for making the position variable variable according to the time interval of the position detector, the load torque and motor speed can be estimated while maintaining stability even at low speeds.

[Effect]

The average acceleration during this period is calculated from the difference between the previous detection value and the current detection value of the speed detection value, which is calculated every time the output signal of the position detector changes.

The average load torque is calculated from this acceleration and the average value of the generated torque in this section. Due to this.

Every time a position detection signal is obtained, the average load torque in the area up to the position detection signal two times before is determined.

Therefore, it is possible to calculate the load torque more accurately than a method of asymptotically estimating the load torque by proportional integration of the deviation between the detected speed value and the estimated speed value. Using this load torque calculation value,
Since speed estimation is performed in the section where no position detection signal is obtained, a speed control system with small speed fluctuations due to load disturbances can be constructed even at low speeds.

Furthermore, in the method of estimating load torque by proportional integration of the deviation between the detected speed value and the estimated speed value, as the time interval of the position detection signal becomes longer, the proportional integral constant for estimating the load torque is made smaller.

The stability of the loop that asymptotically performs load torque estimation can be restored, and more stable speed control can be performed over extremely low speed conditions.

〔Example〕

An embodiment of the present invention will be described below with reference to FIG.

The main components in Figure 1 are a microcomputer (hereinafter abbreviated as microcomputer) 1 that executes speed control of the electric motor.
．． Power amplifier 2. Electric motor 3. The load machine 4 is an incremental encoder (hereinafter abbreviated as encoder) 5 as a digital position detector for detecting the speed of the electric motor 3.
is used.

The motor current is measured by a current detector 6, fed back to a current control circuit 7, and used for current control of the power amplifier 2. The microcomputer 1 performs speed control processing 10 at each predetermined control cycle.
6 is activated and calculates a current command value for the power amplifier 2. This current command value is converted into an analog quantity by a digital/analog converter (hereinafter abbreviated as a D/A converter) 8, and becomes a command value to the current control circuit 7.

On the other hand, output pulses from the encoder 5 are counted by a counter 9. Further, the output of the clock generator 10 is counted by a counter 11, and the count value of the counter 11 is latched in a register 12 for each output pulse of the encoder 5. The microcomputer 1 takes in the count value of the counter 9 and the value stored in the register 12 at each speed control period, and executes speed detection processing 101 for each encoder pulse. Further, the output of the current detector 6 is converted by a voltage/frequency converter (hereinafter abbreviated as V/F converter) 13 into a pulse train of a frequency corresponding to the motor current, which is counted by a counter 14 to Obtain the cumulative value of current.

Since the motor torque is generated in proportion to the motor current, this current measurement value can be regarded as the calculated value of the motor generated torque.

On the other hand, the cumulative value of the motor current, which is the count value of the counter 14, is latched in the register 15 for each encoder pulse.
It is used to calculate the integral value of the motor generated torque between encoder pulses.

The microcomputer 1 takes in the cumulative value of the motor generated torque latched for each encoder pulse from the register 15 and the cumulative value itself from the counter 14 at every predetermined control cycle, and the former is the average of the motor generated torque between the encoder pulses. The latter is used in the calculation process 102 of the value, and the latter is used in the calculation process 103 of the motor generated torque for each control cycle. The microcomputer 1 executes a load torque calculation process 104 for each encoder pulse based on the output of the speed detection process 101 and the output of the average generated torque calculation process 102. on the other hand,
For each control period, a speed estimation process 105 is executed from this load torque calculation value and the output of the instantaneous generated torque calculation process 103, and this output is used to achieve stable and responsive speed control even at low speeds.

Next, this operating waveform is shown in FIG. Here, (a) is the output pulse of the encoder 5, (b) is its count value, (c
) is the count value of the clock pulse for measuring the time interval of the encoder pulse and the value obtained by latching it for each encoder pulse, (d) is the sampling point of the speed control cycle of the microcomputer (the cycle is Ts), ( e) is the detected value of the motor current, (f) is the cumulative value of the motor current obtained by converting the detected value to V/F and counting the output, and the value obtained by latching it for each encoder pulse. g) shows the instantaneous current measurement value of the motor obtained by subtracting the cumulative value of the motor current for each control period.

The microcomputer 1 controls the counter 9. at each sampling time point shown in FIG. 2(d). Register 12. Counter 14. Retrieve data from register 15. Assume now that an encoder pulse is obtained at time j between sampling times i and i-1. At the main sampling point, [(j-1) ~ (j
)] The average speed (IJII(j)) of the section is calculated as follows.

(1) Here, N(j) corresponds to the pulse time interval from (j-1) to (j), and ΔM(j) represents the change in the encoder pulse in the same interval. In this case, ΔM is always 1.

Moreover, ko is a conversion coefficient for speed detection.

On the other hand, the cumulative value of the motor current determined as the count value of the counter 4 can be calculated from [(j-1) to (j )] The integral value of the motor current in the section is calculated by the following formula.

Δ1. (j)=1. (j) −Ijj−1) ・
...(2) Also, by reading out the count value of the counter 4 corresponding to the cumulative value of the motor current at each sampling time and taking the difference from the previous value, the instantaneous current i between the sampling points, (i) can be calculated using the following formula: It is calculated by

In the embodiment of the present invention, the speed detection value cv for each encoder pulse given by equation (1), (j), and the cumulative value Δ of motor current between encoder pulses given by equation (2)
The instantaneous speed at each sampling point is estimated using I+a(j) and the measured value i, (i) of the instantaneous motor current in equation (3). The purpose of this is to perform stable and responsive speed control even at low speeds.

First, the principle of the present invention will be explained with reference to FIGS.

Now, the model of the electric motor drive system is set up as shown in Figure 3. In other words, the generated torque of the electric motor is τyo, the motor speed is ω,
When closed, 0m(t)=-(τ,(t)-τd(t))
...(4). Here, τd is the load torque component, and J is the inertia moment value of the drive system. Also,·
represents the time derivative. In Fig. 3, S represents the Laplace operator. If this drive system model is used, the load torque component τd can be obtained by modifying equation (4), τ, + (1) = τ++ (
t)-Jω, (t) ... It can be calculated from (5). That is, if the generated torque τ1 and the acceleration ω of the electric motor are determined, and the moment of inertia value J at this time is known, the load torque τd can be calculated from equation (5).

Now, the output signal from the digital position detector is
As shown in the figure, it is assumed that it is obtained at the time J 2tJ 1+J.

Now, assuming that the instantaneous velocities at time j-1 and j are found, by integrating equation (5) over the interval [(j-1) to (j)], c d (j) t P (J) = 〒Jj)jp(j)
-J(ω, (j)-ω, (j-1))...(6) holds true. Here, d(j), T, and (j) are the average values of the load torque and motor generated torque in the interval [(j-1) to (j)], respectively, and tp(j) is the time interval of this interval. , ωyo(j) and ω,(j −1) are respectively j
, represents the instantaneous velocity at time j-1.

However, in speed detection from a digital position detector,
At most, only the average speed for each position detection signal can be determined.

For example, in FIG. 4, at time j, [(j-1)~(
j)] The average speed of the section is calculated. Now, [(j −
2) ~ (j −1)], [: (j −1) ~ (j)]
The average speed in each section is (1111 (J-1)
y CL) - (J), and assuming that the speed changes linearly in each section.

...(7) holds true. Here, ωmc52)P 0m (jl)
rω and (j) represent instantaneous velocities at times j-2, j-1, and j, respectively. From formula (7), [(j-2)~(
j-1)] and [(j-1) to (j)].

...(8) becomes.

On the other hand, when formula (6) is expanded and written in the interval [(j-2) to (j)], v a (j 1 ) t p (j 1 ) + 〒
a(j)tp(j)=rjj-1)tp(j-
1) +tm(j)tp(j)-J(ω1(j)-ω, (
j-2)) ...(9) holds true. Substituting the relationship in equation (8) into the second term on the right side of this equation, the load torque τ is constant in the interval [(j - 2) to (j)] and its value is 7d(
j), then? d(j)(t p(j 1
)+tp(j))=〒jj-1)tp(j-1)
+r-(j)tp(j)-2J(ding,(j)-dj-
1)) ...(10). here.

ta(j 1)tp(j 1)+ya(j)t
p(J) meeting r Jj )(t p(j 1 ) + t
p(j)) (11).

That is, the average load torque rd(j) in the interval [(j-2) to (j)] is determined from equation (10). Both sides (
If you divide by tp(j 1)+tp(j))...(1
2) It becomes. In other words, the first term on the right side of equation (12) is [(j
−2) to (j)] is the average value of the torque generated by the motor in the section, and the second term corresponds to the product of the average acceleration in this section multiplied by the moment of inertia value J.

According to equation (12), each time the output signal of the position detector changes,
The average load torque in the section up to the position detection signal two times before can be calculated. From this calculated value ta(j) of the load torque and the motor generated torque (t).

The time change of the estimated value ω,(t) of the motor speed at time t is given by the following equation by substituting Ca(j) into equation (4).

Here, 5Jt) is the estimated speed at time t,
The inclusion m(t) represents its time derivative. Now, if the estimated speed at point j is ω, (j), then the time 1 after point j
The estimated speed value ω at +, (11) is calculated by integrating equation (13) from time j to time tt, as shown in (14). Here, at time j, [(j−
1) ~ (j)] Average value of estimated speed values in the interval S, (j)
and the speed detection value obtained from the output signal of the position detector -jj)
If the estimated speed value is corrected so that it matches, the estimated speed value at time t1 ? , (1,) is.

It is calculated by the following formula.

10'; m (j) + δj (ding, (j)'λ (j))...
(I5) Here, δ1 is a function that is 1 only at time j and zero in other sections, and represents that the estimated speed value is corrected every time a speed detection value from the position detector is obtained. Here, Syo(j) is the average value of the estimated speed values in the interval [(j −1) to (j), and is determined from the following equation.

Note that tp(j) represents the time period of [(j-1) to (j)]. That is, the average value t F (j) of the load torque is calculated for each output signal of the position detector using equation (12), and using this, the estimated speed value ω, (1+) is calculated using equation (15).
), the speed detection delay can be shortened even at low speeds.

In the embodiment shown in FIGS. 1 and 2, the average velocity τ, (
j) and the pulse-to-pulse current integral value ΔI(j) are respectively (1
), is calculated by equation (2), and by equation (3), i
The current measurement value i, (i) at the time point is determined.

The average load torque r-(j) is given by the formula (1z),
...(17) Determined from. Here, kl is a conversion coefficient from a current cumulative value to a 1-lux integral value, and T'ci+ is a period of a clock pulse for measuring time between encoder pulses.

From this, t-(j) is calculated as follows.

(Dingyo(j)-Ding,(j-1)) ...(18) Also, the estimated speed value ωJi+ at the sampling time point i+1
1) converts equation (14) into a discrete time system and obtains Ts ω−(1+1)=(k 1jyo(i) −
x a (j))+';, (i)...(19) Calculated as follows. Here, i y (i) is the difference between the cumulative current values at each sampling time point, and is calculated from equation (3). Ts is the sampling period.

Furthermore, as shown in Fig. 2, when the speed detection value from the encoder pulse is obtained at sampling points i-1 and i, the speed estimate ω, (i) is Ts ; m(i) = ( ktim(i-1) −td(j
−1))+ω, (i −1)+(c, 几(j)′;m(
j))...(20) Calculated as follows. A block diagram at this time is shown in FIG.

As described above, according to this embodiment, the speed can be detected for each encoder pulse by reading out the encoder pulse measurement value and the time measurement value for each encoder pulse at each predetermined sampling period. Accurate speed detection is possible without shortening the sampling period. Also,
Another advantage is that the section average value of the motor-generated torque necessary for estimating the load torque can be easily calculated by integrating the detected torque values at each sampling time and dividing by the number of samplings in that section.

A second embodiment of the invention is shown in FIG. In this example,
At the sampling time i at the time j when the encoder pulse is detected, the average value of the velocity estimates between the encoder pulses S, and the velocity detection value from the encoder pulse, d
Gain K is the deviation from that. , and add this to the load torque calculation value to perform speed estimation. This method is shown in FIG. At this time, at the sampling time j when the encoder pulse is detected, the estimated speed value ω,(i) is calculated by the following equation.

Here, the average value of estimated speed values (j) is calculated as follows.

Note that at the sampling time point (i+1) when no encoder pulse is detected, the estimated speed value is calculated using the following equation as in the first embodiment.

(23) According to the present invention, it is possible to asymptotically match the average value of estimated speed values between encoder pulses, ω, (j), with the detected speed value ω, (j). This closed loop sign Ko
is given by, according to Figure 6. From this, the bandwidth of speed estimation can be limited to this closed loop gain Ko (rad/s).

Therefore, it is possible to obtain a speed estimate ω, (i) from which noise in a high frequency range of Ko (rad/s) or higher is removed, and there is an advantage that a more stable speed estimate can be obtained.

In addition, as a third embodiment of the present invention, since the average speed factor, (, j) and the average acceleration are calculated at time j when the encoder pulse is obtained, the instantaneous velocity at time j is cvjj)
It can be obtained by extrapolation from . Now, from the second term on the right side of equation (8), the average acceleration a, (j) in the interval [(j-2) to (j)] can be regarded as...(25). On the other hand, if the speed changes linearly in the [(j-1) to (j)] interval (time interval tp(j)), us(j) corresponds to the speed at the midpoint of that interval. do. Therefore, the extrapolated value ω2(j) of the instantaneous velocity at time j can be calculated as follows. If the estimated speed value is preset to ω-(j) at time j, the estimated speed value thereafter is...(27
) can be estimated. Here, ω-(j) is the extrapolated velocity value obtained from equation (26), and 〒1(j) in equation (26) is the average acceleration calculated from equation (25).

According to this embodiment, the estimated speed value ω,(i) can be corrected to a value closer to the instantaneous speed, so that the estimated speed value can be obtained in which the phase delay of the detected speed value from the encoder pulse is further improved.

As described above, by calculating the average acceleration from the position detector output signal and estimating the load torque component from this, a stable speed estimate can be obtained even when the interval between the position detector output signals becomes long. I talked about it. On the other hand, even in the case of the conventional speed estimation method, when the interval between the output signals of the position detector becomes long, the gain of load torque estimation is reduced to improve the speed control characteristics when the speed estimation value is fed back and controlled. can be stabilized. Such load torque estimation by the state observer is calculated as follows. Similar to the method described above, the drive system characteristics are modeled as shown in equation (1).

Similarly, assuming that the position detection signal is obtained at time j as shown in FIG. 2, the estimated speed value ωm(t
+) is...(28). At this time,? 'd(j) is the estimated load torque value obtained at time j, and τd(j) is 7d(j)=Kt
Σ(Ding, (j)-7,(j )) + Kp(Ding, (j
) −';, (j ))...(29) Calculated as follows. Here, - (j) is the average speed of the section [(j-1) to (j)] detected at time j, as described above, and (j) is the estimated speed value in the same section. ω, (
t). The first term on the right side of equation (29) is −[(
j-1) to (j)] is the sum of the errors between the detected speed value and the estimated speed value for each speed detection, and corresponds to an integral term. Similarly, the second term corresponds to the proportional term. Furthermore, Kr and Kp correspond to the integral gain and proportional gain of load torque estimation, respectively. According to state observer theory, Kz.

The load torque τ-(j) can be estimated using the characteristic determined by Kp, and using this, the speed S, (t,) in the section where the position detection signal does not change can be estimated as shown in equation (28).

Here, the load torque estimated value (j) is estimated asymptotically every time the output of the position detector changes, so if the interval between the position detection signals becomes long, it becomes impossible to stably estimate the load torque. Therefore, a fourth embodiment of the present invention will be described. That is, by decreasing the estimated gains Kl and Kp in equation (29) as the interval between position detection signals increases, more stable load torque estimation can be performed. Using this load torque estimated value, the speed estimated value for the section where no position detection signal is obtained is obtained by equation (28).

Also, the load torque estimation in the state observer is (2
In Equation 9), it can be executed using only the proportional term.

At this time, the estimated load torque value is ? a(j) = Kp(Ding,(j)-Goyo(j))
...(30) can be executed. In this case, similarly, when the interval between position detection signals becomes longer, the estimated gain K
P cannot be made large. Therefore, by similarly making the estimation gain variable according to the interval of the position detection signals, more stable load torque estimation can be performed.

On the other hand, in load torque estimation using a state observer, (
29) Among the estimated load torque values shown in equation 29), the integral term in the first term on the right side is the load torque amount, and the proportional term in the second term is the load torque for correction to eliminate the discrepancy between the estimated speed value and the detected value. Corresponds to the ingredients. Therefore, in the section where the position detection signal is not obtained, speed estimation is performed using only the estimated value of the load torque amount in the first term on the right side of equation (29), and the position detection signal and speed detection value are obtained. If the speed is estimated by adding the load torque component for speed estimation value correction in the second term on the right side of equation (29) at this point, more stable speed estimation becomes possible. That is, (
29) The first and second terms of equation ↑ar(jL↑6
Let it be p(j). Namely.

shall be. At this time, at time j when the position detection signal is obtained.

...(32) Speed estimation is performed as shown below. Here, Swr (j−
δ) is the velocity estimate before time j. On the other hand, speed estimation in the section where no position detection signal is obtained after time j is as follows.

...(33) Execute from. That is, in this section, speed estimation is performed using only the estimated value τdt(j) of the load torque amount.

Also, using the speed detection value from the position detection signal, (25)
As shown in the formula, by calculating the acceleration by taking into account the time interval between the sections where each speed detection value was obtained, it is possible to more stably calculate the average acceleration even if the average speed between the position detection signals is used. Can calculate. By feeding back this acceleration to the speed control system, speed control performance can be improved.

Further, it is calculated by equation (12) for each output signal of the position detector. The load torque calculation value is stably calculated even at low speeds. By using this as the load torque estimated value and performing feedforward compensation to the speed control system, it is possible to configure a speed control system with less fluctuation due to load disturbance.

Furthermore, for load torque estimation and speed estimation, when based on the shame system model shown in FIG. 1, it is necessary to detect the motor generated torque τ. Normally, in the case of an electric motor, the generated torque τ,
Since τ is considered to be proportional to the motor current, the generated torque τ can be detected by detecting the motor current. Further, normally, the relationship between the torque command value, which is the output of the speed control calculation, and the torque generated by the motor can be approximated by a first-order lag characteristic or the like. Therefore, even if the motor generated torque is not detected, the generated torque can be calculated using the torque command value and the first-order lag characteristic, and this can be used to perform load torque and speed estimation.

In addition, as a method of obtaining the speed detection value from the output signal of the position detector, we have described the case where it is obtained by the reciprocal of the time interval of the position detection signal, but in a state where the position detection signal can be obtained at a sufficiently high speed, the Speed can be calculated as the number of signals. Load torque and speed estimation can be similarly executed at predetermined intervals using this speed detection value.

Next, a fifth embodiment of the present invention will be described with reference to FIG.

This block diagram shows the load torque by the state observer? This is a method for estimating d, and ? 6 is calculated using equation (29). Here, two load torque estimation gains KI and Kp are provided, and the gain when the speed is detected in a low speed state where speed cannot be detected every sampling of speed control is set to K.
r and Kp, and in a high-speed state where the speed is detected at every sampling, gains of K h and K p h are used.

here.

It is. In this way, by switching the load torque estimation method according to the speed state, stable speed estimation can be performed over a wide range of speed states.

As described above, according to this embodiment, there are two tables for the estimated gain, and the estimated gain is made variable by switching the tables depending on whether speed can be detected at each sampling, without increasing the burden on the software. , the estimated gain can be switched.

〔Effect of the invention〕

According to the present invention, the average acceleration is calculated in a responsive manner based on the difference in the output signal of the speed detection value from the output of the position detector, and the load torque component is estimated thereby, so that the load torque is stable even at low speeds. Can be estimated. Thus, by estimating the speed, control with less speed fluctuation can be realized even in response to load disturbances at low speeds.

[Brief explanation of the drawing]

Fig. 1 is a block diagram showing the configuration of an embodiment of the present invention, Fig. 2 is a timing diagram of speed estimation in this embodiment, and Fig. 3 is an immediate-acting system model for explaining the present invention in detail. Block diagram, FIG. 4 is a timing diagram of the position detection signal, FIG. 5 is a block diagram of speed estimation processing in the first embodiment, and FIG. 6 is a block diagram of speed estimation processing in the second embodiment. , FIG. 7 shows a block diagram of speed estimation of the third embodiment. 1...Microcomputer, 3...Electric motor, 5.
... Incremental encoder, 6... Current detector, Fig. 2 (421 (j-11 (] ) Fig. 3 rd Fig. 4 Time

Claims (1)

[Claims] 1. Comprising an electric motor, a digital position detector mechanically coupled to the electric motor, and a torque measuring means for detecting or estimating the torque generated by the electric motor, and generating a speed command signal and a speed feedback signal. In a speed control device for an electric motor that performs speed control calculations, a speed detection value is calculated from the output signal every time an output signal of the digital position detector is generated, and the previous value and current value of the speed detection value are calculated. , an acceleration calculation means for calculating the average acceleration during this period, and an average load torque from the average value of the torque generated by the electric motor in the generation area of the output signal of the digital position detector and the acceleration calculation value. A speed calculation comprising: estimating means, estimating a motor speed from the estimated value of the average load torque and the measured value of the torque measuring means, and performing a speed control calculation using the estimated speed value as a speed feedback signal. A speed control device for an electric motor, characterized in that a speed control device for an electric motor is provided. 2. The speed control device for an electric motor according to claim 1, wherein the detected torque value of the electric motor is estimated from a torque command value determined by speed control calculation. 3. The electric motor according to claim 1, wherein when the speed of the electric motor is sufficiently high, the detected speed value is determined by subtracting the count value of the output signal of the position detector every predetermined period. speed control device. 4. In claim 1, when the output of the position detector changes, a detected speed value calculated from the change in the output of the position detector and an estimated speed value in a section where the output of the position detector does not change. The speed control of an electric motor is characterized in that the difference from the average value of Device. 5. In claim 1, when the output of the position detector changes, a detected speed value calculated from the change in the output of the position detector and an estimated speed value in a section where the output of the position detector does not change. The motor speed is estimated by calculating the difference between the average value of Electric motor speed control device. 6. In claim 1, when the position detector output changes, the average acceleration calculated from the difference from the previous value of the speed detection value and the speed detection value at the time the position sensor output changes. A speed control device for an electric motor, characterized in that the instantaneous speed at the time point is extrapolated, and the estimated speed value at the time point is matched with the speed extrapolated value. 7. It is equipped with an electric motor, a digital position detector mechanically coupled to the electric motor, and a torque detection means for detecting the torque generated by the electric motor, and performs speed control calculations from a speed command signal and a speed feedback signal. The motor speed is estimated from the load torque estimate obtained at the time of the previous change in the position detection value and the motor torque detection value, and this is used to execute speed control. The load torque is estimated by the proportional integral of the deviation between the detected speed value for each change and the average value of the estimated speed value in an interval where the output of the position detector does not change. A speed control device for an electric motor, characterized in that a proportional integral gain of load torque estimation is reduced when the load torque is long. 8. In claim 7, the load torque component is determined by a proportional term of the deviation between the speed detected value for each change of the position detector and the average value of the estimated speed in a section where the output of the position detector does not change. A speed control device for an electric motor, characterized in that the proportional gain of load torque estimation is reduced when the time interval between output changes of the position detector is long. 9. In claim 7, each time the output signal of the position detector changes, the detected speed value at that time and the average value of the estimated speed values in a section where the output of the position detector does not change. The load torque is estimated by the sum of the proportional term and the integral term of the deviation, and in the section where the position detector does not change, speed estimation is performed using only the integral term of the load torque estimated value as the load torque estimated value. A speed control device for an electric motor, characterized by: 10. Claim 1, characterized in that, in addition to the speed feedback signal, an average detected acceleration value calculated at each time point of change of the output signal of the position detector is fed back for speed control calculation. A speed control device for electric motors. 11. The speed control device for an electric motor according to claim 1, characterized in that an average load torque estimate calculated every time the output signal of the position detector changes is fed back to the speed control calculation section. 12. A control device for an industrial robot, characterized in that the speed control device for an electric motor according to claim 1 is used to configure a servo system for positioning each axis of the industrial robot.