Robustness evaluation of deadbeat, direct torque and flux control for induction machine drives

Field oriented control methodologies are one of the dominant industry standards for induction machine drives. While this classical approach has continuously improved, direct torque control (DTC) has emerged as an alternative approach. Deadbeat, direct torque and flux control (DB-DTFC) is a more recent scheme and employs a discrete-time model inversion to achieve decoupled control of both torque and flux with a one PWM period response time. From z-transform-based deadbeat control design theory, it is known that incorrectly estimated system parameters can lead to forced oscillations. Therefore, it was the goal of this research to rigorously evaluate the robustness of DB-DTFC and to investigate if its advantageous properties are maintained under perturbed operating conditions. An in-depth analysis first isolates the observer´s influence from the control law´s sensitivity to detuned parameters at critical operating points. An enhanced flux estimation accuracy model is derived and analyzed numerically. Then, sets of a comprehensive collection of experiments are used to evaluate DB-DTFC´s robustness in comparison to the de facto standard of indirect field oriented control. It is shown that the main contributor to decreased robustness is the flux estimation algorithm - not the DBDTFC controller itself. The experimental results demonstrate nearly ideal deadbeat control performance and a high level of robustness compared to other control methodologies.