Impact Of Exciter And Governor System Model On The Dynamic Security Region Of Power Systems

Dynamic security region(DSR) is the new theory and technology of power system transient stability analysis for dispatching center. Dynamic security region is based on the power system adjustable parameter space, it can be directly used by the dispatching department. At the same time it analysis the system transient stability from the global view, so it can improve accuracy, reduce false. It is different from transient stability region, which is based on the system state variables space.The concept of dynamic security region was introduced many years ago. A lot of researchers has made tremendous contributions to dynamic security region, so it has been made practical by a series of modifications and enhancements in the last couple of years. Based on the predecessors'work, this paper extend the dynamic security region theory from the differential system to differential-algebraic systems. And then apply it to the power system including excitation and speed governing system. Futher more depicts the dynamic security domain by using of quadratic approximation, Q linear approximation, L linear approximation. LO linear approximation.Excitation system can significantly improve power system transient stability. So it must be able to expand the dynamic security region. Construct the DSR including excitation system model, it means that interpret the impact of excitation system on transient stability from the perspective of DSR.Speed governor system also has effect on DSR. And the simulation results shows that the governor can enlarge the DSR used alone. But when the excitation and governor system used together, it shrink a little. This is caused by the water hammer effect of hydroturbine governor (i.e., the input mechanical power should reduce after the fault occur, but in fact it increases firstly then reduces with the water hammer effect).Futher more the quadratic approximation is closest to the true boundary mostly among the different approximations, while the LO linear approximation has the largest error, and the Q linear approximation is more accuracy than the L linear approximation.