Controlador GPC aplicado à malha de velocidade de um motor de indução trifásico

Abstract

This work aims to apply a generalized predictive controller (GPC) strategy based on finite impulse response (FIR) digital filters to the velocity loop of a three phase induction motor through a vector modeling. For such purpose, it is necessary to perform the dynamic modeling of the machine regarding to the electrical and mechanical systems, as well as to demonstrate coupling between electric and mechanical models. The vector control technique is based on the well established Clarke and Park transformations, which are a linear operators that converts the three-phase and sinusoidal stationary system into a rotating, time-invariant system, making the induction machine model similar to the a direct current machine. The direct and quadrature axis currents control flow and torque respectively and are uncoupled from each other. For the control of these currents the method of the root locus has been adopted. Direct shaft current is independently controlled and keeps the machine magnetization. Quadrature shaft current is controlled by a cascade structure along with speed control. Such cascade structure is applicable because of the different settling time between electrical and mechanical variables. The current controller provides a reference for quadrature shaft current due to its relationship with the machine torque. The speed controller is based on a GPC structure based on recursively implemented digital filters. All processing is performed through a digital signal platform (DSP) that performs control structure upgrades through counter logic and the use of internal timers. Finally, simulation results performed in PSIM software are compared with experimental results.