Simplified Finite-State Predictive Torque Control Strategies for Induction Motor Drives

Abstract

This thesis develops a simplified finite-state predictive torque control (FS-PTC) algorithm based on selected prediction vectors (SPVs). This reduces the number of voltage vectors required to be predicted and the objectives to be controlled. The sign of torque or stator flux deviation and the position of stator flux are used to select the prediction vectors. The proposed SPVs strategy also assists reducing the average switching frequency for a two-level voltage source inverter fed induction motor (IM) drive. As a result, the cost function is simplified, as the frequency term is not required. The proposed SPVs based FS-PTC is also applied to a three-level neutral-point clamped inverter driven IM drive. Using the SPVs strategy reduces the computational burden for the proposed three-level inverter fed drive without affecting the system performance. However, an appropriate weighting factor is required for torque and flux errors in the cost function. This leads to the development of a second simplified FS-PTC which does not require complex torque calculations in the prediction loop and hence tuning effort on the weighting factor. A new reference stator flux vector calculator (RSFVC) with an inner proportional-integral torque regulator is employed to convert the torque and flux amplitude references into an equivalent stator flux reference vector. This stator flux reference is used in the cost function for the flux error calculation. The required processing power for the RSFVC-based FS-PTC is further reduced by combining it with the SPVs strategy. Finally, a speed-sensorless simplified FS-PTC of IM supplied from a 3L-NPC inverter is proposed. The sensorless simplified FS-PTC yields improved torque, flux and speed responses, especially at low-speed. The proposed simplified FS-PTC strategies in terms of computational efficiency, cost function design, torque and flux responses, robustness and average switching frequency are validated through experimental results.