DAMPING CONTROL OF POWER SYSTEMS WITH MAGNETIC ENERGY STORAGE
Abstract
The power system dynamic instability problem on the Pacific AC/DC Intertie has been solved by modulating the converter power of the HVDC tie line. In the present research an alternate control approach using magnetic energy storage is explored for more general power systems without HVDC lines. Dynamic equivalents of power systems are first formulated using a component connection model to serve as the basis of an eigenvalue study. The unstable modes of such a linearized power system are factored out and eigenvalue assignment through state feedback is performed. The state feedback gains obtained herefrom are then transformed back into the original system. These gains suggest the use of output feedback using the rates of AC power changes on transmission lines. A Least Squares approximation of the state feedback gains enables the control signals to be obtained by local output feedback. The limits of the converter control voltage may be imposed on the output feedback signals to determine the size of the energy storage unit required. The Ludington pumped storage system is used as a test system. An IC (Inductor-Converter) unit is connected to the system. Unstable eigenvalues are moved to the left half plane using this control design. Effect of the control design after the installation of power system stabilizer is also shown. Both pumping mode and generating mode are tested. A three-machine system is constructed by connecting one machine of the Ludington system to a two-machine/infinite bus system. Three IC units at different locations are also included. Results show that a single IC unit at the mode bus is most suitable for relocating that mode. More IC units may not improve the eigenvalues. For multi-mode relocation, however, use of the same number of IC units as the number of modes will provide best results. The results also suggest that a decentralized control scheme may be better.
Degree
Ph.D.
Subject Area
Electrical engineering|Energy
Off-Campus Purdue Users:
To access this dissertation, please log in to our
proxy server.