No description available.

No description available.

Call number: LD2668 .T4 1961 B45

Call number: LD2668 .T4 1965 F78 / Master of Science

published_or_final_version / Electrical Engineering / Master / Master of Science in Engineering

published_or_final_version / Electrical Engineering / Master / Master of Philosophy

M.Ing. / Previous studies on the four-pulse dynamic compensator showed the existence of a stable limit cycle in the compensating current. Attempts were made to eliminate this limit cycle by placing a small signal active filter in the feedback path of the system, as it influenced the efficiency of the compensator. The efficiency of the active filter depends, on system parameters such as the impedance of the supply network and the load. Thus the small signal active filter eliminated the limit cycle only conditionally. A control method referred to as error modulation, was developed in order to eliminate the limit cycle unconditionally. According to this method the amplitude of the error is used to determine the modulation depth. Results showed that this method completely eliminates the limit cycle. In order to analyse this method analytically, a dual-input describing function for the altered non-linearity was derived. It was possible, by using the describing function, to explain the stability of the system with error modulation. Tests on the system revealed a poor dynamic response. The undamped nature of the system decreased the efficiency of the system. Damping was introduced to the system in such a way, that losses were kept at a minimum, while the dynamic response of the system improved. It has also been experienced that the losses in the system tends to use all the stored energy in the inductor. Previous attempts to stabilise the amount of stored energy in the inductor,did not succeed. 'n Control system using feedback was thus developed to stabilise the energy in the inductor, and the dynamic compensator losses are drawn from the supply.

Thesis--M. Sc., University of Hong Kong. / Mimeographed.

Thesis (M. Phil.)--University of Hong Kong, 1973. / Mimeographed.

M. Tech. Electrical Engineering / Reconfiguration of electric distribution networks is treated as an optimisation problem of a combinatorial nature, where the main goal is to achieve a path in the network with minimal power losses and with a good computation time. There are various ways of reducing active power losses in distribution networks, where network reconfiguration is one of the most important methods. However, it can be a very difficult optimisation problem. The Genetic Algorithm with Rank Weight selection method is therefore proposed. The method is used to search a population of switches for the optimum solution in the network.