Black-box modeling of steam distillation pilot-scale using M-level PRS perturbation

Abstract

This paper is proposed to model the steam temperature of steam distillation pilot-scale using system identification. Multi-level Pseudo Random Sequence (M-level PRS) has been implemented to this system to perturb the input of the system. The linear and nonlinear Auto Regressive with Exogenous input (ARX) model structures are used to estimate and validate the model output of steam distillation pilot plant. Both models will be compared to study the performance and flexibility. The validation test is performed by using auto-correlation function (ACF), cross-correlation function (CCF), model fit and Root Mean Square Error (RMSE) to validate the estimated both ARX models.