Inferential estimation of kerosene dry point in refineries with varying crudes

Zhou, C; Liu, Q; Huang, D; Zhang, J

doi:10.1016/S1570-7946(09)70046-X

Inferential estimation of kerosene dry point in refineries with varying crudes

Downloads

Full text for this publication is not currently held within this repository. Alternative links are provided below where available.

Abstract

A bootstrap aggregated model approach to the estimation of kerosene dry point in refineries with varying crudes is proposed in this paper. Using on-line measurements of process variables, the feed crude oil is classified into one of the three types: with more light component, with more middle component, and with more heavy component. Bootstrap aggregated neural networks are used in developing the on-line crude oil classifier. Historical process operation data are classified into these three groups. A bootstrap aggregated partial least square (PLS) regression model is developed for each data group corresponding to each type of feed crude oil. During on-line operation, the feed crude oil type is first estimated from on-line process measurements and then the corresponding bootstrap aggregated PLS model is invoked. The overall inferential estimation performance of the bootstrap aggregated PLS estimator integrated with feed crude oil classifier is significantly enhanced. © 2009 Elsevier B.V. All rights reserved.