Dynamic Loss Modeling for Loss Minimizing Control of IPMSM using DB-DTFC not Operating in Voltage or Current Limits.

摘要

Deadbeat - direct torque and flux control (DB-DTFC) provides the opportunity of achieving desired torque with minimum losses at each switching period. With DB-DTFC, dynamic loss minimization control can be achieved without compromising fast torque response. This thesis lays the foundation of a flux linkage-based dynamic machine loss model of IPMSMs that is integrated into DB-DTFC achieving dynamic torque and loss minimization control at each switching interval.;A dynamic machine loss model as a function of Volt-sec selection at the switching period level has been developed. This model, which utilizes lambda-I relationship in place of the magnetic B-H relationship, has been evaluated and shown to accurately represent the losses in the machine for sinusoidal loading. This model is evaluated experimentally for different types of cyclical loading and driving cycles.;The stator flux linkage needs to be accurately estimated to provide a precise loss model for dynamic loss minimization control. For this purpose, a reduced parameter sensitivity stator flux observer utilizing disturbance input decoupling (DID) has been developed so that even under varying or inaccurately estimated machine parameters, the stator flux linkage and torque can be accurately estimated. However, this DID stator flux observer is vulnerable to the terminal Volt-sec distortion due to the non-ideal inverter effect at very low speed operations. To mitigate this drawback, a high frequency injection (HFI) based parameter estimation method combined with recursive least squares (RLS) has been developed for accurate torque and flux estimation.;The outcomes of this work are a rigorous documentation of the capabilities and limitations of dynamic loss models and methods that are used to provide precise torque and flux estimation for dynamic loss minimizing DB-DTFC.