Nested NARIMA Model of the Atmospheric Distillation Column

Abstract

Atmospheric distillation column plays a very important role in the crude oil processing. It is multivariate, strongly nonlinear process. Its operation and proper control impacts significantly the overall refinery performance. One may find in literature several papers describing column modeling with full spectrum of approaches staring from the first principle models up to regression black-boxes. Presented model composes of subsequent sections associated with technology. Each section is modeled as the nested NARIMA model. This methodology has been previously tested in several other chemical applications. It is extended by decomposition and coordination of subsequent sub-models in the considered case. Such a structure enables to obtain process nonlinear model with clear technological meaning of all considered elements. Further, it may be directly used in the process of control philosophy design. The procedure is illustrated with real data originating from the industrial installation.