Measurement and simulation of non-linear effects in induction motors in order to determine the flux position without mechanical sensors.

In modern drive applications the combination of inverter-fed induction motors and field oriented control has superseded the conventional dc-motor even if dynamic is important. One disadvantage of this kind of control is the need of a mechanical sensor at the motor shaft to determine the flux position. Especially in the power range down from hundred kW the expenditure of this sensor has more an more influence on the costs of the whole drive. In many applications (especially in traction drives) the mechanical sensor is only needed for the control of the induction machine and therefore eliminating this sensor would lead to a cheaper and more robust control structure. This project deals with a new flux detection method which determines the flux position by estimating the maximum saturation along the air gap of the machine without using a mechanical sensor. This is achieved by measuring the change of the armature current due to a voltage test phasor applied by the inverter. This method will be investigated concerning its applicability to different induction machine designs. By combining measurement and simulation the influence of non-linear effects on the new method will be determined and by that design rules for mechanical sensorless controlled induction motors will be established.

The topic of this project was the sensorless control of inverter fed induction machine drives at zero speed. The term sensorless is used in literature for the elimination of the mechanical shaft sensor which is advantageous not only in terms of costs but also reference to the reliability of inverter fed drives. The goal is to replace this shaft sensor by an intelligent control scheme that works in the whole range of operation (including zero electrical speed). All published sensorless zero speed control methods suffer from the heavy dependence of the control signals on the point of operation. In addition also the design of the machine has strong influence on the performance of the control. These drawbacks till now have limited the application of these methods to laboratory and prevented an industrial application. This project was focused on the reason of that dependence and investigations were made on how the design parameters of induction machines have to be changed in order to increase the performance of the control. The results thus form the basis for an industrial application of sensorless zero speed control of induction machines. The whole know how in this field of research is concentrated at the department. A timely introduction of such a method to industry application would lead to an advantage over the international competitors resulting in a strengthening of the local business base. All modern inverter fed induction machine drives are equipped with a shaft sensor to enable a stable operation in the speed range around zero fundamental frequency. This sensor leads to additional costs leads to an increase of the drive dimensions and deteriorates the reliability of the drive as a breakdown of the sensor usually effects the whole control structure. Recently different control methods have been proposed in literature to replace the shaft sensor even at zero speed operation. They are based on an excitation with high frequency voltage or transient voltage pulses and are able to estimate the rotor or flux position from the reaction of the machine current. The voltage pulses with a duration of some 10s are impressed by the switching of the inverter. All modern inverter fed drives are already equipped with current sensors and micro processors necessary to realise this method. The investigations made in the project were focused on the influence of different induction machine design parameters in the transient electrical behaviour (the reaction on the voltage pulses). This was performed by measurements on specially designed and constructed machines. The signals of test coils placed at different places of the machine revealed a further insight into the transient flux distribution. In a second stage a simulation model of the machine has been developed which also covers the transient effects (transient hysteresis etc.) in the lamination during the test pulses. By comparing the measurements and simulations a connection could be established between the design parameters and the control signal performance. As a result it is possible to increase the performance of sensorless control already in the design process by doing the optimisation of the machine not only reference to the fundamental wave but also reference to the transient electrical behaviour. The project results are, however not only applicable to sensorless control. As it turned out it is also possible with a slight modification to realise a very effective monitoring of the machine reference to possible fault conditions. As drives are more and more introduced into safety critical applications like steer-by-wire, brake-by-wire or the actuation of flight control surfaces the importance of monitoring will increase in the near future.