Measurement and assessment of transient torques in a three-phase squirrel cage induction motor

Abstract

Computer simulation models have been developed to predict the magnitude of transient torques when starting and reswitching induction motors (De Sarkar & Berg, 1970:1031, Ghani, 1988:106 and Krause & Thomas, 1965:1038). In these models the starting or running parameters are used so that the squirrel cage induction motor has been represented by its single cage model and the effects of deep rotor bars have not been considered. During transient conditions, in an induction machine with deep bars, the deep bar effect can significantly influence the rotor time constant of the machine, which is an important parameter during reswitching operations. A model was developed by McCulloch, landy, levy & Macleod that accommodates the change in rotor resistance and leakage inductance as the rotor speed changes and it takes into account the effect of saturating the stator and rotor leakage reactances. During the implementation of the research project the motor dynamic equations have been simulated using the CASED (Computer Analysis and Simulation of Electric Drives ) software package developed by McCulloch et al for the simulation of the dynamic performance of motors with the addition of variable speed drive systems. It allows one to predict behavior with or without the improved bar models. The equations are formulated in the direct - quadrature axis form. Similar models have been developed ( Klingshirn & Jordan, 1970 : 1038, Lipo & Consoli, 1984 : 180, Smith, 1990 : 48 and Siemon, 1992 : 412) but more experimental measurements where required to validate the models. This would mean developing and uSing methods to measure the transient characteristics of the motor. Therefore, the measuring system forms a vital part of this work. Successful measurements were achieved of starting and restarting transients on a 75 kW induction motor. The very acceptable correlation obtained between measured and predicted results, using the deep bar model, was very encouraging and without any doubt shows the advantage of using these improved models when predicting performance. This research project has met its objectives by showing how important it is to use the improved model. It is shown in chapter 3 that if the fixed parameter model is used the correlation between measured and predicted results is very poor. When starting a loaded induction motor direct on line, its rotor windings are subjected to mechanical and thermal stresses due the sudden inrush of cu rrent. The performance of the rotor windings, under these conditions, is affected if the motor has loose or broken rotor bars. Techniques have been developed to detect rotor winding malfunctions i.e. broken rotor bars ( Tavner & Penman, 1987:259). During completion of the project there was an opportunity to assess whether a broken bar in the cage of the motor affects the transient behaviour. If a deviation in the transient behaviour of the motor indicates the presence of broken rotor bar(s), the torque measurement system can be economically used as a diagnostic tool to detect broken rotor bars. For th is reason, measurements on an identical machine having a broken cage where taken. These results are also shown in this thesis and it is interesting to note that it was found that the position of the broken bar affected the transient torque developed by the motor during run-up. Although no reason for this is offered in th is work, the author believes that this finding can form the basis for further intensive research. In the project the background information on the equations used for the simulation of the dynamic performance of a squirrel cage induction machine and the CASED computer simulation package are discussed. A method used to measure the torque-speed curve of the motors is also described. The simulated and measured results of the transient starting- and reswitching torque on the healthy motor are discussed. The reswitching transient torques on the healthy motor and the motor with the broken rotor bar are compared . Finally, the measured results of the starting torque on the motor with the broken rotor bar are shown and evaluated.