F-23 A Review of Predictive Models for Crude Oil Fouling

D. I. Wilson (Department of Chemical Engineering and Biotechnology, Cambridge University Technical Services, University of Cambridge, United Kingdom)

A review of open literature quantitative models that describe and predict the rate of crude oil fouling in heat exchangers suggest that modeling approaches can be classified into three strands: (i) fundamental phenomenological models, supported by detailed experimentation; (ii) semi-empirical models, notably of the “threshold fouling” type, which provide deterministic interpretation of measured or estimated initial fouling rates, and (iii) artificial neural network approaches, which offer several advantages in handling sparse and noisy data sets. Each model type is worthy of further pursuit with a concerted effort to identify the chemical mechanisms responsible for fouling.

This report provides a series of recommendations for improving the protocols in HTRI crude oil fouling tests. Focusing on initial fouling rates will facilitate shorter test runs.