EPE Association: EPE 1997 – 58: Dialogue Session DS4i: IDENTIFICATION, ESTIMATION AND ADAPTIVE CONTROL

EPE 1997 – 58: Dialogue Session DS4i: IDENTIFICATION, ESTIMATION AND ADAPTIVE CONTROL
You are here: EPE Documents > 01 - EPE & EPE ECCE Conference Proceedings > EPE 1997 - Conference > EPE 1997 – 58: Dialogue Session DS4i: IDENTIFICATION, ESTIMATION AND ADAPTIVE CONTROL

   [return to parent folder]

   Fast estimation of unknown resonant frequencies by means of the VeCon chip set
By Ch. Hattrup; M. Oszmann; H. van der Broeck [View]
[Download]

Abstract: A fast control algorithm is presented which estimates the unknown resonant frequency of a series resonant converter and sets the inverter frequency to that value. The procedure is based on the successive measurement of the converter current at certain time instants within the operation period. The algorithm has been implemented in the VeCon chip set and successfully proved experimentally.

   STRUCTURE- AND PARAMETER IDENTIFICATION FOR A TWO-MASS SYSTEM WITH BACKLASH AND FRICTION USING A SELF-ORGANIZING MAP
By F. Schütte; S.Beineke; H. Grotstollen; N. Frohleke; U. Witkowski; U. Rückert; S. Rüping [View]
[Download]

Abstract: A self-commissioning system for high performance speed and position control of electrical drives requires a structure and parameter identification of a nonlinear mechanic as basic building block. This self-commissioning system in combination with a new identification scheme is presented here. The identification is based on extraction of characteristic features from the system response and evaluation of these features by self-organizing neural networks, especially the Self-Organizing feature Map (SOM).

   COMPARISON OF METHODS FOR STATE ESTIMATION AND ON-LINE IDENTIFICATION IN SPEED AND POSITION CONTROL LOOPS
By S. Beineke; F. Schütte; H. Grotstollen [View]
[Download]

Abstract: For high performance speed and position control of drives featuring elasticity observation of inner states becomes necessary, if only motor speed and current can be measured. In the presence of time-varying inertia or friction one line identification of these parameters is also needed. In this paper different approaches for these tasks are compared which base on extended Kalman filters, extended Luenberger observers, basis function networks and their combinations.

   A PRACTICAL IDENTIFICATION SCHEME FOR INDUCTION MOTORS AT STANDSTILL USING ONLY A VS INVERTER AS THE ACTUATOR
By J. Godbersen; P. Thogersen; M. Tonnes [View]
[Download]

Abstract: A method to identify the equivalent circuit parameter values of an induction motor is described. The method takes into account the nonlinear effects of the motor and the inverter, and yields the parameter values at the rated magnetic flux level. The estimations are made at zero speed, i.e. the motor produces no rotational shaft torque during the test sequence. Furthermore the digital computer calculations are held at a level that can be implemented in the control software of standard frequency converters without using an external computer for the commissioning of the drive system.

   FULL-ORDER OBSERVERS FOR FLUX AND PARAMETER ESTIMATION OF INDUCTION MOTORS
By Lennart Harnefors; Hans-Peter Nee [View]
[Download]

Abstract: Design of directly parametrized full-order observers for induction motors is considered. Physics-based observers having properties of the "current" and "voltage" models at low and nominal stator frequencies, respectively, have been designed previously by other researchers. This paper is believed to be the first, not only to show that the same properties can be achieved with a classical (Gopinath) full-order observer, but also to present the proper analytical observer gain selection. Good damping and fast convergence is achieved as well. The proposed observer is compared to the Kalman filter (KF) and it is shown that the dynamic properties of a sensibly designed KF resemble somewhat those of the observer, while the KF has much worse flux estimation properties. The costly extra computations needed to implement the KF are therefore not likely worthwhile.

   REAL-TIME PARAMETER ESTIMATION AND ADAPTIVE CONTROL OF AN IPM MOTOR DRIVE
By R.C. Panaitescu; L.E. Nomm [View]
[Download]

Abstract: In this paper a least-square algorithm for parameter identification and an adaptive control strategy for a Interior Permanent Magnet (IPM) Synchronous Motor drive is presented. A method for estimating the mechanical parameters in an IPM motor was necessary for the Kalman filter based sensorless control method. presented in another paper. In order to maintain a good accuracy in speed and angle estin1ation. the filter must be redesigned dynamically whenever the parameters change. The magnet flux linkage can be determined only in the presence of a good estimation for the load torque (done within the Kalman filter sensorless control). Mechanical parameters are accurately estimated when the condition of persistent excitation is fulfilled.

EPE 1997 – 58: Dialogue Session DS4i: IDENTIFICATION, ESTIMATION AND ADAPTIVE CONTROL
You are here: EPE Documents > 01 - EPE & EPE ECCE Conference Proceedings > EPE 1997 - Conference > EPE 1997 – 58: Dialogue Session DS4i: IDENTIFICATION, ESTIMATION AND ADAPTIVE CONTROL
	[return to parent folder]

	Fast estimation of unknown resonant frequencies by means of the VeCon chip set By Ch. Hattrup; M. Oszmann; H. van der Broeck	[View] [Download]
Abstract: A fast control algorithm is presented which estimates the unknown resonant frequency of a series resonant converter and sets the inverter frequency to that value. The procedure is based on the successive measurement of the converter current at certain time instants within the operation period. The algorithm has been implemented in the VeCon chip set and successfully proved experimentally.

	STRUCTURE- AND PARAMETER IDENTIFICATION FOR A TWO-MASS SYSTEM WITH BACKLASH AND FRICTION USING A SELF-ORGANIZING MAP By F. Schütte; S.Beineke; H. Grotstollen; N. Frohleke; U. Witkowski; U. Rückert; S. Rüping	[View] [Download]
Abstract: A self-commissioning system for high performance speed and position control of electrical drives requires a structure and parameter identification of a nonlinear mechanic as basic building block. This self-commissioning system in combination with a new identification scheme is presented here. The identification is based on extraction of characteristic features from the system response and evaluation of these features by self-organizing neural networks, especially the Self-Organizing feature Map (SOM).

	COMPARISON OF METHODS FOR STATE ESTIMATION AND ON-LINE IDENTIFICATION IN SPEED AND POSITION CONTROL LOOPS By S. Beineke; F. Schütte; H. Grotstollen	[View] [Download]
Abstract: For high performance speed and position control of drives featuring elasticity observation of inner states becomes necessary, if only motor speed and current can be measured. In the presence of time-varying inertia or friction one line identification of these parameters is also needed. In this paper different approaches for these tasks are compared which base on extended Kalman filters, extended Luenberger observers, basis function networks and their combinations.

	A PRACTICAL IDENTIFICATION SCHEME FOR INDUCTION MOTORS AT STANDSTILL USING ONLY A VS INVERTER AS THE ACTUATOR By J. Godbersen; P. Thogersen; M. Tonnes	[View] [Download]
Abstract: A method to identify the equivalent circuit parameter values of an induction motor is described. The method takes into account the nonlinear effects of the motor and the inverter, and yields the parameter values at the rated magnetic flux level. The estimations are made at zero speed, i.e. the motor produces no rotational shaft torque during the test sequence. Furthermore the digital computer calculations are held at a level that can be implemented in the control software of standard frequency converters without using an external computer for the commissioning of the drive system.

	FULL-ORDER OBSERVERS FOR FLUX AND PARAMETER ESTIMATION OF INDUCTION MOTORS By Lennart Harnefors; Hans-Peter Nee	[View] [Download]
Abstract: Design of directly parametrized full-order observers for induction motors is considered. Physics-based observers having properties of the "current" and "voltage" models at low and nominal stator frequencies, respectively, have been designed previously by other researchers. This paper is believed to be the first, not only to show that the same properties can be achieved with a classical (Gopinath) full-order observer, but also to present the proper analytical observer gain selection. Good damping and fast convergence is achieved as well. The proposed observer is compared to the Kalman filter (KF) and it is shown that the dynamic properties of a sensibly designed KF resemble somewhat those of the observer, while the KF has much worse flux estimation properties. The costly extra computations needed to implement the KF are therefore not likely worthwhile.

	REAL-TIME PARAMETER ESTIMATION AND ADAPTIVE CONTROL OF AN IPM MOTOR DRIVE By R.C. Panaitescu; L.E. Nomm	[View] [Download]
Abstract: In this paper a least-square algorithm for parameter identification and an adaptive control strategy for a Interior Permanent Magnet (IPM) Synchronous Motor drive is presented. A method for estimating the mechanical parameters in an IPM motor was necessary for the Kalman filter based sensorless control method. presented in another paper. In order to maintain a good accuracy in speed and angle estin1ation. the filter must be redesigned dynamically whenever the parameters change. The magnet flux linkage can be determined only in the presence of a good estimation for the load torque (done within the Kalman filter sensorless control). Mechanical parameters are accurately estimated when the condition of persistent excitation is fulfilled.