OPTIMIZATION OF A SYSTEM USING STATISTICAL CONTROL THEORY
Date
1970-01
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
ORCID
Type
Degree Level
Doctoral
Abstract
The earlier methods for finding the optimum settings of the variable control parameters of a speed governor of a hydro-electric turbine employed arbitrary performance indices and assumed deterministic system disturbances. In this thesis, the Generalized Mean Square Error, the GMSE, is used as a performance index. The GMSE is a function of the correlation functions of input and output and is ideally suited for situations where the disturbance is non-deterministic. A method is developed to describe the random input disturbance statistically and hence to evaluate the input autocorrelation. The other correlation functions required for the evaluation of the GMSE are then derived using the input autocorrelation function and the frequency response function of the system under study. The model used in this study is a particular example of a non-minimum phase, time invariant feedback control system with a transport delay subject to a random disturbance. It is an electric turbo-generator supplying an isolated load. The turbine is provided with the conventional mechanical type governor. The trends in the variations of the GMSE resulting from changes in the statistics of the input disturbance and in the control parameters are shown.
Description
Keywords
Citation
Degree
Doctor of Philosophy (Ph.D.)
Department
Electrical Engineering