Auto-regressive with exogenous input (ARX) modeling on heating process in steam distillation essential oil extraction system / Nurlaila Ismail

This thesis presents a new ARX model of heating process in steam distillation essential oil extraction system by using system identification approach. Two experiments were carried out; the first is a simple direct heating system and the second is an automation heating system based on the steam distillation essential oil extraction (SDEOE) plant. The experiments were performed at the pilot plants in the Distributed Control "System (DCS) Laboratory in Level 5, Faculty of Electrical Engineering, UiTM. A computer and a data acquisition were used to log the inputoutput data from the experiments. The MATLAB version R2006a was used to estimate design ARX model and the model estimate is done by using linear regression method. In order to validate the estimate model, several methods were used such as Multiple Correlation Coefficient or model fitting (R value), One-Step Ahead Prediction (1-SAP) with its residual and correlation tests of residual which comprises of Autocorrelation Function (ACF) and Crosscorrelation Function (CCF). It is noticed from both experiments that partial data of the whole input-output data is sufficient to produce an adequate ARX model.