JPH0773433B2 - Speed control device - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION [Industrial field of application] The present invention implements robust equivalent disturbance compensation in a speed control system for applying a speed deviation to a control target through a stabilizing device and effectively utilizing its function. The present invention relates to a speed control device for preventing a so-called windup phenomenon.

[Conventional technology]

An example of a general electric motor speed control system will be described with reference to FIG.

FIG. 3 shows a conventional example of a stabilizing feedback control device, in which 1 is a stabilizing device and 2 is a controlled object. here,
R and Y are a set input and a state quantity, respectively. Therefore, in the speed control system example, R is a speed command and Y is a speed detection output. K _T
Is the torque generation coefficient. Further, J and D are equivalent to the inertia and viscous resistance of the controlled object, and S is a differential operator.

In this way, the deviation e between the setting input R and the state quantity Y is stabilized by the stabilizing device 1
The speed control system is stabilized by applying it to the controlled object 2 through. The stabilizing device 1 is generally configured as a P (proportional) I (integrating) control device, and the integral element included as a stabilizing compensating element is an element necessary for zeroing the offset in the steady state.

[Problems to be Solved by the Invention]

By the way, in practical use, there are Leonard and VV
There is an output limit in power actuators such as VF. Especially, in control, current and torque are constantly monitored, and the limiter is used when exceeding a predetermined value. Therefore, while the operating condition is limited by this limiter, the input deviation is integrated in the integral element, and if the deviation goes out of the limiter and the deviation becomes normal, the unnecessary integrated amount is subtracted, and the speed exceeds To shoot. In some cases, an unstable phenomenon may occur. Such a phenomenon is called a windup phenomenon.

Although various measures have been proposed to remove this, those that can be easily applied have not been put to practical use. For example, the document "Journal of the Institute of Electrical Engineers of Japan, D, '88 -July, P678"
The technique described has some effects, but it requires various complicated processes.

In addition, if the integral calculation is stopped while the limiter is limited, an unnecessary amount of integration that exceeds the limiter value will not occur and the wind-up phenomenon can be reduced, but the period before returning to normal operation Indicates that the P1 output of the stabilizer 1 is from a limiter value (eg 150%) to a normal value (eg 75
%), And overshoot occurs during this time.

[Means and Actions for Solving the Problems]

In view of the above points, the present invention has been made based on the basic technical idea of utilizing the equivalent disturbance compensation function that can minimize the influence of parameter fluctuations and load fluctuations of the controlled object. By inputting the difference between the input quantity of and the state quantity multiplied by the inverse function of the mathematical model between the states from the input as an equivalent disturbance to the input of the controlled object, any disturbance applied to the controlled object is collected. It is configured to be able to compensate. Furthermore, by utilizing such an equivalent disturbance compensation function, the stabilizer used for the main loop of speed control is composed of only proportional elements without including integral elements, and the offset is minimized irrespective of steady and transient times. It is possible to eliminate the occurrence of the windup phenomenon based on the integral element.

Hereinafter, the present invention will be described in detail with reference to the drawings.

Here, the present applicant has previously proposed Japanese Patent Application No. 1-159863 “multifunctional controller”.

In this proposal, the equivalent disturbance compensator is described as an example of the speed control system, and the main points are shown here.

That is, FIG. 4 is shown for explanation of the equivalent disturbance compensator according to the present invention, and 3 is the equivalent disturbance compensator. Further, FIG. 5 shows the control target portion after the equivalent disturbance compensation.

In FIG. 4, the equivalent disturbance is calculated by utilizing the information of the designated T ^* (S) [torque command in the case of speed control] and the speed detection output Y, and this is added to the command T ^* (S). It is composed of Where T _L (S) is the load disturbance and ω (S)
Is the rotation speed.

Then, the torque disturbance coefficient K _T , the controlled object 2 and the load disturbance T are compensated by the illustrated equivalent disturbance compensator 3.
_As shown in FIG. 5, all the parameters of _L (S) can be expressed by nominal values, and the load disturbance T _L (S) can be eliminated.
However, Is.

Here, the equivalent disturbance compensation performance during this period will be further described based on the previous patent application "multifunctional control device".
It looks like this:

First, in FIG. 4, when the equivalent disturbance compensator 3 is not added, it is shown as in the equation (01).

T ^* (S) ・_KT - _TL (S) = (JS + D) ω (S)
(01) On the other hand, considering the parameter variation, the following is set. However, Δ indicates a variation.

Then, by substituting equation (02) into equation (01), the equivalent disturbance T _E
(S) becomes formula (03).

The physical content of this equation (03) is that the equivalent disturbance T _E (S) includes all the load disturbance T _L (S) and fluctuations from the nominal value of each constant. By considering that, it is shown that it can be described only by the nominal value of each constant, as in the right side of Expression (03). This equivalent disturbance T
Equivalent disturbance compensation is performed by adding _E (S) to the command through a low-pass filter for noise removal.

Further, since the time constant of the low-pass filter can be made sufficiently smaller than the time constant of the controlled object, FIG.
It can be rewritten as shown.

Therefore, as shown in FIG. 5, it can be described only by the nominal value that is unrelated to the load disturbance and each constant fluctuation, and the rotation output response to the torque command is not affected by the load disturbance and each constant fluctuation, so that the robust and stable operation can be achieved. Can be secured.

In FIG. 4, the primary type is used as the low-pass filter, but the primary type is not always necessary, and the secondary and tertiary types can also exhibit the same performance and may be selected according to the control system.

Further, the following equation (1) is obtained from FIG.

Therefore, it is clear that the load fluctuation and the parameter fluctuation are all canceled in the controlled object.

Next, by utilizing this equivalent disturbance compensation function, offset does not occur even if the stabilizing device of the main loop of speed control is only proportional control, that is, the proportional control output is always operated near zero (that is, equivalent disturbance compensation). The fast integration function is included in the functions).

Now, when the block of the torque generation coefficient K _T is extracted from the command T ^* (S) shown in FIG. 4, it is as shown in FIG. Here, T ^* (S) 'is a torque command after the equivalent disturbance compensation, and C ₀ is an output after the equivalent disturbance compensation. From Figure 6,
The transfer function from the command T ^* (S) to the command T ^* (S) is as follows.

T ^* (S) + _C0 = T ^* (S) '(3) This is the result of equation (3) and is the natural form.

Therefore, the so-called PI compensation element is inserted,
Here, T can usually be selected in the order of 0.5 to 10 (ms). This means that it involves very fast integration.

Here, in FIG. The following is a description of the section. That is, if it is left as it is, it remains as an offset at the addition point of T ^* (S), but if it is considered that there is no viscosity coefficient d of the controlled object, the D component generated in the actual machine should be included in a part of the equivalent disturbance. Because you can It can be melted. Therefore, Is Becomes zero in steady state, and the above equation (4)
It will be controlled only by. This can be further illustrated as shown in FIGS. 7, 8, 10, and 11.

That is, the response when acceleration / deceleration is performed by the torque limiter is shown. FIGS. 7 and 10 are data when the speed control system is PI controlled, and the overshoot at the end of acceleration / deceleration is large.

Further, FIGS. 8 and 11 show data when P control is performed, and the overshoot thereof is greatly reduced.
Here, Figures 7 and 8 are simulations, Figure 10 and
FIG. 11 shows the corresponding actual operation data.

Further, in FIGS. 8 and 11 of the P control example, the speed fluctuation is small even if the load is changed, and the fluctuation becomes zero in the steady state because the above-mentioned high-speed integration function is operating. .

In addition, the document "Journal of the Institute of Electrical Engineers of Japan, D, Vol. 107, No. 1, (Sho 6
2) (Mr. Ohnishi) ”, P84 in Fig. 2 and P85 in Fig. 4 shows a PI (or P) controller in the speed control system. A speed control system is inserted in the minor loop.
Therefore, even if the speed control system is P-controlled and there is no problem if the main loop of the position control is integrated and compensated even if an offset remains, if the speed control system is the main loop, the conventional P As described above, the offset remains unusable in the control.

Further, claim 2 will be described.

As shown in FIG. 8, even if the speed control system is set to P control, a slight overshoot occurs at the end of acceleration. The cause is that the equivalent disturbance output mentioned above includes a filter, and the integral operation is performed even though this is said to be high speed. To eliminate this point, the filter output is reset to a predetermined value at the end of acceleration / deceleration. This will be described in detail with reference to FIG. 12 by taking an example of the actual situation.

FIG. 12 shows the torque and the actual speed when the speed command is changed stepwise and accelerated.

The In FIG. 12, T _B is the torque of the base load at the actual speed N ₁ state at time t ₁ earlier, T _L is the time t ₂ the actual speed at the time point N ₂
This is the torque newly generated as a load when it rises to. Further, T _A1 is a torque contributing to acceleration at time t ₁ , and T _A2 is a torque contributing to acceleration at time t ₂ . Then, at the end of acceleration, the filter output is reset to the value of (T _B + T _L = T _F2 ).

However, there is a point that the torques T _A2 and T _L cannot be measured as they are. The estimation of the torque T _L is obtained by using a so-called Ruenberger or Gopinus observer, but it is not robust.

Therefore, at time t ₁ , T _A1 = T ₁ (150% of the limiter in Fig. 12) -T _B and acceleration ₁ = dN ₁ (t) / dt were measured, and the acceleration at time t ₂ was measured. _{2 2} = dN ₂ (t) / dt can be measured, and the torque T _A2 can be obtained from the equation (5).

That is, the torque T that contributes to the acceleration just before the end of acceleration
_A2 is the torque T _A1 (= T ₁ that contributes to the acceleration immediately after the start of acceleration.
-T _B ) can be measured and this value can be used to obtain the ratio of acceleration ₁ and ₂ immediately after the start of acceleration and immediately before the end of acceleration.

For this purpose, a condition that there is no change in parameters such as inertia in a short time range before and after acceleration is required, but this condition does not pose a problem in an actual system.

Therefore, time t ₂ [(N ^* _2- N) at the end of acceleration
The value of the filter output T _{F2 in the above} can be obtained by the generated torque immediately before time t ₂ , T _{2 in} FIG.
The value obtained by subtracting T _A2 obtained by the formula (5) from the limiter value (150%) of You can reset it to. FIG. 9 shows the data subjected to this treatment by simulation. That is, it can be understood that the overshoot at the end of acceleration is zero.

Then, such a method is more robust than the load torque observer described above. That is, an observer based on the Luenberger method calculates using parameters such as inertia of the rotating body and torque generation coefficient, but the parameter constants used are designed values. In many cases, the design value and the actual value are different, and the calculation result becomes inaccurate. For example, in the winder, the inertia fluctuates by one digit or more, and in the vector control inverter, the generated torque changes depending on the motor temperature.

In this respect, the present invention does not use such a parameter, but can be accurately obtained if the parameter in a short time range such as before and after gear shift is stable, and the difference between the design value and the actual value of the parameter constant can be obtained. It is something that is not bound to.

It should be noted that although the explanation is given for the case of acceleration in FIG. 12, it is needless to say that the same can be said for the case of deceleration. Further, the generated torque during acceleration / deceleration is applied to the limiter, but it is needless to say that this can be similarly explained in the case of the limiter or less.

Furthermore, although we have described the overshoot suppression of the high-speed integration operation of the equivalent disturbance compensator, even if this method is applied to the overshoot suppression by the integration operation of the conventional PI speed control system, the exact same function is obtained. Can be demonstrated.

Furthermore, claim 3 will be described.

Here, the second aspect of the present invention is intended to suppress the actual speed overshoot when the speed command changes, but is intended to suppress the actual speed overshoot when the load suddenly changes. This is the first
It will be described in detail with reference to FIG.

FIG. 13 shows the operation of each part when the load amount is increased stepwise during constant speed operation.

That is, when the load amount increases rapidly, the actual speed decreases and at the same time the equivalent disturbance output increases. As described above, a filter is present in the equivalent disturbance output, and as it is, a velocity overshoot such as a hatched display occurs. The reason for this is that the integration function is achieved even though it is called high speed.

So point A in the figure In regard to point (2), it is focused on that the equivalent disturbance output at that time matches the required load amount, and the output after the equivalent disturbance filter is stored in memory. Further, when the actual speed matches the speed command (point B), the actual speed is reset to the stored value so that neither overshoot nor undershoot occurs in the actual speed.

It should be noted that the way of understanding the equivalent disturbance is affected by the filter constant, so it may be stored with a slight delay from the point A.
This method is also similar to that described in claim 2.
It is obvious that it is useful for overshoot of general integration operation.

In addition, as described in the second and third aspects of the claims, the device according to the present invention always has a command, an actual output, a state quantity such as torque, and the like in order to grasp the command change situation and the load change situation. Because it controls while monitoring the time change of
It goes without saying that changes in the situation can be easily grasped.
Further, in the present description, the feedforward compensator is not mentioned in order to make the description of the present invention easier to understand, but the above-mentioned applicant's proposal, that is, Japanese Patent Application No. 1-159863.
As is clear from the above issue, it is clear that the command response followability can be improved by adding the feedforward compensator.

〔Example〕

FIG. 1 is a third view to facilitate understanding of the technical idea of the present invention.
Similar to the figure, γ is the final command output. In the figure, the same reference numerals as those in FIGS. 3 and 4 indicate parts having the same functions.

Further, FIG. 2 shows an example of the main hardware configuration for the main blocks shown in FIG.

That is, in FIG. 1 and FIG. 2, assuming that the final input output γ is obtained by changing the setting input R by, for example, operating the instruction setter, etc., this final instruction output γ and a part of the state quantity Then, the deviation e of the output Y becomes T ^* (S) after passing through the stabilizing device 1'comprising the P control device. The T ^* (S) is added to the controlled object 2 by passing it through a power actuator and multiplying it by the torque generation coefficient K _T of the gain. On the other hand, in equivalent disturbance compensation, T ^*
(S) 'and Y are input and the calculation result is added to T ^* (S)' which is the output after the stabilization compensation.

In this way, such a practical example has the same function as that of the above-mentioned solving means, and obtains a special action and effect.

〔The invention's effect〕

As described above, according to the present invention, it is possible to provide a device having a simple structure capable of exhibiting the following useful and high functions.

(1) The main control loop does not need the I element, and is composed of only the P element. The stability of the transient response is improved, and the transient overshoot from the acceleration / deceleration while applying the limiter to the constant speed operation ( Windup) reduction, hardware and software simplification, etc. can be obtained.

(2) The model reference function for preventing windup (speed command device for accelerating and decelerating without saturating the PI system) is also unnecessary.

(3) The overshoot of a very slight fool based on the fast integration included in the equivalent disturbance compensator of FIG. 4 is eliminated,
Overshoot can be eliminated for both acceleration / deceleration and load fluctuation.

[Brief description of drawings]

1 and 2 are main configuration block diagrams and main hardware configuration diagrams shown to facilitate understanding of the technical idea of the present invention, and FIG. 3 is a block diagram showing a conventional example of a stabilized feedback control device. 4 and 5 are a block diagram shown for explaining the equivalent disturbance compensating unit according to the present invention and a block diagram showing a controlled object part after the compensation. Further, FIG. 6 is a system diagram shown for explaining the high-speed integration function according to the present invention, and FIGS. 7 and 10 are diagrams showing simulation and actual data of response waveforms when the main control system is PI. 8 and 11 show the main control system with P
Fig. 9 shows the simulation and actual data of the response waveform in the case of, and Fig. 9 shows the simulation when the reset function is added to the equivalent disturbance compensation. Fig. 12 shows the reset value after the command change of the equivalent disturbance compensator. FIG. 13 is a waveform diagram shown for the purpose of determining the above, and FIG. 13 is a waveform diagram shown for explaining resetting of the equivalent disturbance compensating unit after a load change. 1,1 '... Stabilizer, 2 ... Control object, 3 ... Equivalent disturbance compensator.

Claims (5)

[Claims] 1. An actual speed is negatively fed back to a speed command, and
In a speed control device that performs constant speed operation by applying the deviation to the control target through a stabilizing device, an amount obtained by multiplying the applied input amount to the control target and the actual speed by the inverse function of the mathematical model of the control target. A velocity characterized by estimating an equivalent disturbance by using a deviation between the and, and feeding back the estimated amount to an input amount of a controlled object through a low-pass filter, and providing a stabilizing device composed only of a proportional element. Control device. 2. When the speed command changes from a constant speed command to a speed change command and changes to a constant speed command again, the filter output is reset to a predetermined value when the deviation between the speed command and the actual speed is near zero. And the predetermined value (T _F2 ) is However, the torque generated before and after the speed command changed from the constant speed command to the speed change command is T _B and T _{1, respectively,} and the acceleration / deceleration when the actual speed starts shifting is set to ₁ , and the constant speed command becomes a re-constant speed command. The generated torque immediately before the deviation between the value and the actual speed becomes approximately zero is T ₂ , and the acceleration / deceleration at that time is ₂ . The speed control device according to claim 1, wherein the speed control device is obtained. 3. When the actual speed of the controlled object operated by the speed control device changes due to the influence of load disturbance or the like, the filter output in the vicinity where the amount of change in the actual speed per time becomes zero is stored. 2. The speed control device according to claim 1, wherein the filter output is returned to the stored value when the deviation between the speed command and the actual speed becomes near zero by a speed correction function of the speed control device thereafter. 4. The speed control device according to claim 1, wherein a stationary term is set to zero as a mathematical model of the controlled object in estimating the equivalent disturbance. 5. A feedforward compensator for applying the speed command to the output of the stabilizer through an inverse function of a mathematical model of the controlled object.
The speed control device described.