Adaptive fuzzy estimator for a sensor-less IM drive

A novel sensor-less system for induction motors is designed. This novel design is based on an adaptive fuzzy system, which is obtained by mixing an open-loop estimator response with a steady-state estimator. In order to find the best response a ponderating average system is used. The ponderated average depends on the working point of the plant, this point is found without any further measure over the plant. The open-loop estimator response is improved by means of using an adaptive fuzzy controlled filter that selects the optimised cut frequency. The aim of this system is to obtain a good performance with neither great computational charge nor mathematical noise in the estimated process. Besides, the final system is more robust against variations of the internal parameters of the motor. The results validate the entire work, not only in transient and steady state but also in the start-up.