Fractional order integral sliding-mode flux observer for direct field-oriented induction machines

In field-oriented, induction machines, an accurate flux estimation is crucial for high-performance speed control. However, the flux estimation is sensitive to parameter variations such that the control performance will be deteriorated. This paper presents a novel flux observer, fractional order integral sliding-mode (FOISM) flux observer, to estimate the d- and q-axis fluxes in the stationary reference frame. The closed-loop stability is guaranteed by employing the Lyapunov stability theory. Essential properties of fractional operators are also discussed for realizing fractional order integrations and differentiations properly. In addition to numerical analyses and simulations, a DSP/FPGA based experimental platform is set up to evaluate the feasibility of the proposed control framework. Simulation, results indicate that the use of fractional-order schemes leads to better results than the counterparts of integer-order approaches. Also, experimental results demonstrate that the desired, speed and flux tracking of an induction machine can be performed by utilizing the FOISM flux observer.