Sensorless Rotor Velocity Tracking Control for Induction Motors


The natural progression of research in the area of induction motor control is intended to provide more sophisticated control algorithms to accomplish a widely varying range of design objectives (e.g., rotor position/rotor velocity tracking, compensation of parameter uncertainty, etc.) while maintaining a high degree of reliability. That is, the control algorithm accompanied by the necessary hardware and software should achieve the desired control objective with a minimum possibility of failure. Incorporated in reducing this possibility of failure is the necessity that the hardware needed for execution of the control scheme be as simplistic as possible. The use of sensors in the control architecture inherently increases the possibility of failure (e.g., alteration of the mechanical system to accommodate sensor placement, sensor calibration and accuracy, data acquisition and processing, etc.) of the control process due to the additional hardware; hence, the benefits of sensorless control schemes are obviously apparent.

Research Objective:

Many practical approaches to sensorless control of the induction motor, many of the control strategies for sensorless control of the induction motor display one or more of the following shortcomings: i) the lack of a rigorous mathematical development or stability proof for the closed-loop system, ii) the nonlinearities of the electromechanical system are often neglected (i.e., the system model is linearized), iii) and the full order model of the electromechanical system is not utilized in the control development (i.e., the stator current dynamics and/or the mechanical system dynamics are often neglected). To overcome these shortcomings, we propose a sensorless observer/control algorithm that achieves semi-global exponential rotor velocity/rotor flux tracking for the full-order, nonlinear system model of an induction motor actuating a mechanical subsystem (i.e., only stator current measurements are required).


We construct a nonlinear, rotor velocity observer that exploits a ``back-emf'' term appearing in the stator current dynamics. The nonlinear, rotor velocity observer is endowed with feedback structure which facilitates the potential for improved rotor velocity tracking transient performance. That is, as opposed to open-loop rotor velocity observer, an observer feedback gain can be adjusted to enhance closed-loop performance. We then design the desired torque signal based on the structure of the observed mechanical subsystem to promote rotor velocity tracking. The desired rotor flux trajectory in conjunction with the desired stator current trajectory and an auxiliary control variable are then designed to ensure that the desired torque is delivered to the mechanical subsystem and that rotor flux tracking is achieved. We then design the stator voltage control inputs to ensure stator current tracking. We conclude the control development with a Lyapunov type stability analysis for the composite observer-controller system.

The Experimental Setup:

An experiment was conducted on a three phase induction motor (Baldor Electric Co., Model M3541) powered by three linear amplifiers (Techron, Model 7570-60) to test the performance of the proposed controller. Three hall effect sensors (Microswitch, Model CSLB1AD) were used to measure the stator phase currents. A QNX based real time Photon-windows environment developed in-house serves as the user-interface required to implement the control algorithm. The control algorithm is computed on a Pentium processor thus eliminating the need for a separate DSP board. The sampling frequency was selected to be 1500 Hz. The MultiQ board (8 A/D, 8 D/A, and 6 encoder channels) manufactured by Quanser Consulting was used to output the three phase voltages to the induction motor. The shaft of the induction motor was directly coupled to a separately excited direct current (SEDC) motor (Baldor Electric Co., Model CD3433) to provide a constant load torque.

the experimental setup

Some Experimental Results:

Select and click to view some of the experimental plots:


For more information on this research, please refer to the following publication:

M. Feemster, P. Aquino, D. M. Dawson, and D. Haste, "Sensorless Rotor Velocity Tracking Control for Induction Motors," Proceedings of the IEEE American Control Conference, San Diego, CA, June 1999, pp. 2158-2162.