A new approach for sensorless speed control of induction motors and its application

Induction motors are the most preferred electric motors due to having rigid structure, ease ofmaintenance and low cost. Although they were hardly used in digital applications until 1980sbecause of having not only slow response time but also complicated control systems,nowadays they are used in applications of servo systems since vector control algorithm hasenabled to have improved response time. The costs can be reduced further compared to otherelectric drive systems with the sensorless (without velocity feedback) control of inductionmotor.Despite the advantage of reduced costs, the complicated structure of induction motorinvolving variable parameters causes problems in transient response. This fact has limited theuse of sensorless control for induction motors in industrial applications. There have beenvarious efforts from the industry and academia on the structure of observers estimating themotor velocity with the aim of eliminating this limitation. In this thesis, by reducing theadverse effect caused by the variation of the rotor time constant, which is the most effectiveparameter, realization of the sensorless control of induction motor is aimed. For this purpose,a new observer, which is less affected by the parameter variation, is designed with thecombination of sliding mode approach and Luenberger observer. The new observer is thenused in the rotor flux oriented vector control of induction motor. The control hardware isrealized by using a motion control card on a computer. By means of the analog-to-digitalconverter on the motion control card, the current and voltage values of the motor aretransferred to the control algorithm software. The actual speed of the motor obtained from theencoder is compared with the estimated speed computed by the software. Experimental resultshave shown that the estimated speed converged to measured speed in steady state.Prior to the realization of the system, the simulations of the rotor flux oriented control byusing the proposed observer are performed in MATLAB. The observer constants and thecontrol parameters are adjusted according to the simulation results. The comparative resultsand related overall conclusions are presented. Last but not least, C programming language isused for the real-time applications.