JP7293905B2 - friction compensator - Google Patents

Description

The present invention is intended for a servo control system in which a controlled variable such as the position and attitude of a driven object follows a target value, in particular to suppress the influence of mechanical friction that occurs when the speed (operating direction) of the driven object is reversed. Friction compensation technology.

In the servo control system, in the feedback control system that corrects the error between the position command of the driven body and the position detection value, after the position error is detected, a driving force corresponding to the error is applied to the controlled body. responds with a delay from the time of occurrence of the position error. In addition, the frictional force inherent in the drive system of the servo control system exhibits non-linear characteristics particularly when the speed (operating direction) of the driven body is reversed.
These phenomena cause a significant deterioration in the motion control accuracy of the servo control system.

For a driven body that requires high-precision position control, it is important to reduce the influence of frictional disturbances on control accuracy during speed reversal. Friction compensation techniques and model-based feedforward friction compensation techniques exist.
For example, it is possible to estimate disturbances including friction using a disturbance observer and perform disturbance suppression compensation by feedback control to suppress the effects of friction. known to be effective.
A technique is also generally used in which disturbance estimates are calculated using disturbance models corresponding to a plurality of state quantities, and a disturbance correction amount is generated from the sum of these disturbance estimate values.

In recent years, model-based development has spread to the servo industry, and feedforward friction compensation technology has been widely adopted.
For example, FIG. 14 is a functional block diagram of a servo control system that includes a velocity feedback control system, a position feedforward compensator, and a friction compensator, and is disclosed in Patent Document 1.
In the figure, 210 is a position controller, 220 is a speed controller that generates a current command i ^* , 230 is a feedforward compensator, 240 is a friction compensator that estimates a friction force (friction torque) F, 241 is a Coulomb friction A parameter identification unit that identifies various parameters such as force, a friction estimation unit 242 that estimates the frictional force F based on various parameters, a current command conversion unit 250 that converts the frictional force F into a current compensation command _idis , and a 300 A driver (driving device) 400 composed of a power converter or the like indicates a machine tool as a controlled object including an electric motor and a driven body.

The parameter identification unit 241 in FIG. 14 generates the actual machine friction characteristics based on the current command i _c ^* and the position of the driven body of the machine tool 400. From the actual machine friction characteristics, the Coulomb friction force F _c , the maximum static The frictional force F _s , the maximum frictional force ratio α _i of each element, and the spring constant k _i are identified as parameters. Then, the identified parameters are employed in the friction model, and the friction model and the speed of the driven body are used to calculate the verification friction characteristics indicating the relationship between the position of the driven body and the frictional force.
Then, parameters such as the Coulomb friction force F _c and the maximum static friction force F _s are adjusted until the error between the verification friction characteristic and the actual machine friction characteristic falls within the allowable range. _{to the friction} estimator 242 _.

The friction estimator 242 estimates the frictional force F using the latest identification parameter input, and the current command conversion unit 250 converts the frictional force F into a current compensation command _{i_dis} and outputs it.
By correcting the output i ^* of the speed controller 220 using this current compensation command i _dis and the output of the feedforward compensator 230, the friction-compensated current command i _c ^* is given to the driver 300. .

Japanese Patent No. 6214948 ([0022] to [0029], [0042] to [0046], Figures 1 to 3, Figure 11, etc.)

The prior art according to Patent Literature 1 described above has the following problems.
First, considering the dynamic characteristics of position and speed, the structure of the friction model described as formulas (1) to (4) in Patent Document 1 is complicated, so it is difficult to model with high accuracy and identify parameters. there is a risk of becoming
On the other hand, the difference between the verification friction characteristics calculated based on the parameters identified under certain conditions and the actual machine friction characteristics depends on the system identification method, the accuracy of the position/speed detector, noise, etc. There will always be errors. Although Patent Literature 1 states that each parameter is adjusted so that the above error is within the allowable range, it does not disclose any specific procedure.

Further, in Patent Document 1, since the dependence characteristics of the Coulomb friction force on speed and acceleration cannot be expressed, there are cases where robustness such as reproducibility of control performance cannot be guaranteed with changes in operating conditions. Moreover, under special conditions where the speed is reversed and the driven body moves in the opposite direction, even if the Coulomb friction force and maximum static friction force are adjusted, the difference between the verification friction characteristics and the actual machine friction characteristics may not fall within the allowable range.
Furthermore, considering that there is a response delay in the drive system of the machine tool, even if the above error is within the allowable range by adjusting the parameters, it is possible to always solve the decrease in control accuracy due to the influence of friction. is not limited.

Therefore, the problem to be solved by the present invention is to use a friction model that is as simple as possible to simultaneously consider both static friction characteristics and dynamic characteristics, so that the friction force can be calculated with high accuracy even when the speed (movement direction) of the driven body is reversed. To provide a friction compensator capable of realizing appropriate friction compensation by estimation.

In order to solve the above-mentioned problems, the invention according to claim 1 provides a friction control system for compensating for the frictional force acting on the driven body in a servo control system that causes the control amount of the position and attitude of the driven body to follow a target value. A compensator,
Friction compensation for the friction force added to the output of the speed feedback control system using a friction model including the speed of the driven body, the rise coefficient of the friction force with respect to this speed, the Coulomb friction force, and the viscous friction coefficient Equipped with friction compensation amount estimation means for estimating as an amount,
The friction compensation amount estimating means is
Using the input/output information of the servo control system and the friction model, the rise coefficient, the Coulomb friction force, and the viscous friction coefficient are identified as parameters of the friction model, and the identified parameters are used as initial values for different acceleration conditions. a model parameter calculator that updates and outputs the parameters under
The friction compensation amount is initialized using the parameter output from the model parameter calculation unit, and the Coulomb friction force is updated so that an error in the control amount near the speed reversal of the driven body is minimized. Then, obtain the relationship between a plurality of accelerations corresponding to a plurality of speed test signals and the rise coefficient, and determine the friction compensation amount based on this relationship, the speed and acceleration of the driven body, and the Coulomb friction force. a model-based friction compensator that estimates
characterized by comprising

The invention according to claim 2 is the friction compensating device according to claim 1, wherein the model parameter calculation unit calculates the frictional force based on the actual machine frictional characteristic indicating the relationship between the speed or acceleration of the driven body and the frictional force. It is characterized by identifying the parameters of the model.

The invention according to claim 3 is the friction compensating device according to claim 1 or 2, wherein the friction model continuously detects a discontinuity in which the speed of the driven body changes from positive to negative or from negative to positive. It is characterized by being a model approximated by a function.

The invention according to claim 4 is the friction compensation device according to any one of claims 1 to 3, wherein the hysteresis width in the relationship between the acceleration of the driven body and the rise coefficient is used as a parameter of the friction model. characterized by comprising

According to the present invention, the amount of friction compensation is dynamically determined according to the operating conditions, and model-based feedforward friction compensation is performed, thereby reducing control errors due to the effects of friction during speed reversal and providing appropriate friction compensation. can be realized.

1 is a functional block diagram showing one configuration example of a servo control system according to an embodiment of the present invention; FIG. FIG. 4 is a diagram showing an example of velocity-frictional force characteristics in an embodiment of the present invention; FIG. 2 is a functional block diagram of a model-based friction compensator in FIG. 1; FIG. 4 is a diagram showing velocity-frictional force characteristics with hysteresis characteristics in an embodiment of the present invention; 2 is a functional block diagram of a model parameter calculator in FIG. 1; FIG. It is a figure which shows the relationship between the acceleration and a rise coefficient in embodiment of this invention. FIG. 4 is an explanatory diagram of a speed test signal in the embodiment of the present invention; 4 is a flow chart showing parameter tuning operation in the embodiment of the present invention; It is a figure which shows the relationship between the acceleration and a rise coefficient in embodiment of this invention. 4 is a flow chart showing a procedure for calculating a friction compensation amount in the embodiment of the present invention; FIG. 10 is a diagram showing measurement results of frictional force according to acceleration near the time of speed reversal, for an actual machine; FIG. 7 is a diagram showing estimation results of frictional force (friction compensation amount) according to acceleration near the time of speed reversal when the embodiment of the present invention is applied; FIG. 5 is a diagram showing simulation results of velocity followability with and without friction compensation according to the embodiment of the present invention; 1 is a functional block diagram of a servo control system described in Patent Document 1; FIG.

An embodiment of the present invention will be described below with reference to the drawings. In addition, the technical scope of the present invention is not limited at all by the embodiments described below.
First, FIG. 1 is a functional block diagram showing one configuration example of a servo control system including the friction compensator of this embodiment. A system is assumed in which a current command or the like is given to a function G) to control the position of the driven body.

In FIG. 1, the deviation between the position command sent from the host control device and the position output of the controlled object 30 is calculated by the subtracting means 81, and the position feedback controller 50 operates to make the deviation zero, thereby increasing the speed. Output directive. The position output of the controlled object 30 is differentiated by the differentiating means 91 (s is the Laplace operator) and converted into a detected speed value. The subtracting means 82 calculates the deviation between the speed command and the detected speed value, 20 operates to make the deviation zero and outputs a current command (or torque command).

In addition, the position command is input to an inverse system 50 for feedforward compensation (assumed to be G ⁻¹ for the transfer function G of the controlled object 30 ), and its output is sent to the velocity feedback controller 40 by the addition means 83 . Added to the output.
The output of the adding means 83 is given to the controlled object 30 via the adding means 84, the subtracting means 85 and the adding means 86. The adding means 84 adds the output of the model-based friction compensator 70, which will be described later, to the output of the adding means 83, and the subtracting means 85 subtracts the disturbance estimated by the disturbance observer 40 from the output of the adding means 84. The adding means 86 adds the output of the subtracting means 85 and the disturbance signal d.

Furthermore, the position command is input to the differentiating means 93 and the second-order differentiating means 92 to calculate a velocity command and an acceleration command, which are applied to the friction model of the model-based friction compensator 70 for compensating for control errors caused by friction. is entered. The acceleration command, which is the output of the second-order differentiating means 92 , is input to the model parameter calculator 60 , and the parameters of the friction model identified by this calculator 60 are input to the model-based friction compensator 70 . The model-based friction compensator 70 calculates the frictional force (friction compensation amount) to be compensated based on the identified parameters, the speed command and the acceleration command, and outputs it to the adding means 84 .
Here, the model parameter calculator 60 and the model-based friction compensator 70 constitute friction compensation amount estimating means in the claims.

The model parameter calculator 60 identifies a plurality of parameters using, for example, a simple friction model as shown in Equation 1 below, and the model-based friction compensator 70 estimates the frictional force T _f based on these parameters. do.
[Formula 1]
_Tf =Erf(αω) _Tcf +Dω
Here, Erf is the error function, α is the rise coefficient representing the slope of rise of the friction force, ω is the speed of the driven body, _Tcf is the Coulomb friction force, and D is the viscous friction coefficient.

While the Coulomb frictional force changes discontinuously from positive to negative or from negative to positive with the zero speed of the driven body as a boundary, the frictional force _Tf can be approximated using the error function Erf as shown in Equation 1. By doing so, as shown in the speed-friction force characteristics of FIG. 2, static friction characteristics that smoothly change with a predetermined slope as the speed approaches zero can be obtained.
Note that the function of the model parameter calculator 60 will be described later together with FIG.

As shown in the functional block diagram of FIG. 3, the model-based friction compensator 70 of FIG. ing.

The model-based friction compensation amount calculator 71 calculates the frictional force to be compensated based on the parameters of the friction model identified by the model parameter calculator 60 and speed information (speed command and acceleration command).
The phase adjustment unit 72 adjusts the velocity-frictional force characteristic of FIG. 2 to a characteristic having a predetermined hysteresis as shown in FIG. 4, for example. That is, when the speed reverses from negative to positive and from positive to negative, the phase of the frictional force is adjusted so that the frictional force becomes equal to the Coulomb frictional force _Tcf , and the speed-frictional force characteristic is Smooth with a given slope. As a result, the responsiveness of the servo control system can be enhanced to compensate for the hysteresis characteristic of the frictional force of the actual machine, and the friction compensation amount before and after the speed reversal can be stabilized. The hysteresis width can be arbitrarily set by adjusting the phase of the frictional force by the phase adjustment unit 72, and the hysteresis width may be included in the parameters of the friction model.
Of course, instead of the above, friction modeling considering hysteresis characteristic compensation and phase compensation may be performed.

Returning to FIG. 3, the gain adjustment and output limiter 73 following the phase adjuster 72 is for maintaining numerical stability of the frictional force to be compensated and preventing overcompensation. Specifically, each parameter identified by the model parameter calculation unit 60 is substituted into the friction model of the above-described Equation 1 to convert the initial value of the calculated frictional force to a magnitude that is multiple times, for example, about three times.
An output filter 74 subsequent to the gain adjustment and output limiter 73 has, for example, a low-pass filter characteristic. (more specifically, the output of the adding means 83) and the result is input to the adding means 85.

Next, functions of the model parameter calculator 60 will be described with reference to FIG.
As shown in FIG. 5, the model parameter calculation unit 60 identifies various parameters of the friction model by system identification using the friction model of Equation 1 and previously acquired input/output information of the servo control system (step S1 ). Specifically, for example, various parameters are identified based on actual machine friction characteristics that indicate the relationship between the speed or acceleration of the driven body and the frictional force. Using these identified parameters as initial values, different acceleration conditions are given and dynamic parameters are calculated by tuning (S2). Next, the relationship between the acceleration and the parameter is recorded (S3), and the acceleration information during actual driving is collated with the recorded relationship to dynamically update the parameter (S4). The parameters obtained are output to the model-based friction compensator 70 .

Here, Equation 1 is intended for a friction model that considers only static characteristics of frictional force with respect to velocity. For example, even if the speed-friction force characteristics are adjusted as shown in FIG. 4, the rise coefficient of the friction compensation amount is a constant value. The frictional force in the very small speed region near the time of reversal cannot be reproduced accurately.
For this reason, in the present embodiment, parameters are dynamically obtained by tuning the model parameter calculation unit 60 described above so as to reflect the dynamic behavior of the frictional force in the micro-velocity region near the time of velocity reversal. .

The tuning operation by this model parameter calculator 60 will be further described with reference to FIGS. 7 and 8. FIG.
For example, using the trapezoidal speed test signal shown in FIG. 7 (±ω _r is the maximum speed and T is the acceleration time), the operable speed range of the servo control system (range from −ω _r to +ω _r in FIG. 7) A first velocity test signal and a second velocity test signal with different accelerations are predefined.

Then, first, the estimated frictional force T _f of Equation 1 is initialized using the identified parameters (step S11 in FIG. 8). After the initialization is completed, the servo control system is driven by the first speed test signal (S12). Next, while adjusting the set value of the Coulomb friction force _Tcf among the parameters of the friction model (at step S13), the control amount error (for example, the position target value of the driven body and the position detection value error) is minimized (S14), and when the control amount error is minimized (Yes at S14), the Coulomb friction force is updated with the set value at that time (S15).
Next, the rise coefficient α of the friction compensation amount is adjusted (S16). Then, the rise coefficient α is repeatedly adjusted until the error in the control amount near the speed reversal is minimized (S17), and when the error in the control amount is minimized (Yes in S17), the rise coefficient It is stored as the coefficient _α1 (at S18).

Subsequently, the servo control system is driven by the second speed test signal to adjust the rise coefficient α (at S19 and S20), and the adjustment of the rise coefficient α is repeated until the error in the control amount near the speed reversal is minimized. (S21), when the control amount error becomes minimum (Yes in S21), the rising coefficient α at that time is stored as the second coefficient _α2 (S22).
Using the first coefficient _α1 and the second coefficient _α2 saved by the above processing and the upper and lower limits of the rise coefficient, the relationship between the acceleration a and the rise coefficient α is obtained by a first-order linear approximation function as shown in FIG. is obtained (same S23).

Next, FIG. 10 is a flow chart showing the operation of estimating the friction compensation amount by the model-based friction compensator 70 of FIG.
First, the operating information such as the current speed and acceleration of the servo control system in operation is acquired (step S31), each parameter of the friction model is updated using this operating information (step S32), and these parameters Based on this, the friction compensation amount is calculated by the formula 1 and output (at steps S33 and S34).
Specifically, from the first coefficient _α1 and the second coefficient _α2 obtained by the processing of FIG. Since the relationship with the coefficient α has been obtained and the Coulomb friction force has been obtained in step S15 of _FIG . be able to.
Here, if there is a change in the characteristics of the servo control system over time or a change in the surrounding environment, dynamic parameter calculation (step S2 in FIG. 5) by the tuning operation of the model parameter calculation unit 60 described above is performed again. It can be handled if implemented.

11A and 11B are diagrams showing the measurement results of the frictional force according to the acceleration in the vicinity of the speed reversal for the actual machine. According to FIG. 11, it can be seen that the smaller the acceleration, the larger the fluctuation of the frictional force near the speed reversal.
On the other hand, FIG. 12 is a diagram showing the estimation result of the frictional force (friction compensation amount) according to the acceleration near the speed reversal when the embodiment of the present invention is applied. In the embodiment of the present invention, the estimated frictional force _Tf is obtained by Equation 1 using the relationship between the acceleration a and the rising coefficient _α shown in FIG. It is possible to dynamically obtain a smoothly changing friction compensation amount regardless of the magnitude, and to reduce control errors due to the influence of friction.

Furthermore, FIG. 13 is a diagram showing simulation results of velocity followability with and without friction compensation according to the implementation of the present invention.
When friction compensation is performed according to the embodiment of the present invention, the speed changes continuously without being affected by friction before and after the speed reversal. It can be seen that there is a period during which speed change is hindered by friction.

As described above, according to the present embodiment, the friction compensation amount can be estimated with high accuracy for different accelerations in consideration of the static and dynamic characteristics of friction. and speed control.

10: Position feedback controller 20: Velocity feedback controller 30: Control object 40: Disturbance observer 50: Inverse system 60: Model parameter calculator 70: Model-based friction compensator 81, 82, 85: Subtracting means 83, 84, 86 : addition means 91, 93: differentiation means 92: second-order differentiation means

Claims (4)

A friction compensator for compensating for frictional force acting on a driven body in a servo control system that causes a control amount of the position and attitude of the driven body to follow target values,
Friction compensation for the friction force added to the output of the speed feedback control system using a friction model including the speed of the driven body, the rise coefficient of the friction force with respect to this speed, the Coulomb friction force, and the viscous friction coefficient Equipped with friction compensation amount estimation means for estimating as an amount,
The friction compensation amount estimating means is
Using the input/output information of the servo control system and the friction model, the rise coefficient, the Coulomb friction force, and the viscous friction coefficient are identified as parameters of the friction model, and the identified parameters are used as initial values for different acceleration conditions. a model parameter calculator that updates and outputs the parameters under
The friction compensation amount is initialized using the parameter output from the model parameter calculation unit, and the Coulomb friction force is updated so that an error in the control amount near the speed reversal of the driven body is minimized. Further, obtaining relationships between a plurality of accelerations corresponding to a plurality of speed test signals and the rise coefficient, and calculating the friction compensation amount based on the relationships, the speed and acceleration of the driven body, and the Coulomb friction force. a model-based friction compensator that estimates
A friction compensator comprising: In the friction compensator according to claim 1,
The friction compensator according to claim 1, wherein the model parameter calculator identifies the parameters of the friction model based on actual machine friction characteristics indicating the relationship between the velocity or acceleration of the driven body and the friction force. In the friction compensation device according to claim 1 or 2,
A friction compensation device according to claim 1, wherein the friction model is a model that approximates a discontinuity in which the speed of the driven body changes from positive to negative or from negative to positive with a continuous function. In the friction compensation device according to any one of claims 1 to 3,
A friction compensator according to claim 1, wherein a hysteresis width in relation between the acceleration of the driven body and the rise coefficient is included as a parameter of the friction model.

Images