Design and Advanced Control of Switched Reluctance Motors
Publikation: Forskning › Ph.d.-afhandling
The introduction of mainly power electronics and cheap micro computers have made the Switched Reluctance Machine (SRM), which is in focus in this thesis, a feasible alternative to traditional electrical machines like the induction- and DC-motor which have been the dominating electrical machines in a century.
Even so that the classical SRM is a very simple construction, characterized with that both the stator and rotor have salient poles and that only the stator has conducting material, the engineering physics required to analyze, design, improve and operate the SRM is untraditional and very complicated.
A main goal for this thesis here is to construct user-friendly tools for investigation and improvement of the SRM technology. As tangible main tools are a SRM design and simulation computer program (SRDaS), a General Purpose Control System (GPCS) with two flexible four phase converters and a static characterization system developed.
To simulate and analyze the electromagnetic performance of different variations of SRMs, in for instance SRDaS, is a general dynamical model derived, which also takes into account SRMs having permanent magnets. The parameters for the models are obtained with 2D-Finite Element Analysis (FEA), measurements from the developed static characterization system and with special developed SRDaS calculation module. This module is based on simple FEA routines together with some classical analytical functions/routines. The different methods to acquire parameters are all compared and it is seen that the 2D-FEA method has problems due to large end-effect, when the rotor is in the unaligned position. With the measured model parameters the developed dynamical model is also verified. A good agreement is seen between dynamic SRDaS simulations and measurements.
One of the real big problems with SR technology is the reputation of the high acoustic noise emission compared to other electrical machines. Therefore, extra focus to analyze this specific subject is done. The major origin to the high level of acoustic noise is due to highly concentrated normal forces with a high harmonic content which easily triggers resonances in the SRMs stator yoke. In order to better understand and reduce the acoustic noise already in the early design state of a new SR-product is a unique model developed and implemented in SRDaS. The input to the model is the physical layout of the SRM together with the control strategy and the output is the sound pressure at the top of the stator yoke. This is presented as both a technical A-weighted value and as a subjective value in form of the sound from the PC having the SRDaS program installed. The developed acoustic models are also verified in practice on several different SRMs with good results.
The control of SRMs is also in focus, where both basic and advanced control are considered. Some hands-on design methods for speed and current controllers are developed and simulated in SRDaS, and finally verified in practice on the GPCS. A simple position and speed sensorless method is also analyzed in details both theoretically and in practice on the GPCS. With an implementation on a specific SRM with a speed controller the motor was still operating sensorless at a minimum speed of around 500 rpm.
To operate the SRMs in an efficient way the settings of commutation angles are very important. Optimization tools for setting them are thus constructed in both the GPCS and in SRDaS. With the constructed tools it is found that the optimal operation point was not so sensitive to the commutation angles. From the results gathered from the optimization tools some simple strategies to set the commutation angles are also proposed.
The developed optimization tools are also used in dynamometer test where the performance of three practical SRMs have been mapped with special attention to the efficiency. One of the three tested motors has dimensions and ratings quite similar to a standard induction motor, and from the test it is seen that the nominal efficiency of the SRM is 83 % compared to only 72 % for the induction motor.
During the work with this thesis some follow-ups are done which were not specified in the main-goals. But these is very important contributions in the SR-technology. The follow-ups are "Shark-motors" which is a new way to construct SRMs with both axial and radial forces. "Spacers" which stiffens up the stator structure with reduced acoustic noise as a consequence. Finally, "E-core transverse flux machines" which is SR-machines made from standard single-phase transformer laminations.
|Udgiver||Institut for Energiteknik, Aalborg Universitet|