JPH0764605A - Adaptive controller - Google Patents

Abstract

PURPOSE:To provide an adaptive controller which can improve its control performance to the disturbance by setting the parameter of a compensator in order to eliminate the estimated/updated parameter by application of the adaptive control and by securing the coincidence between the control system output and a speed or control command. CONSTITUTION:The target value is supplied to an adder 1, and the feedback value omega and the deviation (e) are transmitted. A compensator 2 sends a manipulated variable V to an estimator 4 as well as to a controlled system consisting of an amplifier 31, a motor 32 and an encoder 33 based on the deviation (e). The encoder 33 transmits the actual speed of the motor 32 as the feedback value omega and supplies it to the adder 1 and the estimator 4. The estimator 4 operates the parameter of a servo motor and the estimated value of disturbance (torque) based on the variable V and the value omega, and the compensator 2 controls its compensation value in order to eliminate the calculated parameter. This processing can be controlled as if the amplifier 31 of the controlled system, the motor 32 and the encoder 33 were not provided. Furthermore the compensator 2 can secure the coincidence between a speed command and the output of a control system.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION The present invention relates to the processing of a servo system by CP.
The present invention relates to an adaptive control device using a software servo system performed by a U / DSP, and particularly to a control system that compensates for disturbance. The present invention relates to adaptive control applied to an adaptive control device in order to improve the limit of control performance by adding a disturbance compensation control system to a conventional control system, and more particularly to a mechanism driven by a servo motor.

[0002]

2. Description of the Related Art In response to changes in the environment surrounding a control system,
There is an adaptive control system that changes the control input to best fit the new environment. Since the conventional feedback control system matches the control amount to a desired value under unpredictable disturbance, as shown in FIG. 9, the adder 1, the compensator 2, and the controlled object 3 to which the target value ωref is input are provided. The output ω of the controlled object 3 is fed back to the adder 1 to form a closed loop. Since the disturbance cannot be measured directly, its effect can be known only by observing the controlled variable. In this case, there is a time delay until the controlled variable is affected by the disturbance. The compensator 2 works to make the deviation e between the target value ωref and the feedback value ω zero. Adjustment of the manipulated variable V based on the deviation e,
There is a time delay in the change of the control amount ω based on the operation amount, and the suppression of the disturbance is dealt with with a time delay of one closed loop. This causes the control system to become unstable. In such a control system, when the disturbance can be directly measured, good control characteristics can be expected if the control system is configured so as to cancel the influence in advance (feedforward control system).

The conventional adaptive control is not a system that copes with disturbance. Therefore, sufficient control performance was not obtained when disturbance was applied. JP-A-1-12931
In the vibration suppression method for software servo disclosed in Japanese Patent Publication No. 3, it takes time to stabilize and converge the system when a disturbance is introduced.

[0004]

The present invention has been proposed in view of the above circumstances, and an object thereof is to provide an adaptive control device of a control system capable of improving control performance against disturbance. is there.

[0005]

According to the present invention, the parameters of a servo motor are estimated / applied by applying adaptive control.
By updating and estimating / updating the disturbance (torque), and setting the compensator parameters so that these parameters are canceled by the compensator using these parameters, the control system output matches the speed or position command. ing.

[0006]

According to the present invention, the disturbance is estimated and the setting of the parameter of the servo motor is changed. Further, the disturbance is measured to improve the estimation accuracy, and the setting is changed so as to cancel the disturbance, and at the predetermined timing. By changing the parameter setting at a predetermined timing, the system vibration due to the disturbance is prevented. This improves control performance against disturbance.

[0007]

DESCRIPTION OF THE PREFERRED EMBODIMENTS A case where the adaptive control means of the present invention is applied to a mechanism driven by a servo motor will be described. The DC motor that moves the load with the total inertia J can be expressed by the following equation.

[Equation 1]

[Equation 2] (Here, R: resistance, L: inductance, Kt: torque constant, Ke: induced voltage constant, B: damping coefficient,
v: supply voltage, i; armature current, Td: disturbance torque, ω: speed, J: total inertia. ) Equations (1) and (2) can be obtained by performing the Laplace transform operation.
Is given by the following transfer function.

[Equation 3] The corresponding transfer function in the z region using the zero-order hold is expressed by the following relationship using velocity: ω, voltage: v, and disturbance torque: Td.

[Equation 4] (Here, b ₁ is a coefficient related to 0th-order hold, z is a discrete value expression.) The discrete dynamic equation of the system can be written in the time domain using the following recurrence equation. Where the system factor is θ
It is represented by i.

[Equation 5] This can be written as

[Equation 6] Where the new factor is

[Equation 7] And the system equation can be expressed in vector notation.
Here, the vector θ is composed of all known signals. And (″) written in the upper right shows the replacement (transversion) of the vector.

[Equation 8] here,

[Equation 9] Is. If the disturbance torque changes constantly, the signal vector factor has an unknown term. And in many cases the disturbance torque result is unknown. As a result, the estimation process is Φ
It is non-linear due to the unknown term of the vector.

Now, the signal and factor vector can be revised as follows.

[Equation 10] -K ₂ Td (K) is the remaining result given by ω (K + 1) -Φ (K-1), so

[Equation 11] Where the new θ and Φ vectors are defined below.

[Equation 12] The recursive estimation method is given by the following equation. The ^ in this equation represents the estimate, and the vector G represents the gain.

[Equation 13] Since all elements of the Φ vector are unknown in this equation, the unknown factors Td (k) and Td (k-1) are replaced by these residual (error) characteristics.

[Equation 14] The recursive equation for obtaining the vector gain G is similar to the Kalman filter equation.

[Equation 15]

[Equation 16] Here, P (K) is a common (mutual / complementary) variation matrix.
It is initialized by setting P (0) equal to the diagonal matrix. The elements of the factor vector θ are initialized by some initial estimation. Controller design can be performed by using the estimation of system factors instead of the actual values.

The desired reference velocity during the next sample time is equal to the prior velocity estimate obtained by:

[Equation 17] The compensation for the disturbance torque will be described below. The transfer function is

[Equation 18] There are two control targets as can be understood by looking at the transfer function of the motor shown by. On the other hand, the current compensation is

[Formula 19] Only compensation has been made. This is because the disturbance torque: Td has not been conventionally measured.

FIG. 1 is a diagram showing the concept of the adaptive control apparatus of the present invention, which is an example using one estimator. The process here is to estimate the parameters of Gp (z). The target value ωref is input to the adder 1 and the deviation e from the feedback value ω
Is output. The compensator 2 operates the manipulated variable v based on the deviation e.
Is output to the control target including the amplifier 31, the motor 32, and the encoder 33 that detects the motor output at the rotation speed, and the estimator 4. The encoder 33 outputs the actual speed as a feedback value ω, and the feedback value ω is fed back to the adder 1 and also input to the estimator 4 as the other signal. Estimator 4 is manipulated variables v and the estimated value K of K ₁ and a feedback value omega ₁ ^, estimated values a ₁ of a ₁
^, Estimated values of a ₂ a ₂ ^, estimated value K ₂ Td of K ₂ Td ^, c
₁ estimate c ₁ ^ calculated, and controls the compensation amount of the compensator 2.

The processing performed here is POLE-ZERO CANCEL
ATION and can be equivalently changed as shown in FIG.
To further summarize this, as shown in FIG. 3, it is as if the amplifier 31, the motor 32, and the encoder 33 of the control target 3 are controlled.
Can be controlled so that it does not exist, and operates as if there is one integrator. The estimator 4 estimates θ ₁ , θ ₂ , θ ₃ , θ ₄ , and θ ₅ and calculates v (k) in the equation (13) based on the estimation result to calculate the compensator 2
Set with.

[Embodiment 2] In this embodiment, the disturbance torque T
It will be explained that d is measured by current and compensated. As shown in FIG. 4, in this embodiment, two compensators A and B are provided.
Use one compensator. The target value ωref and the feedback value ω are input to the adder 1, and the deviation e is output to the compensator A21. The compensator A21 calculates the manipulated variable v using the estimated values θ ₁ ^ to θ ₆ ^ relating to θ ₁ to θ ₆ from the estimator 4, and outputs it. The manipulated variable v is output to the second adder 11 and the estimator 4, and the output formula (18) of the compensator B22 is input to the adder 11. The compensator B22 outputs θ ₁ ~ from the estimator 4.
The manipulated variable v is calculated using the estimated values θ ₁ ^ to θ ₆ ^ for θ _6.
(K) Calculate and output to the adder 11.

The compensator B22 will be described in more detail with reference to FIG.

[Equation 20]

[Equation 21] In the constant speed state, the acceleration α is almost 0,

[Equation 22] And the ratio of the disturbance Td to the output v of the adder 1 is

[Equation 23] And from this equation, the output v of the adder is

[Equation 24] It is represented by.

The gain Gc of the compensator A is

[Equation 25] It is represented by. Where Tg: generated torque, Ta: acceleration torque, Td: disturbance torque (load), Kt: torque constant,
Ke: induced voltage constant, I: motor current, V: supply voltage,
Vd: compensator output, L: inductance, R: resistance,
J: inertia, α: acceleration, B: damping constant, Gc: compensator, T′d: = Tg−Ta = Tg−Jα≈Kt
I-Jα In a low speed state, acceleration α = 0 and T′d = K
It can be interpreted as tI.

By this compensation, the disturbance Td applied to the controlled object as shown in FIG. 6 is the output Co of the compensator.
Is canceled by, and a substantially flat result ω is obtained. According to this correction method, as compared with the method shown in FIGS. 1 to 4, since the disturbance is measured, it is possible to obtain the further effect of improving the quick response and stability.

[Embodiment 3] This embodiment is an example of a control device used in a mechanism such as a motor for driving a transfer roller of a copying machine, which always causes a large disturbance at a certain timing. When such a mechanism is controlled by the conventional adaptive control, the parameters of the compensator are changed after the disturbance has occurred, so there is a time delay in the control and there is a delay in the convergence of the hunting that occurs for a large disturbance. In the present embodiment, there is a problem that it takes time, and for an extremely large disturbance whose timing and its magnitude are known in advance, parameters are estimated and identified in advance, according to the timing at which the disturbance occurs, Feedforward control is performed by changing the parameters of the feedforward compensator according to the disturbance.

This embodiment will be described with reference to FIGS. 7 and 8. This control device includes an adder 1 that inputs a target value ωref and a feedback value ω from a controlled object 3 and outputs a deviation e thereof,
The compensator A21 that calculates the control amount Va of the controlled object based on the deviation e and outputs the result, the feedforward compensator 23 that calculates the feedforward compensation amount Vf based on the target value ωref, and the compensator A21. The output Va and the output Vf of the feedforward compensator 23 are added to obtain the supply voltage V to the controlled object 3, and the supply voltage V and the feedback value ω are added to obtain the final supply voltage Vc. Obtaining adder 11, controlled object 3, feedback value ω and disturbance Td
Also, at the timing when the supply voltage V is input and the disturbance whose magnitude is known in advance is input, the compensator 23 is input in advance.
And a parameter estimator 41 for estimating so as to change the parameter. The controlled object 3 is (Kt / (Ra +
LaS)) 34 and (1 / (Js + B)) 35, and the output ω is fed back to the adder 11 via the induced voltage constant (Ke) 99.

The calculation in the parameter estimator 41 is performed as follows, and the output is given by the following equation (20).

[Equation 26] The parameters are changed and identified in the feedforward compensator 23 by the following equation (21), and the timing thereof is controlled by a timer (not shown).

[Equation 27] Further, the compensator A21 is compensated by the following equation (22). This compensation is performed by fixed coefficient compensation.

[Equation 28] Here, in order to bring E (s) to 0, G
₂ (s) Gp (s) ≈1 or G ₁ (s) G
There are two methods of setting p (s) → ∞. In this case, bringing G ₁ (s) to ∞ causes many troubles and limits of the limiter, so G ₂ (s) Gp
Control is performed by setting (s) ≈1. That is, G
₂ (s) = 1 / Gp (s) is set. As described above, the disturbance is an irregular disturbance (for example, indicated by Td ₁ ), the timing of its occurrence and termination, and the magnitude of which is predicted in advance and the magnitude of which is extreme (for example, indicated by Td ₂ ). In some cases, if the same compensator parameter is always used for coping, for example, when an extreme disturbance indicated by Td ₂ occurs, control is performed with a parameter that does not correspond to the disturbance, and sufficient control performance is obtained. May not be obtained. Further, even if the control is performed while performing the estimation in real time, it may take time until the parameter estimation is completed, that is, the convergence, and it is possible that sufficient control performance cannot be obtained. Therefore, when the situation of disturbance occurrence can be considered by dividing it into Td ₁ and Td ₂ described above, that is, Td ₂
When the timing of occurrence and the timing of termination and their magnitude are known in advance, the parameters given to the compensator at Td ₁ and Td ₂ are changed, or the initial value used for estimation is changed to achieve fine grained control. It is possible to obtain performance. In this case, the setting is changed immediately before Td ₂ occurs and immediately after Td ₂ . Further, the reason for using the feedforward control is to reduce the time delay due to the control target meter. As shown in FIG. 8, the compensation operation by the feedforward compensator 23 operates in advance of tf at the time of occurrence of a disturbance whose timing is known in advance, and compensates for a large disturbance without a time delay. You can

[0019]

As described above, according to the present invention, it is possible to obtain sufficient control performance corresponding to the disturbance even if it is applied.

[Brief description of drawings]

FIG. 1 is a diagram showing a concept of a control device according to the present invention.

FIG. 2 is a diagram showing a concept of a control device according to the present invention.

FIG. 3 is a diagram showing a concept of a control device according to the present invention.

FIG. 4 is a conceptual diagram of a control device according to a second embodiment of the present invention.

FIG. 5 is a conceptual diagram of a control device according to a second embodiment of the present invention.

FIG. 6 is a diagram showing effects of the control device according to the second embodiment of the present invention.

FIG. 7 is a conceptual diagram of a control device according to a third embodiment of the present invention.

FIG. 8 is a diagram showing an operating state of a control device according to a third embodiment of the present invention.

FIG. 9 is a diagram showing the concept of a conventional feedback control device.

[Explanation of symbols]

1, 11, 12: adder, 2, 21, 22, 23: compensator, 3: control target, 4, 41: estimator.

Claims (4)

[Claims] 1. An adaptive control device for a control system, means for modeling and adapting a control amount for a control system including disturbance, so that a deviation between an output obtained by this means and an output of the control system approaches zero. Means for modifying the parameters of the model. 2. In an adaptive control device for a control system, means for modeling and adapting a control amount for disturbance, means for modeling and optimizing the control system, and outputs and control system outputs obtained by these means. And a means for correcting the parameters of the model so that the deviation of the model becomes close to zero. 3. An adaptive control device for a control system having a feedforward compensator, means for modeling and adapting a control amount for disturbance, means for modeling and adapting the control system, and these means Means for modifying the parameters of the model so that the deviation between the output of the control system and the output of the control system approaches zero, and means for adding the output signal of the feedforward compensator obtained by the means. An adaptive control device characterized in that 4. In a control system adaptive controller having a feedforward compensator, the parameters of the model are set so that the deviation between the means for modeling the disturbance and the controlled object by adaptive estimation and the deviation of the control system output approaches zero. An adaptive control device comprising: means for correcting; and means for previously correcting compensator parameters as needed for large periodic disturbances.