Electricity: motive power systems – Positional servo systems – 'reset' systems
Reexamination Certificate
1999-11-29
2001-05-22
Nappi, Robert E. (Department: 2837)
Electricity: motive power systems
Positional servo systems
'reset' systems
C700S037000
Reexamination Certificate
active
06236182
ABSTRACT:
The present invention relates to a method for determining the optimum gain for the integral-action component of a closed-loop or automatic speed control.
DESCRIPTION OF RELATED ART
U.S. Pat. No. 5,157,597, describes a method to detect oscillations in servo system operation and to modify the loop gain of closed loops. In the process, a characteristic ratio is fixed between the position-control-loop gain and the position deviation in such a way that the loop gain assumes a large value for certain position deviation values. The position deviation is periodically determined during operation, and the main oscillation component of the position deviation is determined. The loop gain of the servo system is adjusted in such a way that the main oscillation component falls within a specified frequency range.
In this scheme, however, although the frequency range in which the servo system begins to oscillate is adjusted, the tendency to oscillate is not diminished. Thus, controller parameters are not optimized, and the adjusted frequency range can shift under different load conditions, which can lead to damaging oscillation in the controlled subassemblies.
U.S. Pat. No. 4,549,123 describes a method for tuning a controller. A nonlinear subassembly, whose output signal has either a constant negative or positive amplitude, is introduced in series to and before the PID controller. The controller parameters are subsequently modified in such a way that the transfer function of the nonlinear subassembly is multiplied by the common transfer function of the controller and the controlled system, to yield the value −1 for a specific amplitude and a specific frequency. The modified control loop can then be brought into self oscillation. Various parameters are then calculated using the Ziegter and Nichols formulas, and the controller is tuned as a function of the calculated parameters.
A drawback of this system is that it requires that the entire controller structure be altered. No provision is made for simply modifying the controller parameters.
EP Patent 347 465 B1 describes a servo motor that is more likely to oscillate at low rotational speeds than at high rotational speeds. The servo motor is provided with a control loop, whose gain in the P-and I-branch of the controller is adjusted as a function of the speed. As a result, at low rotational speeds, the servo motor does not begin to oscillate and, at high rotational speeds, the torque is not unnecessarily limited.
This design has the disadvantage that optimal starting values must already be present in order to have a stable operation. These starting values are then merely adapted to a modified electro-motor speed. The reference does not explain how the starting values are determined.
Optimal controller parameters of a PI-controller in a speed control loop can be automatically determined according to non-prepublished Patent Application DE 197 34 208.6 of the applicant. To adjust a gain in the proportional branch of the controller, a first test signal is fed to the automatically controlled electro-motor, and the resultant speed characteristic is determined as a function of time in a first measuring interval. This speed characteristic is analyzed using a specific weighting function. Using a small gain in the proportional and integral branch of the controller as a baseline, the gain in the proportional branch is continually increased, until a beginning oscillation of the control loop is detected using the weighting function.
The adjusted gain in the proportional branch is then multiplied by a factor of less than one, more particularly of about 0.45, and is adjusted. A second test signal is then fed as a reference input variable to the control loop, to adjust the gain in the integral branch of the controller. The time characteristic of the electro-motor's rotational speed is determined once more in a measuring interval and weighted using a second weighting function. A deviation from the set-point rotational speed is then determined in the measuring interval using an error function. Using as a baseline a gain in the integral branch at which oscillations are present, the gain is progressively reduced until a minimum of the error function is reached. The gain, determined in this manner, is then adjusted in the integral branch.
The disadvantage of this scheme is that when the gain in the integral branch is adjusted, oscillations experienced by the mechanical components coupled to the electro-motor do not decay quickly in the measuring interval. As a result, an excessively low gain exists on a regular basis in the integral branch. This unnecessarily degrades the dynamic response of the entire system that includes the control loop and the machine components.
SUMMARY OF THE INVENTION
The present invention in one aspect is a method for automatically determining the gain in the integral branch of the controller of a speed control loop. The computed gain reliably prevents the entire driving assembly from oscillating and, at the same time, precisely controls the required electro-motor torque being made available, so that the set-point rotational speed is reached quickly. The gain is determined in a way that allows for influences caused both by the electro-motor, as well as by other subassemblies of the driving assembly. Moreover, the method can be implemented simply and cost-effectively.
The method of the present invention allows additional integral-action time of the controller when determining the gain in the integral branch of the controller. The gain in the integral branch of the controller is initially determined in accordance with the method according to applicant's DE 197 34 208.6, incorporated herein by reference in its entirety, and subsequently is selected in proportion to the integral-action time of the controller. This is achieved by multiplying the gain for the integral branch, determined as explained in DE 197 34 208.6, by the integral-action time.
REFERENCES:
patent: 4109190 (1978-08-01), McNaughton
patent: 4549123 (1985-10-01), Hagglund et al.
patent: 4754391 (1988-06-01), Suzuki
patent: 5063335 (1991-11-01), Baechtel et al.
patent: 5157597 (1992-10-01), Iwashita
patent: 5223778 (1993-06-01), Svarovsky et al.
patent: 5263113 (1993-11-01), Naitoh et al.
patent: 5475291 (1995-12-01), Yoshida et al.
patent: 5587896 (1996-12-01), Hansen et al.
patent: 5684375 (1997-11-01), Chaffee et al.
patent: 5736824 (1998-04-01), Sato et al.
patent: 5973467 (1999-10-01), Eguchi
patent: 6127793 (2000-10-01), Kerner
patent: 36 18 337 (1990-06-01), None
patent: 43 33 146 (1994-03-01), None
patent: 44 42 834 (1995-06-01), None
patent: 196 17 867 (1997-11-01), None
patent: 197 34 208 (1999-02-01), None
patent: 197 48 993 (1999-05-01), None
patent: 0 347 465 (1989-12-01), None
Dr. Johannes Heidenhain GmbH
Duda Rina I.
Kenyon & Kenyon
Nappi Robert E.
LandOfFree
Method and circuit arrangement for determining an optimum... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Method and circuit arrangement for determining an optimum..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method and circuit arrangement for determining an optimum... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2494607