Adaptive granular control for HVDC transmission system

Considerable progress in control technology has been reported on solving a rapidly growing range of problems with a decision system. This occurs especially in those niches of control engineering where the classical approaches have been faced with some difficulties or have appeared to perform inherently weakly. In a non-liner and complicated plant system like High Voltage Direct Current (HVDC), it is very difficult to mathematically and accurately model its behaviors. It is a good approach to address this problem by measuring the input and output, and establish a relationship between the input and output. However, the relationship between input and output is not always one-to-one. When we apply, any control to such a system, accurate control parameters are not easily obtainable by calculations. The classical PI control, which has been widely used in HVDC system, is one of these examples.; In this thesis, we study the application of rough set theory and granular computing techniques to the control of the HVDC systems. We propose an adaptive rough-set-based control scheme in place of the classical PI control. On the one hand, adaptive rough control deals with the dynamics and complexity of responses from a HVDC system in the operation points, by adjusting its control parameters adaptively. On the other hand, granular computing handles the issue of the unavailability of an accurate frequency domain model, which is usually required by adaptive control but is hardly established in a HVDC system. Our study includes a brief introduction to rough sets, rough control, and granular computing, both theory and application.; We also evaluate in the thesis the performance of adaptive rough control by simulation. The focus of our experiments is on the constant current control. The result shows there are many improvements offered by the rough control scheme in comparison with the conventional HVDC control scheme (i.e. classical PI control).