JPH06208404A - Automatic adjusting unit for feedback gain - Google Patents

Abstract

PURPOSE:To provide an automatic feedback gain adjusting unit capable of determining optimum feedback gain having excellent adaptiveness and giving stable motion in combinations between various kinds of motors for controlling high speed positioning and loads. CONSTITUTION:The automatic feedback gain adjusting unit 20 is provided with computing means 14, 16 for individually computing a positional deviation and torque command storing means 15, 17 for successively and individually storing the computed results of the means 14, 16, a determining means 18 for comparing the values of plural evaluation functions stored in the means 15 and determining gain reducing the values of the evaluation functions, and a torque oscillation judging means 19 for judging the existence of oscillation corresponding to the determined gain based upon the contents of the means 17. The change of gain is repeated so that the evaluation function value of a positional deviation is reduced within a range free from oscillation and the optimum gain can be finally determined.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a feedback gain automatic adjustment unit for automatically adjusting the feedback gain of a motor for high-speed positioning control.

[0002]

2. Description of the Related Art In recent years, in a motor for high-speed positioning control, digital positioning feedback control has been used instead of the conventional analog positioning feedback control, and it has become easy to change each feedback gain. Therefore, even if the motor changes over time or the load fluctuates, the gain can be adjusted to the optimum value each time by changing the gain in software.

FIG. 4 shows an example of the conventional feedback gain automatic adjustment unit. A table 1 on which components and a circuit board are placed is driven by rotation of a ball screw 3 directly connected to an electric motor 2, and the electric motor 2 is controlled by a torque control unit 4 and a position control unit 5 indicated by a dotted line. The function of automatically adjusting the feedback gain of the position control unit 5 is the feedback gain automatic adjustment unit 6 shown by the alternate long and short dash line.

The torque control unit 4 supplies a current 2a to the electric motor 2 based on the encoder pulse of the rotary encoder 7 so that the electric motor torque follows the torque command 5a from the position control unit 5.

The position control unit 5 feeds back the position data θ and the speed data ω from the torque control unit 4 so that the motor position (angle) θ follows the position command 5b, and outputs the torque command 5a to the torque control unit. To do. Torque command 5a
Is calculated by multiplying the difference Δθ between the position command 5b and the position data θ by the position feedback gain (Kp) 8 and the difference between the speed data ω and the speed feedback gain (Kvp) 9.

Feedback gain automatic adjustment unit 6
Is composed of a target value command unit 10 for automatic gain adjustment, an evaluation function calculation unit 11, an evaluation function storage unit 12, and an evaluation function minimum gain determination unit 13.

The automatic gain adjustment target value command unit 10 has a function of outputting a position command 5b to the position control unit 5 in order to drive the electric motor 2 and calculate an evaluation function during automatic gain adjustment. In the evaluation function calculation unit 11, one or more of the error square area of the position deviation Δθ, the speed ω, and the torque command 5a are output for a certain period of time after the position command is output from the target value command unit 10.
Calculate the weighted sum. The evaluation function storage unit 12 stores a pair of gain and evaluation function at that time. The evaluation function minimum gain determination unit 13 determines the gain corresponding to the minimum evaluation function among the evaluation functions stored in the evaluation function storage unit 12 as the optimum gain.

As the determining means, means for calculating an evaluation function for all combinations of gains and determining the gain from which the evaluation function is the smallest, or an evaluation function for gains starting from an arbitrary gain and in the vicinity thereof There has been a means of repeatedly changing the gain so as to reduce the evaluation function while calculating

[0009]

However, in the above-described conventional feedback gain automatic adjusting unit, the position deviation Δθ or the position θ, the speed ω, the torque are output for a certain period of time after the position command is output from the target value command unit in the evaluation function calculator. Since one or more weighted sums of the error squared areas of the command 5a are obtained, the minimum gain of the evaluation function varies depending on how the weights are determined. Therefore, the optimum gain is determined by the combination of a certain electric motor and load, but there is a problem that vibration may occur after automatic adjustment in the combination of another electric motor and load.

In view of the above-mentioned problems, the present invention determines an optimum feedback gain that provides excellent responsiveness and stable vibration-free behavior in a short time without changing the weight of the evaluation function for each combination of the motor and the load. Have the means to
An object is to provide an automatic feedback gain adjustment unit.

[0011]

A feedback gain automatic adjustment unit of the present invention has a function of outputting a position command 5b to a position control unit for actually driving a motor and calculating an evaluation function at the time of automatic gain adjustment. A function for independently calculating a position deviation evaluation function for transient evaluation and a torque command evaluation function for stable evaluation; a means for separately storing a position deviation and a torque command for a combination of gains; A function to determine the gain that minimizes the evaluation function of position deviation from the evaluation function of position deviation with respect to the gain in the vicinity of the gain, and a function to determine the presence or absence of motor vibration from the evaluation function of the torque command based on the gain thus determined , The motor is reciprocated for each gain combination, the evaluation function is calculated and stored, and the gain is not changed in the direction of decreasing the evaluation function. Et al., Front stop gain changes begin to vibrate, and determines the optimal feedback gain.

[0012]

The automatic feedback gain adjusting unit according to the present invention judges the transient evaluation by the position deviation evaluation function and the presence or absence of vibration of the electric motor, that is, the stability in the stationary state by the torque command evaluation function. Therefore, the weight of each evaluation function is not required. Therefore, no adjustment is required even when the combination of electric motor and load is changed.

[0013]

Embodiments of the present invention will be described with reference to FIGS. 1, 2 and 3.

FIG. 1 is a block diagram of the feedback gain automatic adjustment unit 20 of this embodiment including the entire structure.

A table 1 on which parts and a circuit board are placed is driven by rotation of a ball screw 3 directly connected to an electric motor 2, and the electric motor 2 has a torque control unit 4 and a position control unit 5 indicated by a dotted line.
Controlled by. The feedback gain automatic adjustment unit 20 indicated by the alternate long and short dash line automatically adjusts the feedback gain of the position control unit 5.

The torque control section 4 feeds back an encoder pulse of the rotary encoder 7 and supplies a current 2a to the electric motor 2 so that the electric motor torque follows the torque command 5a from the position control section 5.

The position control unit 5 feeds back the position data θ and the speed data ω from the torque control unit 4 so that the motor position (angle) θ follows the position command 5b, and the torque control unit 4 receives the torque command 5a. Output. Torque command 5a
Is calculated by multiplying the difference Δθ between the position command 5b and the position data θ by the position feedback gain (Kp) 8 and the difference between the speed data ω and the speed feedback gain (Kvp) 9. Except for the feedback gain automatic adjustment unit 20, it is the same as the conventional example of FIG.

Feedback gain automatic adjustment unit 2
0 is a target value command unit 10, a position deviation evaluation function calculation unit 14, a position deviation evaluation function storage unit 15, a torque command evaluation function calculation unit 16, a torque command evaluation function storage unit 17, an evaluation function minimum gain at the time of automatic gain adjustment. It is composed of a determination unit 18 and a torque vibration determination unit 19.

The automatic gain adjustment target value command unit 10 outputs a position command 5b to the position control unit in order to actually drive the electric motor 2 and calculate an evaluation function during automatic gain adjustment. The output method of the position command may be stepwise, or a position command with a trapezoidal velocity waveform to limit the acceleration may be used, but in order to more realistically adjust the gain, the actual command is used. Position command.

In the position deviation evaluation function calculation unit 14, the motor 2
In order to evaluate the transient characteristics of, the error square area of the position deviation during the transition from the output of the position command to the stable region is calculated. However, the calculation time from the start of automatic adjustment to the completion is unified. Also, in order to suppress variations in the evaluation function, 1
One round trip for each combination of two gains, that is, the sum of the evaluation functions of the return and the return.

The position deviation evaluation function storage unit 15 stores the position deviation evaluation function obtained by the position deviation evaluation function operation unit 14 and a gain as a set, and serves as a database for gain determination.

The torque command evaluation function calculation unit 16 calculates the error square area of the torque command for a fixed time after the position command is output, in order to evaluate the motor 2 in a stationary state. However, like the position deviation evaluation function, the sum of the evaluation functions for one round trip for the combination of gains is used to unify the calculation time from the start of automatic adjustment to the completion.

The torque command evaluation function storage unit 17 stores the evaluation function of the torque command obtained by the torque command evaluation function operation unit 16 and the gain as a set, and serves as a database for gain determination.

Based on the combination of the gain from the position deviation evaluation function storage unit 15 and the evaluation function, the evaluation function minimum gain determination unit 18 calculates the position deviation evaluation function from the position deviation evaluation function for the gain near the current gain. Determine the minimum gain.

The torque vibration determination unit 19 determines whether or not the motor vibration is occurring at the gain determined by the evaluation function minimum gain determination unit 18 according to the evaluation function of the torque command stored in the torque command evaluation function storage unit 17. Judge whether the limit is not exceeded.

Next, FIG. 2 shows the process of determining the optimum gain from the initial gain on a gain plane in which the horizontal axis represents position gain (Kp) 8 and the vertical axis represents velocity gain (Kvp) 9. When a point where the evaluation function of the position deviation is the minimum is plotted with respect to an arbitrary position gain (Kp) 8, the position deviation evaluation function minimum line 20 in the lower right direction is obtained.

Further, when a limiter (boundary line) that the motor 2 vibrates when the evaluation function of the torque command becomes larger than this is shown on this plane, the vibration boundary line 2 descends to the left.
It becomes 1.

The minimum line 20 of the evaluation function of position deviation and the vibration boundary line 21 always intersect on the gain plane, and if the gain at this intersection is selected, the motor 2 converges earliest without vibration.

To find this intersection, position gain (Kp)
8. Velocity gain (Kvp) 9 is divided into arbitrary step widths to form a two-dimensional array, and while varying the gains at the step widths, first find the position deviation evaluation function minimum line 20, and then on the position deviation evaluation function minimum line 20. It is possible to search by tracing the route and stopping before the vibration boundary 21. It should be noted that the position deviation evaluation function minimum line 20
The position of the vibration boundary line 21 is unknown during the automatic gain adjustment, and the intersection can be found without knowing it.

It is now assumed that the initial gain is point P2. K
The evaluation functions of the positional deviations of the three points P1, P3 and P2, which have constant p and Kvp on both sides are compared. Since the evaluation function of the position deviation at the point P3 is smaller, the gain is moved to the point P3.
It is assumed that the same operation is repeatedly reached at the point P5. At point P4, point P5, and point P6, the position deviation evaluation function becomes the minimum at the center point P5, so the gain is the position deviation evaluation function line 2
It has reached 0. At this time, if the torque command evaluation function at the point P5 is equal to or larger than the vibration limit, Kp is decreased by one and similarly repeated until a position deviation evaluation function line is found within a range where vibration does not occur. Since the vibration boundary line 21 descends to the left, the vibration region 22 can be removed.

Next, in order to reduce the position deviation evaluation function while tracing the position deviation evaluation function minimum line 20, a point P5 is used.
And the lower point P6, the lower right point P7, and the right adjacent point P8 are compared with each other, and the gain change is repeated so that the evaluation function becomes smaller. In this case, the point P
7 and repeat in the same manner until point P9. Next, even if the point P11 is selected by comparing the evaluation functions of the position deviations of the point P9 and the points P10, P11 and P12 in the vicinity thereof, the evaluation function of the torque command at the point P11 exceeds the limiter and the vibration area 22 is displayed. Since it is included, the optimum gain is determined as point P9. However, when the evaluation function of the torque command does not exceed the limiter and the evaluation function of the position deviation becomes the minimum on the minimum line 20 of the position deviation evaluation function, the point is set as the optimum gain.

Next, FIG. 3 shows a flowchart of a process for determining the optimum gain. The position gain (Kp) 8 and the velocity gain (Kvp) 9 are handled as an array, and are divided by a constant step width and are numbered.

First, in # 1, the initial gain is obtained and the gain
No. is assigned to (i, j), and in # 2, Kvp No. is 3 points including points (i, j-1) and (i, j + 1) on both sides,
Position deviation evaluation function JΔθ and torque command evaluation function J _TL
To measure. When the measurement result already exists in the storage unit, the data is used as the measurement result. Next, in # 3, it is checked whether or not JΔθ at the center of the three points becomes the minimum and the gain is on the position deviation evaluation function minimum line 20. If JΔθ at the center of 3 points is not the minimum, proceed to # 4 and JΔθ out of 3 points
The gain is moved to the point where is minimum, and the gain No. is assigned to (i, j). After that, the process proceeds to # 2 and repeats until JΔθ at the center of the three points becomes minimum, and when it becomes minimum, #
Go to 5. # 5 in the investigation whether JT _L in the current gain does not exceed the limiter of J _TL, if over Kp
Decrease No. 1 by one, return to # 2 and repeat until J _TL does not exceed the limiter. If J _TL does not exceed the limiter, proceed to # 7.

In # 7, the current gain No. (i, j),
(I + 1, j), (i, j + 1), (i + 1, j + 1)
Measure JΔθ and J _TL at 4 points. When the measurement result already exists in the position deviation evaluation function storage unit 15 and the torque command evaluation function storage unit 17, the data is used as the measurement result. Next, in # 8 and # 9, the gain No. is set to J of 3 points other than (i, j).
_If all of _TL are smaller than the limiter of J _TL and JΔθ at (i, j) is not the minimum of 4 points, proceed to # 10 and J out of 4 points.
Substitute the gain No. with the smallest Δθ for (i, j), and
Return to 7. Otherwise, in # 11, the gain of the current gain No. is determined as the optimum gain and the process ends.

By the above method, the evaluation function of the position deviation can be made as small as possible within the range where vibration does not occur, and the feedback gain is automatically adjusted and the feedback gain is automatically adjusted without vibration during the stationary state.

The constants used in this method are the initial gain and J
_Although it is a _TL limiter, the initial gain can be anywhere within a range where the electric motor 2 moves reasonably, and the J _TL limiter changes the load a little because the J _TL is greatly different when the electric motor 2 is vibrating and when it is not vibrating. However, it may be fixed.

[0037]

According to the automatic feedback gain adjusting unit of the motor for high-speed positioning control of the present invention, the responsiveness at the time of transient is separated from the evaluation function of the position deviation, and the vibration at the stable is separated from the evaluation function of the torque command. There is no weight of the evaluation function that is commonly used. Therefore, the adjustment of weight is no longer an issue. Also, any Kp on the gain plane
Since the gain is changed through the line where the position deviation evaluation function is the minimum, even if the line is changed due to changes in the motor or load, if the automatic adjustment is performed again, the position deviation will be changed along that line. The optimum gain is determined by moving the gain in the direction in which the evaluation function of becomes smaller and selecting the gain before the vibration.

Therefore, even if the type or load of the electric motor changes, the feedback gain can be automatically adjusted in a short time without changing the constant for automatic gain adjustment.

[Brief description of drawings]

FIG. 1 is a block diagram of a feedback gain automatic adjustment unit according to an embodiment of the present invention.

FIG. 2 is a schematic explanatory diagram on a gain plane regarding determination of an optimum gain according to the present embodiment.

FIG. 3 is a flowchart diagram for determining an optimum gain according to the present embodiment.

FIG. 4 is a block diagram of a conventional example.

[Explanation of symbols]

 2 electric motor 4 torque control means 5 position control means 10 target value command means during automatic gain adjustment 14 position deviation evaluation function calculation means 15 position deviation evaluation function storage means 16 torque command evaluation function calculation means 17 torque command evaluation function storage means 18 evaluation function Minimum gain determination means 19 Torque vibration determination means 20 Feedback gain automatic adjustment unit

Claims (1)

[Claims] 1. In a feedback gain automatic adjustment unit of a motor for high-speed positioning control, target value command means for commanding a target value during automatic gain adjustment, and a position deviation evaluation function for calculating an error square area of position deviation as an evaluation function. A calculation means; a position deviation evaluation function storage means for sequentially storing the values of the evaluation functions calculated by the position deviation evaluation function calculation function; and a plurality of position deviation evaluation function values stored in the position deviation evaluation function. Gain determining means for comparing and determining the smaller gain of the evaluation function value, torque command evaluation function calculating means for calculating the error squared area of the torque command when having a torque command type driver as an evaluation function, and the torque command Torque command evaluation function storage means for sequentially storing the values of the evaluation function calculated by the evaluation function calculation function, and determined by the gain determination function The automatic feedback gain adjustment unit, further comprising: a torque vibration determination unit that determines whether or not there is vibration of the electric motor based on the evaluation function of the torque command stored in the torque command evaluation function storage function for the obtained gain.