Date of Award
2004
Document Type
Master Thesis
Degree Name
Master of Engineering (Research)
Department
Electronic Engineering
First Advisor
Dr. Joe Connell
Abstract
The field of system identification and process parameter estimation is an area that has developed rapidly during the past few decades. The approach to the system identification problem can be influenced by many different aspects which include the specific purpose of the identification. In this thesis, system identification for the purpose of controller design is the central theme.
Initially, the work will focus on a number of approaches to real system identification. Particular attention will be made to the classical non-parametric techniques and the more modern parametric estimation techniques using both open- and closed-loop real-time data. A voltage controlled dc motor plant will be used to help benchmark the various techniques.
The work will then progress to the area of control relevant identification, which aims at achieving optimal control performance for a given system. The inter-relation between system identification and controller design is examined. Results from the dc motor plant will show that it is beneficial to use a number of controller design techniques to determine the best closed-loop response and hence the most suitable combination of model and controller.
The knowledge gained and lessons learnt from the motor plant will then be applied to the main target of the work which is system identification and control of a pilot-scale milk pasteurisation plant. A comparison between non-parametric identification and both open- and closed-loop parametric identification methods will show that the non-parametric transient step response technique produces the most accurate and the simplest model. Also highlighted is that reduced plant model complexity produces significant improvements in both open- and closed-loop performance.
Cascade feedback control simulations are also performed as a further addition to the single feedback control of the pasteurisation. Results show vast improvements in the overall control of the plant and the plant output performance.
Recommended Citation
Murphy, Karolyn, "Modelling and System Identification for Real Systems" (2004). Theses [online].
Available at: https://sword.cit.ie/allthe/150
Access Level
info:eu-repo/semantics/openAccess