This paper presents a digital implementation of deadbeat-direct torque and flux control (DB-DTFC) for induction machines at low switching frequencies (SFs). For high SFs, existing discrete-time flux observers and Volt-second-based inverse torque models used for DB-DTFC achieve acceptable flux estimation accuracy and fast torque response. However, the flux estimate is less accurate when the SF is reduced and the DB-DTFC performance degrades. This paper develops a more suitable flux observer and Volt-second-based inverse torque model that minimizes flux estimation error and improves torque control for DB-DTFC. Simulation and experimental results are provided to evaluate the performance of the proposed observer and torque model at very low SFs. Consequently, digital implementation of low-SF DB-DTFC on high-power induction machines is feasible. Index Terms-Deadbeat-direct torque and flux control (DB-DTFC), flux observer, low-switching-frequency (SF) operation.NOMENCLATURE ( ) s Stationary reference frame. ( ) r Rotor reference frame. ( ∧ ) Estimation quantity. ( ) * Reference quantity. ( · ) Time derivative operator. ( ) s Quantity on the stator side. ( ) r Quantity on the rotor side. ( ) qdx Complex vector quantity. (k) Quantity at the kth sampling time. (s) Quantity in the Laplace domain. s Laplace operator. v qds Stator voltage complex vector. i qds Stator current complex vector. i qdr Rotor current complex vector. λ qds Stator flux linkage complex vector. i qdr Rotor flux linkage complex vector. R s Stator resistance. R r Rotor resistance. L m Magnetizing inductance. L ls Stator leakage inductance. L lr Rotor leakage inductance. σ = 1 − (L m L m /L s L r ) Total leakage factor. τ r = R r /L r Rotor time constant. ω r Electrical angular velocity of rotor. ω br = (R r /L r ) − jω r Rotor break frequency. T e Electromagnetic torque. T s Sampling time.

This paper presents a digital implementation of deadbeat-direct torque and flux control (DB-DTFC) for induction machines at low switching frequencies. For high switching frequencies, existing discrete time flux observers and Volt-sec.-based inverse torque models used for DB-DTFC achieve acceptable flux estimation accuracy and fast torque response. However, the flux estimate is less accurate when switching frequency is reduced and DB-DTFC performance degrades. This paper develops a more suitable flux observer and Volt-sec.-based inverse torque model that minimizes flux estimation error and improves torque control for DB-DTFC. Simulation and experimental results are provided to evaluate the performance of the proposed observer and torque model at very low switching frequencies. Consequently, digital implementation of low switching frequency DB-DTFC on high power induction machines is feasible.

Speed sensorless vector control for induction motors using an adaptive flux observer makes it difficult to output high and stable torque at a standstill condition when the output voltage error mainly due to dead-time of the inverter. This voltage error leads to a speed estimation error. Consequently, axis misalignment occurs, and a torque value that tracks the reference command value cannot be output. A method has been proposed to compensate for this voltage error by using a current command value; however, the compensation accuracy is limited because of the forward voltage drop and nonlinear characteristics of power switching devices. This study applies the conventional dead-time compensation to the model reference adaptive system, and the effect of varying the amount of compensation, instead of investigating the true dead-time length, on the start-up is evaluated using a 2.2 kW experimental system and through dynamic analysis.

No abstract