The design of periodic excitations for dynamic system identification
McCormack, Anthony Sean (1995) The design of periodic excitations for dynamic system identification. PhD thesis, University of Warwick.
WRAP_THESIS_McCormack_1995.pdf - Requires a PDF viewer such as GSview, Xpdf or Adobe Acrobat Reader
Official URL: http://webcat.warwick.ac.uk/record=b1400560~S15
System identification techniques are developed for modelling linear and nonlinear systems. The main results of the work are concerned with the design and utilisation of periodic perturbation signals in general areas of time- and frequency-domain system identification. A design strategy is given for a new class of perturbation signals, together with examples of their use in system identification applications. Signal processing procedures are developed for the practical treatment of drift disturbances and transient effects, and also for the detection of nonlinear contributions to the measurement data. The techniques rely completely on the periodicity of the excitation, and so the advantageous properties of periodic input signals are considered in detail. The use of periodic excitations in discrete- and continuous-time nonlinear system identification is also reported, with the identification methods illustrating the worth of frequency-domain measurements in this area. An automatic tuning procedure for PID controllers is also developed, which illustrates an application of system identification techniques to control problems.
|Item Type:||Thesis or Dissertation (PhD)|
|Subjects:||Q Science > QA Mathematics
T Technology > TK Electrical engineering. Electronics Nuclear engineering
|Library of Congress Subject Headings (LCSH):||System identification, Perturbation (Mathematics), Signal processing, Electronic excitation|
|Institution:||University of Warwick|
|Theses Department:||School of Engineering|
|Supervisor(s)/Advisor:||Godfrey, Keith ; Flower, John|
|Sponsors:||Engineering and Physical Sciences Research Council (EPSRC) (GR/H14267)|
|Extent:||xxii, 167 p.|
Actions (login required)