A New Deadbeat Model Predictive Control of Induction Motor Drives without Speed Sensor

In order to eliminate the weighting factor and reduce the computation burden in conventional model predictive torque control (MPTC), this paper proposes a new deadbeat model predictive control. By analyzing the differentiation of torque and stator flux amplitude analytically, a unique stator voltage vector reference is calculated based on the principle of deadbeat torque and flux control (DBMPC) without solving quadratic equation in prior method. A new cost function consisting of voltage vector error is used to select the best voltage vector directly, eliminating the weighting factor in conventional MPTC. Moreover, a different vector selection way is introduced, in which the best vector is determined only by the position of reference voltage vector in the complex vector plane. As a result, the time-consuming enumeration in traditional MPTC is avoided. Finally, a speed adaptive stator flux observer with novel gain matrix is proposed to achieve speed sensorless operation, which has higher speed and flux estimation accuracy than conventional fixed gains. The effectiveness of the proposed method is confirmed by the presented experimental results.