PROCESS CONTROL SYSTEM IDENTIFICATION

• Main Author: MUSTAFA, NOR FAEIZAH
• Format: Final Year Project
• Language: English
• Published: Universiti Teknologi Petronas 2006
• Subjects: TK Electrical engineering. Electronics Nuclear engineering
• Online Access: http://utpedia.utp.edu.my/7257/1/2006%20-%20PROCESS%20CONTROL%20SYSTEM%20IDENTIFICATION.pdf; http://utpedia.utp.edu.my/7257/

System identification is a method for generating workable dynamic response models based on an observed dataset from an actual system. It is used to give the input-output relationship of the dynamic response. The objective of this project is to design and implement System Identification for Liquid System Pilot Plant. The project will also make comparisons between the conventional and intelligent modeling technique. The project concentrates on the conventional technique known as empirical modeling and intelligent modeling by means of System Identification Toolbox. In empirical model building, models are determines by making small changes in the input variable about a nominal operating condition. The model developed by using this method provides the dynamic relationship between selected input and output variables. Matlab providesthe SystemIdentification Toolboxthat helps to ensure the observedtest data represents the dynamics of the system under investigation. It provides tools for creating mathematical models ofdynamic systems based on the observed input-output data. For the intelligent technique, two model predictors, ARX and ARMAX, are used to obtain the best model. From the analysis, it shows that the ARX models exhibit quite the same characteristics as the models obtained from the empirical technique. By using the System Identification Toolbox, the ARMAX structures are the best models in representing the actual system. After model validation tests, all models from both the conventional and intelligent technique are capable of reproducing observed data with minimum predictive error.