Control of vibration in switched reluctance motor drives.

摘要

Switched Reluctance Motor (SRM) drive system has attracted considerable attention in industry and academic research. However, the levels of acoustic noise emitted by switched reluctance motor drive are noted as being above those of competing variable speed drives.; The dominant source of acoustic noise in switched reluctance motor drive has been shown to be radial vibrations of the stator. The radial vibrations of the stator arise from the torque production mechanism in switched reluctance motor drive. In order to obtain a satisfactory performance of the switched reluctance motor, phases are turned off around aligned position where radial forces are maximum. Further, in order to avoid negative torque generation, phase current has to be shut down as fast as possible thereby producing a high rate of change in radial forces.; Using a two-dimensional finite-element method, electromagnetic behavior of the SRM drive is studied first to predict the static characteristics of SRM. These static data will be then used in the development of a non-linear dynamic SRM model to predict the steady state as well as the dynamic performance of the SRM. The predicted dynamic performance compares favorably with the experimental results over a wide range of torque-speed variation.; This dissertation introduces a design philosophy for switched reluctance motor to improve the vibrational behavior of the machine. The effects of the stator dimensions on the vibration of the SRM are investigated. Moreover, using a time-stepping finite element analysis, transient acoustic noise emitted by radial vibration of the SRM is computed. The predicted values are in agreement with those obtained from measurement.; A new current profiling technique for mitigating the radial vibration in SRM drive is also presented. However, it has been found that a radical reduction of vibration by profiling the tail current might reduce the efficiency of the drive. Finally, using a neural network based controller, this current profiling scheme is applied to the entire torque-speed region. The trained ANN will recreate the optimal control parameters on-line in real time. This algorithm is successfully applied to the SRM drive and experimental results are given to support the proposed algorithm.