| dc.description.abstract | The recent advancements in power electronics have resulted in widespread use of electronically controlled energy sources and loads in both AC and DC distribution systems. Power electronics also plays an important role in the emerging and future transportation systems like electric vehicles, electric ships, and more electric aircraft. Distributed power systems based on power electronic converters are increasingly used in on-board power systems. Furthermore, power electronics-based microgrids are regarded as key technologies to be part of the future smart distribution systems. However, instability is still a major issue in the design of such systems, mainly because of the nonlinear and chaotic behaviour of power electronics systems. These instabilities may be initiated either by the interaction between power electronics converters themselves or between the converters and other elements of the power system. Modelling and stability analysis of such systems are complex and cumbersome. In this Thesis, the main focus has been to develop stability analysis tools that can capture the potential instabilities in power electronics-dominated distributed power systems.
Most of the established research in stability analysis of the power electronics is based on averaged modelling of converters. The classical stability tools based on the averaged modelling have limited validity and accuracy, since the switching frequency and the nonlinear dynamic of the system are usually neglected in these methods. Therefore, the dynamic model of the power electronics system is improved in this research, taking into account the nonlinear system dynamics originating from the switching frequency and intrinsic nonlinearities of the system. For this objective, a discrete-time model of the distributed systems with pulse width modulated (PWM) converters is proposed as a basis for nonlinear stability analysis of the system. Discrete-time eigenvalues of the system are then calculated using the proposed model and are used to predict the system instabilities. Moreover, time-domain characteristics of the system are drawn from the discrete-time method, and are used for bifurcation analysis of the system. To improve the asymptotic stability of the system, an active stabilizer is proposed and is included in the system model. This stabilizer adds additional damping to the controller of the energy source to suppress the unstable oscillations of the system.
The proposed method is applied to a simple, but general, distributed system architecture with power electronics-based sources and loads. The analyses show that the proposed method is able to identify the well-known slow scale dynamics, as well as the fast scale dynamics of the system which cannot be identified by the averaging based methods. The discrete-time method is compared with the conventional method and the potential advantages of the proposed method are discussed. All concepts and techniques proposed in this Thesis are verified by theoretical analysis as well as numerical simulation and laboratory experiments. According to the theoretical and experimental results, the discrete-time method can provide an accurate measure of the stability margin of the system. Therefore, instead of trial and error, the proposed method can be applied both to predict the instabilities and to guarantee the stability of the system during the design process. | nb_NO |
