Nonlinear adaptive control for robust wide speed range operation of IPMSM drives
Schoonhoven, Garin Matthew
SubjectPermanent magnet synchronous machines (PMSM)
Motor control algorithms
Permanent magnet motors
MetadataShow full item record
Various applications, including robotics, spindle drives, machine tools, etc. rely on accurate, reliable controllers to deliver the required drive performance. With recent advances in magnetic materials and semiconductor technology, machines such as the permanent magnet synchronous machine (PMSM) family of ac drives have seen a rise in popularity, owing to the high power density, efficiency and relative longevity as compared to conventional ac motors. In particular, interior permanent magnet synchronous machines (IPMSM) are characterized by all the features of the PMSM family, with the additional possibility of improved efficiency due to rotor construction, making them ideal for critical applications with high performance demands. Notably, despite the advantageous aspects of PMSM motors in general, control of this class of ac machines is complex if full performance potential is to be realized. In order to achieve optimal efficiency while permitting wide speed range operation, it is crucial to design controllers that are capable of delivering this high performance. Due to the nonlinearity of magnetic flux distribution during operation, the parameters of the PMSM may vary significantly. Thus, a high performance controller must be capable of optimizing efficiency while maintaining excellent response characteristics from set-point or loading variations. As a result of the nonlinear flux distribution caused by rotor/stator magnetic field interactions, direct control of PMSM in the stator reference frame is not possible as the level of mathematical complexity renders it infeasible. Expression of the PMSM stator variables in the rotating rotor reference frame permits the effective decoupling of machine variables into velocity and torque control components. This is roughly analogous to separately excited direct current (DC) motors, where control of the rotor speed (field magnetization) and shaft torque (armature current) are decoupled as a function of the design. Analysis of the PMSM model in the rotating reference frame shows that the “d” and “q” axis currents are principally responsible for indirect air gap flux control and developed shaft torque, respectively. Traditional linear type control techniques based on proportional-integral-derivative (PID) controllers are able to achieve moderate success in controlling the PMSM family. The performance achieved is however typically within a narrow operational band and without the ability to adapt to parametric variation or optimize efficiency. This restriction makes PID type controllers non-ideal for more demanding applications that require highly accurate control and high efficiency regardless of load, temperature, machine age or operating environment. Therefore, this thesis presents a robust nonlinear control algorithm utilizing an adaptive back-stepping technique with flux control for optimizing developed torque and improved operational range. Further, global asymptotic stability of the proposed controller is assured through Lyapunov’s stability criterion in conjuncture with criterion supported by Barbalat’s lemma. The proposed control algorithm ensures that the machine operates at precise command speeds, coping with system uncertainties and disturbances, while reducing losses and enabling operation over a wide speed range. Simulation of the proposed system is carried out in MATLAB/Simulink, as well as in a cosimulation environment utilizing MATLAB/Simulink and PSIM. The first scenario implements an ideal mathematical system model with the controller in Simulink; whereas the second scenario uses PSIM to host the dynamic system model, with MATLAB/Simulink hosting the controller. This co-simulation permits rapid, accurate system analysis, by employing more accurate software models for switching elements, synchronous machine and any reactive elements not reflected in the basic mathematical model. Simulation results from both methods indicate excellent performance and robust operation, with excellent disturbance rejection. Real-time implementation of the system is realized utilizing the DS1104 digital signal processor (DSP) in conjuncture with an IPMSM commutated by a three-phase two-level insulated gate bipolar transistor (IGBT) inverter, with a direct current (DC) generator as dynamic load. Performance of the proposed controller have been verified through experimental implementation for a range of operating conditions.