Search

During the last few years, the control of induction machine by the field orientation method has become the preferred technique which yields very good performances. This method allows to define the dynamic structure of an induction machine which is similar to that of D.C machine. This is obtained by controlling the angular position of the rotor flux so that the flux and torque can be decoupled. The process to be controlled is composed of an autopiloted induction machine fed by a voltage inverter. In this paper the authors study the influence of different problems due to the numerical processing of informations by the computer on the system performances. These problems, usually neglected, are closely related to the performances and the dynamic of the system to be controlled. It should be noted that the vector control method is very sensitive to the precision with which the angular position of the rotor flux has been determined. The use of a computer in the regulation chain implies a sampling time and a double quantization concerning the measured and control parameters. In the continuous model, the input parameters allow to determine the exact amplitude and orientation of the rotor flux vector, but the model discretization and the input sensors produce an uncertainty of the rotor flux. The induced error on the rotor flux angular position depends essentially on the resolution of the different sensors. In this paper the authors show the influence of the following elements on the system stability and performances:-Quantization of input parameters -Discretization and time delay produced by the sampling of measured signals -Time delay produced by the inverter sampling.The Authors present a control scheme (fig 2.1) which takes into account the quantization induced by the different sensors. It can be observed that the proposed control structure is different from the conventional one. The main difference consists in replacing the three control signals of the PWM inverter (Vslref, Vs2ref, Vs3ref) issued from Vsdc and Vsqc, by a cyclic ratio signal. In this way the control is simplified by eliminating a coordinate transformation. A numerical simulation of the process Fig.2.1 is done. This simulation takes into account the scaling factors due to the numerical data processing, the sensors resolution and the error caused by the data truncation during the processing. The control realization is based on the digital techniques developed around MC 68000 microprocessor.