JP2003131705A - Servo control method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To solve the large problem of being unable to shift to the next work until a residual vibration is suppressed when the residual vibration occurs at the time of positioning. SOLUTION: This two degree of freedom control method in which a feed- forward control unit 2 has a model of mechanism, and in which a controller is constituted so that the model performs a desired operation, and at least one state quantity of the feed-forward control unit 2 is made as an instruction to a feedback control unit 1. The feedback control unit 1 has an observer processing part 13, the feed-forward control unit 2 also has an observer processing part 27, the deviation between an estimated value of the observer processing part 27 of the feed-forward control part 2 and an estimated value of the observer processing part 13 of the feedback control part 1 is calculated, the deviation is multiplied by the gain 14, and the result is added to the state quantity of the feedback control unit 1.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a two-degree-of-freedom control method for machine tools, robot arms, and the like, and particularly to a servo for suppressing vibration when there is a modeling error and for suppressing vibration during feedback of a speed estimated value by an observer. It relates to a control method. The present invention also relates to a servo control method for performing failure diagnosis and abnormality detection using the estimated value of the observer, performing repetitive learning control, and identifying the actual machine model.

[0002]

2. Description of the Related Art In a conventional servo control method, in order to improve a command response, a two-degree-of-freedom control is used in which a feedforward controller controlling a command response and a feedback controller controlling a disturbance response can be independently designed. . Among them, Japanese Unexamined Patent Publication No. 8-168280 discloses the following method as "speed control device for electric motor and speed and position control device". Hereinafter, a brief description will be given with reference to the drawings. In Japanese Patent Laid-Open No. 8-168280, first, the configuration shown in FIG. 10 is given as a conventional example. 1 in FIG.
Is a feedback (hereinafter referred to as FB) control unit, and 2
Is a feedforward (hereinafter referred to as FF) control unit. The meanings of the symbols shown in the figure are as follows. Kp: Position loop proportional gain of feedback control unit Kv: Speed loop proportional gain of feedback control unit Ki: Speed loop integral gain of feedback control unit Kp2: Position loop proportional gain of feedforward control unit Kv2: Speed loop of feedforward control unit Proportional gain Jm2: Motor inertia of model of feedforward control unit S: Laplace operator θref: Position command value θ: Motor position v: Motor speed u: Motor operation amount θff: Feedforward model motor position vff: Feedforward model motor speed uff: Feedforward model motor operation amount Here, 3 is a torque control circuit that creates a command for operating the motor according to the motor operation amount, and 4 is a motor and 5 is a control target. A position / speed detector 6 detects the position and speed of the motor. According to this method, the command response and the disturbance response can be designed independently. However, with this configuration, there is no problem when the controlled object is a rigid body, but when the rigidity of the torque transmission mechanism such as the ball screw or speed reducer used for the XY table or robot arm is low, the feedforward control gain is increased to improve the target value response. It is described that there is a problem that the response becomes oscillating when raising the. On the other hand, Japanese Patent Laid-Open No. 8-168280 discloses the configuration shown in FIG. In FIG. 9, the meaning of each symbol is shown in FIG.
The same as 0. The meanings of other symbols are shown below. Ks2: Torsional angle FB gain Ksd2: Torsional angular velocity FB gain K2: FF model spring constant JL2: FF model load inertia 24, 25, and 26 are added to the configuration of FIG. In this way, the spring element K is used as a mechanical model.
2 and the load inertia JL2 are taken into consideration, the FF controller is configured so that even when the rigidity of the torque transmission mechanism is low,
It is described that a highly responsive control with extremely few overshoots and vibrations can be realized.

[0003]

However, in the above-mentioned conventional "motor speed control device and speed and position control device", when there is an error between the actual control target and the FF model, residual vibration occurs. As an example, FIG. 7 shows a simulation result when the model is displaced. Figure 8
8A shows the case where the model matches the actual machine, and FIG. 8B shows the case where the actual load inertia is 1.5 times the load inertia of the FF model. The parameter values in this simulation are shown below. Kp: 40 (1 / S), Kv: 375 (1 / S), Ki: 20 (1 / S) Jm: 1.16e-4 (kg ・ m ^ 2), JL: 4.53e-4 * 1.5 (kg・ M ^
2), K: 0.23 (N ・ m / rad) Jm2: 1.16e-4 (kg ・ m ^ 2), JL2: 4.53e-4 (kg ・ m ^ 2),
K2: 0.23 (Nm / rad) Residual vibration occurs as shown in FIG. 8B even though it is originally desired to make the response as shown in FIG. 8A. In this way, if residual vibration occurs during positioning,
Until the residual vibration is subsided, the next work cannot be started, which is a big problem. Therefore, an object of the present invention is to provide a servo control method that can solve the above-mentioned problems and that does not cause residual vibration during positioning.

[0004]

According to a first aspect of the present invention, there is provided a servo control method, wherein a feedforward controller has a mechanical model, and the controller is configured so that the model performs a desired operation. In a two-degree-of-freedom control method in which at least one state quantity of the control unit is a command of the feedback control unit, the feedback control unit has an observer processing unit, and the feedforward control unit also has an observer processing unit. The difference between the estimated value of the observer of the forward control unit and the estimated value of the observer of the feedback control unit is calculated, and the difference is multiplied by a gain and added to the state quantity of the feedback control unit. The servo control method according to a second aspect of the present invention is characterized in that estimated values of the observer of the feedforward control unit and the observer of the feedback control unit are filtered. The servo control method according to claim 3, wherein a deviation between an estimated value of the observer of the feedforward control unit and an estimated value of the observer of the feedback control unit is filtered, and a gain is multiplied to obtain a state quantity of the feedback control unit. It is characterized by adding. The servo control method according to claim 4, wherein the feedforward control unit has a mechanical model, and the controller is configured so that the model performs a desired operation, and at least one state quantity of the feedforward control unit is set. In the two-degree-of-freedom control method using the feedback control unit as an instruction, the feedback control unit has an observer processing unit, the feedforward control unit also has an observer processing unit, and an estimated value of the observer of the feedforward control unit. Then, the deviation of the estimated value of the observer of the feedback control unit is obtained, and the deviation is monitored to diagnose the state of the mechanism. The servo control method according to claim 5, wherein the feedforward control unit has a mechanical model, the controller is configured so that the model performs a desired operation, and at least one state quantity of the feedforward control unit is set. In a two-degree-of-freedom control method using a command from a feedback control unit, the feedback control unit has an observer processing unit, and the feedforward control unit also has an observer processing unit, and an estimated value of the observer of the feedforward control unit and feedback are provided. It is characterized in that the deviation of the estimated value of the observer of the control unit is obtained, and the learning control is repeatedly performed so that the deviation becomes small. The servo control method according to claim 6, wherein the feedforward processing unit has a mechanical model, and the controller is configured so that the model performs a desired operation, and at least one state quantity of the feedforward control unit is set. In the two-degree-of-freedom control method using the feedback control unit as an instruction, the feedback control unit has an observer processing unit, the feedforward control unit also has an observer processing unit, and an estimated value of the observer of the feedforward control unit. And the deviation of the estimated value of the observer of the feedback control unit is obtained, and the parameter of the mechanical model is identified while monitoring the magnitude of the deviation.

[0005]

BEST MODE FOR CARRYING OUT THE INVENTION Embodiments of the present invention will be described.
It will be described with reference to the drawings. Example 1 FIG. 1 is a diagram showing the configuration of the first exemplary embodiment of the present invention. In FIG. 1, the meaning of each symbol is FIG. 9 and FIG.
It is the same as that described in 0. The newly added part in FIG. 1 is an observer 13 and an F shown in FIG.
B gain G and three observers configured in the FF portion shown by 27. The difference between the state quantity estimated by the observer of 27 and the state quantity estimated by the observer of 13 is multiplied by the gain G shown in 14 and fed back to the FB controller. Next, a configuration example of the observers shown in 13 and 27 is shown below. The equation (1) is an actual machine model equation in a continuous system, and the equation (2) is a discretization of the equation (1) in the form of an observer. Actually, the equation (2) is used, and the k + 1st variable is sequentially calculated using the kth value.

[0006]

[Equation 1]

FIG. 7 shows a simulation result using this method. The value of each parameter is the same as the value shown in the simulation of FIG. It can be seen from FIG. 8B that the residual vibration is rapidly damped. As described above, according to the method of the first embodiment, it is possible to suppress the vibration even when there is a modeling error between the FF model and the actual machine. Here, the rigid body disturbance observer was introduced as the observer, but the observer of other configurations can naturally obtain the same result. Further, in the present embodiment, a model of two inertia system is used as the actual machine model,
Even if any model is configured according to the actual machine, the same effect as this embodiment can be obtained.

Embodiment 2 Here, a second embodiment of the servo control method according to claim 1 will be described with reference to FIG. With respect to the problem that the feedback gain cannot be increased due to the control calculation delay, a method of compensating the calculation delay by forming an observer and feeding back an estimated speed value of the observer is generally known. Also in this case,
As shown in FIG. 2, an observer is provided in both the FF section and the FB section,

[0009]

[Equation 2]

And the FB section

[0011]

[Equation 3]

The use of is advantageous in that the calculation delay can be improved without inducing vibration.

Third Embodiment Here, a servo control method according to claims 2 and 3 will be described. The structure is shown in FIG. Reference numeral 15 in FIG. 3 represents a filter. As the filter, any of a low pass filter, a high pass filter, a band pass filter, a notch filter, etc. may be used.
Now, consider a case where a certain periodic disturbance acts on the motor.
In this case, in order to suppress the vibration, we usually use the disturbance estimation observer,

[0014]

[Equation 4]

A method is known in which FB is obtained by multiplying by a gain after filtering. However, when the observer is configured with a rigid system, in the case of a mechanism having a spring element, not only the periodic disturbance but also the torsion torque of itself is estimated as the disturbance (equation (4)). When FB is applied, vibration may occur or the locus may worsen during acceleration or deceleration. In order to solve this problem, in this method, the FF section also has an observer,

[0016]

[Equation 5]

From the FB section

[0018]

[Equation 6]

By subtracting, the estimated value of the torsional torque is canceled, and as shown in the equation (5), F is really
Only the periodic disturbance that you want to B can be extracted.

[0020]

[Equation 7]

As described above, according to the third embodiment, it is possible to extract only the true disturbance from the disturbance estimated value, and if it is FB after being filtered, it is possible to have a good control characteristic. .

Fourth Embodiment Here, a servo control method according to a fourth aspect will be described with reference to FIG. If the controlling mechanism fails, is damaged, or collides with other equipment, F
It can be detected by monitoring the disturbance estimated value of the observer in the B section. In this method, the observer is configured in both the FF section and the FB section to monitor the deviation of the estimated disturbance value.
As explained in, even if there is a torsional torque,
Only true disturbance can be extracted, and as a result, diagnosis can be performed with high accuracy. Reference numeral 7 in FIG. 4 is an abnormality monitoring unit that determines the expression (6).

[0023]

[Equation 8]

If the conditional expression (6) is satisfied, the size of the limit may be set so that it is determined to be abnormal.

Embodiment 5: Here, a servo control method according to claim 5 will be described with reference to FIG. In the case of periodic vibration or the like, iterative learning control may be effectively used. Here again, with the configuration shown in FIG. 5, the difference between the disturbance estimation values of the observer in the FF section and the FB section is obtained, and the gain G
It is possible to eliminate the periodic vibration by multiplying by and adding to the torque command value at the time of the next trial, and repeating the operation. Also here, since the observer is configured in both the FF part and the FB part and the deviation of the disturbance estimated value is used,
Even if there is a torsional torque, only the true disturbance can be extracted, and as a result, it is possible to learn with high accuracy.

Embodiment 6 Here, a servo control method according to claim 6 will be described with reference to FIG. Reference numeral 8 in FIG. 6 denotes a parameter identification unit that receives the observer deviation. Also in this case, observers are respectively formed in the FF section and the FB section. Using the two deviations as the evaluation criteria, the parameters are determined or modified so that the value becomes smaller. If there is an error, the deviation will be large, and when the models match, the deviation will be small. As the observer estimated value, the load position estimated value, the load speed estimated value, the disturbance estimated value,
Any estimation value such as a torsion angle estimation value and a torsion angular velocity estimation value may be used. Further, even when a person adjusts the parameter while checking the waveform, there is an effect that the estimation values are compared with each other to facilitate the adjustment.

[0027]

According to the servo control method of the first aspect, the observer is provided in both the FF section and the FB section, and the deviation between the observer estimated value of the FF section and the estimated value of the observer of the FB section is calculated. It is possible to suppress the residual vibration even when there is a modeling error by extracting only the error amount of the above and multiplying the value by the gain and performing FB. Claim 2
According to the described servo control method, the observer is provided in both the FF section and the FB section, and the observer estimated value of the FF section and the FB
Since the deviation of the estimated value of the observer of the part is obtained, only the error component of the model can be extracted, and only the special frequency component can be extracted from the estimated value of the observer and the phase can be adjusted, which enables highly accurate control. . According to the servo control method of the third aspect, the observer is provided in both the FF section and the FB section, and the deviation between the observer estimated value of the FF section and the estimated value of the observer of the FB section is calculated. In addition to being able to extract, only a special frequency component can be extracted from the observer estimated value and the phase can be adjusted, so that control can be performed with high accuracy. According to the servo control method of claim 4, the observer is provided in both the FF section and the FB section, and the difference between the observer estimated value of the FF section and the estimated value of the observer of the FB section is calculated. Since it is possible to extract only the disturbance of, abnormalities can be accurately diagnosed. According to the servo control method of claim 5,
Since observers are provided in both the FF and FB sections and the deviation between the observer estimated value of the FF section and the estimated value of the observer of the FB section is obtained, only the model error and the true disturbance can be extracted, so learning can be performed accurately. be able to. According to the servo control method of the sixth aspect, the observer is provided in both the FF section and the FB section, and the deviation between the observer estimated value of the FF section and the estimated value of the observer of the FB section is calculated. Since it can be extracted, the model can be accurately identified.

[Brief description of drawings]

FIG. 1 is a block diagram showing a configuration of a first embodiment of the present invention.

FIG. 2 is a block diagram showing a configuration of a second embodiment of the present invention.

FIG. 3 is a block diagram showing a configuration of a third exemplary embodiment of the present invention.

FIG. 4 is a block diagram showing a configuration of a fourth exemplary embodiment of the present invention.

FIG. 5 is a block diagram showing a configuration of a fifth embodiment of the present invention.

FIG. 6 is a block diagram showing a configuration of a sixth exemplary embodiment of the present invention.

FIG. 7 Simulation results of the method of the present invention (when the models do not match)

8A is a simulation result of a conventional method (when the models match), and FIG. 8B is a simulation result of the conventional method (when the models do not match).

FIG. 9 is a block diagram showing a configuration of a conventional method.

FIG. 10 is a block diagram showing a conventional configuration in a conventional method.

[Explanation of symbols]

1 Feedback control unit 2 Feedforward controller 3 Torque control circuit 4 motor 5 controlled objects 6 Position speed detector 7 Abnormality monitoring section 8 parameter identification section 11 Position loop processing 12 Speed loop processing 13 FB (feedback) section Observer processing section 14 gain 15 filters 16 dead time 21 FF (feed forward) position loop processing 22 FF speed loop processing 23 FF motor model 24 FF model spring constant 25 FF model inertia part 26 FF Torsional angle feedback section 27 FF section Observer processing section

─────────────────────────────────────────────────── ─── Continuation of front page (51) Int.Cl. ⁷ Identification code FI theme code (reference) H02P 5/00 H02P 5/00 K X F term (reference) 5H004 GA08 GA14 GA28 HA07 HB07 HB08 JB22 KA41 KB13 KB32 KB39 KC28 KC34 KC43 KC45 KD62 LA02 LA13 5H550 AA18 BB05 GG01 GG03 GG10 JJ04 JJ24 JJ25 JJ26 LL01

Claims (6)

[Claims] 1. A feedforward control unit 2 has a mechanical model, and a controller is configured so that the model performs a desired operation. At least one state quantity of the feedforward control unit 2 is fed back to a feedback control unit. In the two-degree-of-freedom control method using the command of 1, the feedback control unit 1 has an observer processing unit 13, the feedforward control unit 2 also has an observer processing unit 27, and the observer of the feedforward control unit 2 The deviation between the estimated value of the processing unit 27 and the estimated value of the observer processing unit 13 of the feedback control unit 1 is calculated,
A servo control method characterized in that the deviation is multiplied by a gain 14 and added to the state quantity of the feedback control section 1. 2. The servo control method according to claim 1, wherein the estimated values of the observer processing unit 27 of the feedforward control unit 2 and the observer processing unit 13 of the feedback control unit 1 are filtered. 3. The deviation between the estimated value of the observer processing unit 27 of the feedforward control unit 2 and the estimated value of the observer processing unit 13 of the feedback control unit 1 is input to the filter 15 and multiplied by the gain 14. 3. The servo control method according to claim 1, further comprising adding to the state quantity of the feedback control unit 1. 4. The feedforward controller 2 has a mechanical model, and a controller is configured so that the model performs a desired operation, and at least one state quantity of the feedforward controller 2 is feedback controlled. In the two-degree-of-freedom control method using the command of the unit 1, the feedback control unit 1 has an observer processing unit 13, the feedforward control unit 2 also has an observer processing unit 27, and the feedforward control unit 2 A servo control method, wherein a deviation between an estimated value of the observer processing unit 27 and an estimated value of the observer processing unit 13 of the feedback control unit 1 is obtained, and the state of the mechanism is diagnosed by monitoring the deviation. 5. The feedforward control unit 2 has a mechanical model, and a controller is configured so that the model performs a desired operation. At least one state quantity of the feedforward control unit 2 is fed back to the feedback control unit. In the two-degree-of-freedom control method using the command 1, the feedback control unit 1 has an observer processing unit 13, the feedforward control unit 2 also has an observer processing unit 27, and the observer processing unit of the feedforward control unit 2 27 estimated values and the feedback control unit 1
The servo control method is characterized in that the deviation of the estimated value of the observer processing unit 13 is obtained, and the learning control is repeatedly performed so that the deviation becomes small. 6. The feedforward control unit 2 has a mechanical model, and a controller is configured so that the model performs a desired operation, and at least one state quantity of the feedforward control unit 2 is fed back to the feedback control unit. In the two-degree-of-freedom control method using the command of 1, the feedback control unit 1 has an observer processing unit 13, the feedforward control unit 2 also has an observer processing unit 27, and the observer of the feedforward control unit 2 The deviation between the estimated value of the processing unit 27 and the estimated value of the observer processing unit 13 of the feedback control unit 1 is calculated,
A servo control method, wherein the parameter of the mechanical model is identified while monitoring the magnitude of the deviation.