JPH1189267A - Control constant adjusting device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a control constant adjusting device the identification error of which does not become larger, even when the speed deviation of a motor is almost zero as in the case where the motor rotates at a constant rotating speed. SOLUTION: A control constant adjusting device is provided with a speed control section 12 which controls the rotating speed of a motor; an estimating section 13 which simulates the control section 12; an identifying section 14 which identifies an inertia from the ratio of a value SFVe which is obtained by performing time integration on an absolute value obtained by passing the speed deviation of the control section 12 through a prescribed filter in a prescribed section to another value SVe', which is obtained by performing time integration on the absolute value of the speed deviation of the estimating section 13 in the same section, and an adjusting section 15 which adjusts a control gain based on the found inertia, so that the device may perform inertia identification operation obtained through the time integration, under such a condition that the absolute value of the speed deviation in the control section 12 has a prescribed value α or larger and the speed in the estimating section 13 is not zero.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a control device for a robot, a machine tool, etc., and more particularly to a control device for adjusting a control constant when inertia varies during operation.

[0002]

2. Description of the Related Art As a conventional device for adjusting a control constant,
For example, Japanese Patent Application No. 8-230713 related to the present applicant,
There is a “motor control device”. The motor control device determines a torque command so that the input speed command matches the actual motor speed and controls the motor speed, and the motor control device controls the motor speed so that the model speed matches the motor speed. An estimating unit that simulates a speed control unit; and an identification unit that identifies inertia based on a ratio of a value obtained by time-integrating a speed deviation of the speed control unit in a predetermined interval, and a motor speed in the speed control unit. And the inertia identification is performed only when the speed of the model in the estimating unit coincides with a non-zero value. The identification of the inertia is a value obtained by time-integrating the absolute value of the speed deviation of the speed control unit in a predetermined section. And a value obtained by time-integrating the absolute value of the speed deviation of the estimating unit in the same section. Therefore, in the motor control device of the prior application, the speed deviation Ve of the speed control unit is
Since the magnitude of the load inertia is identified based on the time integral value of, and the gain of the feedback control loop is automatically adjusted based on the identified value, optimal operation is guaranteed even if the load inertia changes. Has advantages. FIG. 5 is a diagram for explaining the details of the operation of the identification unit of the conventional motor control device, and the switching of the integrator is performed as shown in FIG. That is, from time t0 to t1 (section 1), the inertia J is calculated using MODE0 in the upper part of FIG. 5, and at time t1, the operation moves to MODE1 in the lower part of FIG.
The values | SFVe | of the integrator used at 0 and | SVe '| are cleared to zero, and the integration is started again. MODE1 from time t1 to t2 (section 2)
To calculate the inertia J at time t2
0 and at the same time the value of the integrator | S used in MODE1
FVe | and | SVe '| are cleared to zero and integration is started again. Shifting the section (section 1 → section 2 → section 3 → section 4 → section 5 →...) While repeating these series of operations enables identification of inertia responsive to the speed deviation affected by load fluctuation. Thus, the control gain can be adjusted. The inertia identification calculation in this case is performed under the condition that “only when the speed Vfb in the speed control unit and the speed Vfb ′ in the estimation unit match with a non-zero value” as described above.
The identification unit is provided with two integrators, each of which integrates the value FVe obtained by filtering the speed deviation Ve in the speed control unit and the absolute value of the speed deviation Ve 'in the estimation unit, and integrates them between predetermined sections [a, b]. The inertia J is calculated by calculating the ratio of each integral value.

[0003]

However, in the above-mentioned prior art, the magnitude of the load inertia is identified only on condition that the speed Vfb in the speed control unit and the speed Vfb 'in the estimation unit match with a non-zero value. Therefore, the identification calculation is performed even when the speed deviation is almost zero, such as when the motor is rotating at a constant rotation speed, and there is a problem that the identification error increases.
Further, in the prior art, since the integrator is switched at predetermined fixed time intervals, the integrator overflows,
There is a problem that the value of the integrator is too small and the identification error increases. Accordingly, the present invention prevents the identification error from increasing without performing the identification calculation when the speed deviation is almost zero, such as when the motor is rotating at a constant rotation speed, and has a large speed deviation. To provide a control constant adjustment device that does not cause an overflow and that can accurately identify inertia without clearing the integrator even when the speed deviation is almost zero, and that can perform tuning in real time with high accuracy. The purpose is.

[0004]

In order to achieve the above object, according to the first aspect of the present invention, a torque command is determined so that an input speed command matches an actual motor speed, and a speed for controlling the motor speed is determined. A control unit, an estimating unit that simulates the speed control unit so that the speed of the model matches the motor speed, and an absolute value of a value obtained by passing a predetermined filter to a speed deviation of the speed control unit in a predetermined interval. An identification unit that performs an inertia identification calculation from a ratio of a value obtained by time integration of the value in time and an absolute value of the speed deviation of the estimation unit in the same section, and an inertia identified in the identification unit and an inertia in the estimation unit. An adjustment unit that adjusts a control gain based on a value obtained by passing a predetermined filter to the ratio of the control unit. The method is performed under the condition that the absolute value of the speed deviation of the section is equal to or larger than a predetermined value α and the speed in the estimating section is not zero. With such a configuration, the condition that the inertia identification calculation is performed when the absolute value of the speed deviation in the speed control unit is equal to or larger than the predetermined value α and the speed in the estimation unit is not zero is imposed. When the speed deviation is almost zero, such as when the motor is rotating at a constant speed, the identification calculation is not performed, and the identification error can be prevented from increasing. The invention according to claim 2 is characterized in that the value of the predetermined value α is specifically an integer multiple of the minimum resolution of the speed deviation. According to this configuration, it is possible to set a condition that enables accurate inertia identification calculation without increasing the identification error. According to a third aspect of the present invention, the identification unit specifically integrates the absolute value of the value obtained by passing the speed deviation of the speed control unit through a predetermined filter in a predetermined section over time. And a second integrator for time-integrating the absolute value of the speed deviation of the estimating unit in the same section. The first and second integrators perform integration under the same conditions. One of the integrators has a value β / 2 that is half of the upper limit value β and the other integrator has a zero value. When the value of the one integrator reaches the upper limit value β, the integration is cleared to zero and the integration used for the identification operation is performed. The integrator alternately switches from one to the other between the first and second integrators. According to this configuration, switching of each integrator is performed not by setting a fixed time but by setting an upper limit value β. Therefore, the value of the integrator used for the identification calculation always takes a value between β / 2 and β, and the speed deviation is large. The value does not cause overflow,
Even when the speed deviation is almost zero, it is possible to prevent the identification error from becoming large, and it is possible to accurately identify the inertia.

[0005]

Embodiments of the present invention will be described below with reference to the drawings. FIG. 1 is a block diagram of a control constant adjusting device according to an embodiment of the present invention. FIG. 2 shows FIG.
3 is a detailed block diagram of each unit shown in FIG. FIG. 3 is a diagram illustrating details of the identification unit shown in FIG. FIG. 4 is a diagram showing an example of a simulation response by the control constant adjusting device shown in FIG. In FIG. 1, reference numeral 11 denotes a command generation unit, which outputs a speed command Vref to a speed control unit 12. The speed control unit 12 outputs the motor speed Vfb in the speed control unit to the estimation unit 13 and outputs the speed deviation Ve to the identification unit 14. The identification unit 14 receives the speed deviation Ve of the speed control unit 12 and the speed deviation Ve ′ of the estimation unit 13 to obtain the inertia J, and outputs the inertia ratio J / J ′ to the adjustment unit 15. The adjusting unit 15 receives the inertia ratio J / J ′, determines the proportional gain Kv and the integral gain Ki in the speed controller 12 based on the values passed through a predetermined filter, and determines the value of the integrator in the speed controller 12. Adjust Next, details of each unit will be described with reference to FIG. The speed control unit 12 forms a speed loop such that the speed Vfb matches the speed command Vref from the command generation unit 11, and here, as a PI (proportional-integral) control, a current controller that drives a motor by using a torque command Tref. , The speed Vfb is output to the estimation unit 13, and the speed deviation Ve is output to the identification unit 14. It is assumed that a load JL is attached to the motor, and the speed Vfb is output from the motor. The estimating unit 13 uses the speed Vfb in the speed control unit 12 as a command and
The control target is modeled by 1 / j's, the torque command Tref is output to the control target 1 / J's, and the speed deviation Ve 'is output to the identification unit 14. The identification unit 14 includes a first integrator 141, a second integrator 142, and an inertia calculation unit 143. The first integrator 141 receives the speed deviation Ve output from the speed control unit 12, and adds the speed command Vref of the speed control unit 12 to the speed deviation Ve.
, The absolute value of the value FVe obtained through a filter using a transfer function as a model up to the speed Vfb is obtained, and a predetermined interval [a,
b] to perform time integration, and obtain a time integrated value | SFVe |. The second integrator 142 calculates the absolute value of the speed deviation Ve 'output from the estimating unit 13, performs time integration in a predetermined section [a, b], and obtains a time integration value | SVe' |
Ask for. The inertia calculation unit 143 calculates the time integrated value | SFVe | and the time integrated value | SV
From the ratio to e ′ | and the inertia J ′ of the estimating unit 13, the inertia J of the speed control unit 12 is obtained from the following equation (1). J = (| SFVe | / | SVe '|) × J' (1) However, the identification calculation of the expression (1) is performed when the absolute value | Ve | of the speed deviation Ve of the speed control unit 12 is obtained. Has a magnitude equal to or greater than the predetermined value α, and the speed Vfb ′ in the estimation unit 13 is not zero;
It is conditional. Here, the inertia identification principle is briefly mentioned.
To accurately determine the inertia from the torque command or the time integral of the motor current, the transfer function from the torque command or the motor current to the speed is expressed only by the inertia. If the speed is not zero, the time integral value of the torque command or the motor current The inertia can be easily obtained from the ratio between the speed and the speed. Utilizing this relationship, the same speed command is input to the actual speed controller and its model, and under the condition that the actual speed matches the model speed with a non-zero value, each torque command or motor current In the conventional method, the inertia was obtained from the time integral value and the speed of the motor. However, in this case, since the torque command or the motor current includes a friction compensation component and a disturbance compensation component in addition to the command response component, in order to eliminate these influences, the speed deviation is time-integrated instead of the torque command or the motor current. The actual speed is used as a model speed command so that the condition that the actual speed matches the model speed with a non-zero value can be satisfied as much as possible.
However, if only the actual speed of the conventional method and the model speed match at a value other than zero, identification calculation is performed even when the speed deviation is almost zero, such as when the motor is rotating at a constant speed. This causes a problem that the identification error becomes large. Therefore, in the present embodiment, it is assumed that the actual speed and the model speed match,
The condition is improved so that the inertia J is identified only when the absolute value | Ve | of the speed deviation Ve has a value equal to or larger than the predetermined value α and the speed Vfb ′ in the estimation unit 13 is not zero. Therefore, when the speed deviation is almost zero, such as when the motor is rotating at a constant speed, the identification calculation is not performed, and the identification error does not increase. The predetermined value α is set to an optimal value that is an integral multiple of the minimum resolution of the speed deviation. Next, a method of setting the predetermined integration section [a, b] will be described. Speed deviation Ve in the speed control unit 12
Is integrated with two sets of integrators 141 and 142 that respectively integrate the value FVe obtained by filtering and the absolute value of the speed deviation Ve ′ in the estimating unit 13. Conventionally, as shown in FIG. The inertia was identified by switching the integration of MODE1. However, in the integration of the conventional example, the switching of the integrator is performed from the section 1 to the section 2 (t1 to t2),
Section 3 (t2 to t3), Section 4 (t3 to t4),.
Since the speed deviation Ve takes a large value, an overflow occurs. Since the speed integrator is cleared when the speed deviation Ve is almost zero, the identification value is also zero when the integrator is cleared. In this embodiment, the inconvenience has been made as shown in FIG. The integrator shown in FIG.
The DE0 integrator has an initial value β / 2, which is half the upper limit value β. Until the value of the integrator of MODE0 reaches the upper limit value β (section 1), the operation of Expression (1) is performed using MODE0, and at the same time the value of the integrator of MODE0 becomes the upper limit value β, MODE1 in the lower part of FIG. Then, the values of the integrators | SFVe | and | SVe '| used in MODE0 are cleared to zero, and the integration is started again. After moving to MODE1 (section 2), the operation of equation (1) is performed in MODE1, and
When the value of the integrator of DE1 reaches the upper limit value β, MODE again
0 and at the same time the value of the integrator | S used in MODE1
FVe | and | SVe '| are cleared to zero and integration is started again. Shift the section while repeating these series of operations (section 1 → section 2 → section 3 → section 4 → section 5 → ...
By performing (), overflow does not occur even when the speed deviation Ve takes a large value, and even when the speed deviation Ve is almost zero, the integrator is not cleared and the inertia can be identified with high accuracy. When the section [a, b] in which the speed deviation is integrated with respect to time is described using the section 3 in FIG. 3 as an example, time a corresponds to time t1, and time b corresponds to each time from time t2 to t3. . Based on the value of the inertia ratio (J / J ′) obtained in the identification unit 14 thus passed through a predetermined filter, the adjustment unit 15 controls the speed controller 1
2, and updates the value of the integrator in the speed controller 12 while updating the proportional gain Kv and the integral gain Ki. Next, an example of response by simulation is shown in FIG. FIG.
FIG. 4A is a response example without using the present invention for comparison, and FIG. 4B is a response example using the present invention. FIG.
Then, position control is performed, and in the figure are as follows. Is a position command Pref, which is a positioning command at a constant speed of 500 rpm. Is the motor position Pfb, is the motor speed Vfb, is the speed deviation Ve, is the total load inertia JN set in the simulation.
(JN = JM: motor + JL: load) and the ratio JN / J 'of the inertia J' in the estimating unit 13, where is the inertia J identified in the identifying unit 13 and the estimating unit 1
3 is a value obtained by passing a primary filter having a time constant of 100 ms in the adjustment unit 15 to the ratio J / J 'of the inertia J' in 3 and is displayed for each control cycle. Here, the simulation is performed with J ′ = JM, and Vfbｆ
0 and when the condition of Ve> α is not satisfied (1)
The calculation of the equation was not performed and the identification value in the previous control cycle was used. or,
The adjusting unit 15 changes the proportional gain Kv in the speed control unit according to the inertia ratio J / J ′ by Kv = Kv ×
Is updated by the adjustment unit at each control cycle, and the integral gain Ki is set to a value suitable for the proportional gain Kv. As is apparent from FIG. 4, in FIG. 4A, which is an example of a response in the case where the present invention is not used, the diagram representing the motor speed Vfb rises from 0 to 500 rpm and then continues the constant speed motion. Since the identification unit 13 performs the identification operation during the entire period, an identification error occurs as described above. On the other hand, in the response example 4 (b) using the present invention, since the identification unit 13 does not perform the identification calculation during the motor speed Vfb during the constant speed motion period,
No identification error occurs. As described above, according to the present embodiment, when the speed deviation is almost zero, such as when the motor is rotating at a constant speed, the identification calculation is not performed, so that the identification error increases. And a stable response can be obtained.

[0006]

As described above, according to the present invention, when the speed deviation is almost zero as in the case where the motor is rotating at a constant speed, the identification calculation is not performed, and the identification error is large. However, since the integrator is switched by setting the upper limit value, no overflow occurs even if the speed deviation Ve takes a large value, and even if the speed deviation Ve is almost zero, the integrator is not cleared and the inertia is accurately removed. Since the identification can be performed, it is possible to provide a control constant adjustment device that can accurately perform tuning in real time.

[Brief description of the drawings]

FIG. 1 is a block diagram of a control constant adjusting device according to an embodiment of the present invention.

FIG. 2 is a detailed block diagram of each unit shown in FIG.

FIG. 3 is a diagram illustrating details of an identification unit shown in FIG. 2;

FIG. 4 is a diagram showing an example of a simulation response by the control constant adjusting device shown in FIG. 1;

FIG. 5 is a diagram illustrating details of an identification unit of a conventional control constant adjustment device.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 11 Command generation part 12 Speed control part 13 Estimation part 14 Identification part 141 1st integrator 142 2nd integrator 143 Inertia calculation part 15 Adjustment part

Claims (3)

[Claims] A speed controller that determines a torque command so that an input speed command matches an actual motor speed and controls the motor speed; and a speed controller that controls a model speed to match the motor speed. An estimating unit that simulates a unit, and a value obtained by time-integrating an absolute value of a value obtained by passing a predetermined filter to a speed deviation of the speed control unit in a predetermined section and an absolute value of the speed deviation of the estimating unit in the same section. An identification unit that performs an inertia identification operation from a ratio of the time-integrated value, and an adjustment unit that adjusts a control gain based on a value obtained by passing a ratio of the inertia identified in the identification unit and the inertia in the estimation unit through a predetermined filter. An adjustment unit, comprising: a control constant adjustment device, comprising: an inertia identification operation performed by the identification unit, wherein the absolute value of the speed deviation of the speed control unit has a magnitude equal to or larger than a predetermined value α; The control constant adjusting device is operated under the condition that the speed in the estimating unit is not zero. 2. The control constant adjusting device according to claim 1, wherein the predetermined value α is an integral multiple of a speed deviation minimum resolution. 3. The control constant adjusting device, wherein the identification unit includes a first integrator that time-integrates an absolute value of a value obtained by passing a speed filter of the speed control unit through a predetermined filter in a predetermined interval, and A second integrator that integrates the absolute value of the speed deviation of the estimator with respect to time in the same section, wherein the first and second integrators perform integration under the same conditions. Has a value β / 2 that is half of the upper limit value β and the other integrator has zero. When the value of the one integrator reaches the upper limit value β, the integrator is cleared to zero and the integrator used for the identification operation is 2. The operation of switching from one to the other is alternately repeated between the first and second integrators.
The control constant adjusting device described in the above.