Variable-Speed, Robust Synchronous Reluctance Machine Drive Systems

Research output: ResearchPh.D. thesis

Abstract

The synchronous reluctance machine drive is getting more and more interests from the industrial side, since it can provide higher system energy efficiency than traditional inverter-fed induction machine drive systems with similar production cost. It is considered as a good candidate for super premium efficiency machine and commercial products are available in the market. The research work in this dissertation aims at developing a simple, compact and robust synchronous reluctance machine drive system that can provide satisfactory performance with optimized system energy efficiency at various working conditions.
Field oriented control assisted with various position estimation algorithms is in-vestigated. Position sensing via machine flux linkage is implemented with the assistance of a widely used flux observer. Experiments show that it may not always work properly and system oscillation will occur when there is a failure of the flux observer. Solutions from both flux observer enhancement and controller bandwidth adjustment point of views are proposed and verified experimentally. Both of them can work satisfactorily. The enhanced flux observer could provide better perfor-mance at steady state conditions, while the latter, which is realized by using adap-tive PI controllers, is more robust especially during the dynamic transients.
Moreover, position sensing via machine anisotropies based on high frequency signal injection is investigated, since it could possess the feature of machine inductance independency. High frequency signal injection based on duty cycle shifting is introduced, which can achieve signal injection without sacrificing the maximum allowable fundamental voltage for torque production, so that the machine induct-ance independent position estimation algorithm can be implemented in the whole machine operation range. As a supplement to the proposed voltage lossless injection method, a machine inductance independent position estimation algorithm, which is suitable for arbitrary injection, is developed and tested. Compared with the flux linkage based position estimation algorithm, the proposed machine inductance in-dependent position estimation algorithm shows better dynamic response against the step load changes but requires more accurate signal sampling in steady state. Therefore, the flux linkage based position estimation algorithm with adaptive PI controllers is recommended.
In recent years, there is an increasing trend to replace the electrolytic capacitor in the frequency converter with film capacitor, which has a longer expected service lifetime and no explosion risk. Furthermore, it is possible to achieve a compact converter design by using film capacitor, since the capacitance needed is less due to its high ripple-current-tolerant capability. This is known as small dc-link drive system, which often suffers from stabilization problems. Active damping methods to stabilize the dc-link voltage are investigated in the dissertation. The essential idea and technique details to implement active damping control are discussed. Control strategies, which are of great practical values such as minimized system current stress control and optimized voltage utilization control, are proposed and verified experimentally.
The existing active damping methods are mainly based on the variation signal injection. A common problem of these methods is that they require on-line calculation of the compensation gain factors, which are system parameters and operating condition dependent. To overcome these difficulties, a “virtual positive impedance” active damping method is proposed, which can ensure the system stability without the knowledge of the system parameters and operating conditions. This method has a clear physical meaning and provides a new way to achieve active damping control. It is verified experimentally and satisfactory results are achieved.
Finally, the preferred flux linkage based position sensorless control with adaptive PI controllers is used to cooperate with different active damping controls to drive the small dc-link synchronous reluctance machine drive system. The possible inter-action between sensorless control and active damping control is investigated. The investigation results show that they can cooperate well in steady state conditions. While for step load change conditions, the existing active damping control with d-axis injection can affect the performance of the sensorless control. However, the proposed “virtual positive impedance” active damping control can cooperate with the sensorless control well. Thus, the flux linkage based position sensorless control with adaptive PI controllers and “virtual positive impedance” active damping technique is recommended as the solution of the simple and robust variable-speed syn-chronous reluctance machine drive system.
Close

Details

The synchronous reluctance machine drive is getting more and more interests from the industrial side, since it can provide higher system energy efficiency than traditional inverter-fed induction machine drive systems with similar production cost. It is considered as a good candidate for super premium efficiency machine and commercial products are available in the market. The research work in this dissertation aims at developing a simple, compact and robust synchronous reluctance machine drive system that can provide satisfactory performance with optimized system energy efficiency at various working conditions.
Field oriented control assisted with various position estimation algorithms is in-vestigated. Position sensing via machine flux linkage is implemented with the assistance of a widely used flux observer. Experiments show that it may not always work properly and system oscillation will occur when there is a failure of the flux observer. Solutions from both flux observer enhancement and controller bandwidth adjustment point of views are proposed and verified experimentally. Both of them can work satisfactorily. The enhanced flux observer could provide better perfor-mance at steady state conditions, while the latter, which is realized by using adap-tive PI controllers, is more robust especially during the dynamic transients.
Moreover, position sensing via machine anisotropies based on high frequency signal injection is investigated, since it could possess the feature of machine inductance independency. High frequency signal injection based on duty cycle shifting is introduced, which can achieve signal injection without sacrificing the maximum allowable fundamental voltage for torque production, so that the machine induct-ance independent position estimation algorithm can be implemented in the whole machine operation range. As a supplement to the proposed voltage lossless injection method, a machine inductance independent position estimation algorithm, which is suitable for arbitrary injection, is developed and tested. Compared with the flux linkage based position estimation algorithm, the proposed machine inductance in-dependent position estimation algorithm shows better dynamic response against the step load changes but requires more accurate signal sampling in steady state. Therefore, the flux linkage based position estimation algorithm with adaptive PI controllers is recommended.
In recent years, there is an increasing trend to replace the electrolytic capacitor in the frequency converter with film capacitor, which has a longer expected service lifetime and no explosion risk. Furthermore, it is possible to achieve a compact converter design by using film capacitor, since the capacitance needed is less due to its high ripple-current-tolerant capability. This is known as small dc-link drive system, which often suffers from stabilization problems. Active damping methods to stabilize the dc-link voltage are investigated in the dissertation. The essential idea and technique details to implement active damping control are discussed. Control strategies, which are of great practical values such as minimized system current stress control and optimized voltage utilization control, are proposed and verified experimentally.
The existing active damping methods are mainly based on the variation signal injection. A common problem of these methods is that they require on-line calculation of the compensation gain factors, which are system parameters and operating condition dependent. To overcome these difficulties, a “virtual positive impedance” active damping method is proposed, which can ensure the system stability without the knowledge of the system parameters and operating conditions. This method has a clear physical meaning and provides a new way to achieve active damping control. It is verified experimentally and satisfactory results are achieved.
Finally, the preferred flux linkage based position sensorless control with adaptive PI controllers is used to cooperate with different active damping controls to drive the small dc-link synchronous reluctance machine drive system. The possible inter-action between sensorless control and active damping control is investigated. The investigation results show that they can cooperate well in steady state conditions. While for step load change conditions, the existing active damping control with d-axis injection can affect the performance of the sensorless control. However, the proposed “virtual positive impedance” active damping control can cooperate with the sensorless control well. Thus, the flux linkage based position sensorless control with adaptive PI controllers and “virtual positive impedance” active damping technique is recommended as the solution of the simple and robust variable-speed syn-chronous reluctance machine drive system.
Original languageEnglish
PublisherDepartment of Energy Technology, Aalborg University
Number of pages176
StatePublished - 2015
Publication categoryResearch
ID: 229257604