Encoderless Parallel Predictive Torque Control for Induction Machine Using a Robust Model Reference Adaptive System

This paper proposes a robust model reference adaptive system (MRAS) estimator incorporating online parameter identification algorithm for parallel predictive torque control (PPTC) scheme. In contrast to conventional predictive torque control (PTC), the proposed PPTC is designed to facilitate the determination of weighting factor that modifies the relative importance of the control objectives, by simultaneously evaluating the torque and flux tracking error terms. The weighting factor is fine-tuned via an adaptive selecting mechanism, to obtain the attractive feature of reduced torque and current tracking errors. Besides, an encoderless MRAS-based estimator is employed for rotor speed and stator flux estimation to reduce the hardware complexity. However, this proposed encoderless PPTC algorithm still suffers from weak robustness against inevitable disturbances caused by parameter variations, which is an important issue to be further investigated. In this paper, an encoderless MRAS estimator combined with online parameter identification algorithm is proposed as an effective solution to accurately estimate the rotor speed and predicted stator flux-linkage, that improves robustness against mismatched parameters for the proposed encoderless PPTC scheme. The feasibility of the proposed algorithm is verified through the obtained experimental measurements, that achieves precise speed tracking capability as well as reduced torque and current tracking errors.

Automated summary

This paper proposes a robust model reference adaptive system (MRAS) estimator for parallel predictive torque control (PPTC) scheme. It determines the weighting factor that modifies the relative importance of control objectives by evaluating torque and flux tracking error terms. The paper also reduces hardware complexity by employing an encoderless MRAS-based estimator for rotor speed and stator flux estimation.