In this paper a sensorless control strategy has been implemented and experimentally validated to control independently the electromagnetic torque and the reactive power of a doubly-fed induction machine. The machine acts as a grid connected generator at variable speed operations around the synchronous speed. This control strategy is based on a model-predictive control strategy with the governing equations represented in a synchronous reference frame aligned to the stator flux. A modulation stage has been introduced in order to overcome the well-known issues of the model-predictive control strategies, such as high current ripple and the non-constant switching frequency. The stator flux vector has been identified by using a programmable low-pass filter achieving an acceptable estimate which is not affected by the integer offset derived problems, and it does not require high number of calculations as by using a full-order observer. The rotor position feedback signal needed to implement the control strategy has been estimated by using an extended Kalman filter. The machine under test is a 7.5kW doubly-fed induction machine. Simulations are carried out by using the software Matlab/Simulink 2018b showing how the electromagnetic torque and the reactive power of the machine can be successfully controlled without the need of the encoder signal; furthermore low current ripple and high dynamic response can be achieved making the studied control strategy suitable for grid-connected variable speed operation such as wind energy conversion systems. © 2019 EPE Association.
Sensorless and modulated model-predictive control for a doubly fed induction machine
Zanchetta P.
2019-01-01
Abstract
In this paper a sensorless control strategy has been implemented and experimentally validated to control independently the electromagnetic torque and the reactive power of a doubly-fed induction machine. The machine acts as a grid connected generator at variable speed operations around the synchronous speed. This control strategy is based on a model-predictive control strategy with the governing equations represented in a synchronous reference frame aligned to the stator flux. A modulation stage has been introduced in order to overcome the well-known issues of the model-predictive control strategies, such as high current ripple and the non-constant switching frequency. The stator flux vector has been identified by using a programmable low-pass filter achieving an acceptable estimate which is not affected by the integer offset derived problems, and it does not require high number of calculations as by using a full-order observer. The rotor position feedback signal needed to implement the control strategy has been estimated by using an extended Kalman filter. The machine under test is a 7.5kW doubly-fed induction machine. Simulations are carried out by using the software Matlab/Simulink 2018b showing how the electromagnetic torque and the reactive power of the machine can be successfully controlled without the need of the encoder signal; furthermore low current ripple and high dynamic response can be achieved making the studied control strategy suitable for grid-connected variable speed operation such as wind energy conversion systems. © 2019 EPE Association.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.