Abstract
Due to their exhibition of nonlinearities, electrical power systems, such as shipboard power systems, provide an interesting research environment. Nonlinear systems are characterized by properties such as highly interdependent components and thus it is not hard to observe that power systems may display such phenomena. Nonlinear dynamics, such as bifurcations and chaos, have been shown to cause voltage collapse, angle divergence, and other detrimental issues in power systems. Power systems of the future, such as those envisioned for NGIPS, will potentially have much higher degrees of structural flexibility (from and electrical point of view), which allows them to adapt in the case of a fault, or drastic change in load demand. The design process of power electronics is driven by practical industry applications to optimize the performance of power electronic devices, i.e., higher reliability and performance. Bifurcation analysis can be used as a tool to design power electronic systems because the performance of the system is determined by the parameters chosen in the design process, and bifurcation points are sensitive to these parameters values. Although in general bifurcations should be avoided, designing too far from the bifurcation point may decrease the performance of the system. Since a linear model is required for the use in designing conventional controllers, the averaged circuit is linearized about an acceptable operating point. These averaging methods (state space averaging, iterative mapping, etc.) intend to replace the more complex nonlinear, time-varying dynamical system with a simpler averaged linear system. In the converter model design process, accuracy is often sacrificed for simplicity. Considerable research has been applied in developing linearized power electronic models that can be used in traditional small-signal analysis such as closed loop stability and transient response characteristics. Although the use of the linearized models of switching converters to test system performance is well developed, they may not well-predict nonlinear behavior that is present. If the nonlinear effects are not accounted for, the switching converter could be designed for optimal performance in the linearized state but still be non-optimal from a global viewpoint.
|
