Fouling in pre–heat trains of refinery crude distillation units causes major energy inefficiencies,
resulting in increased costs, greenhouse gas emissions, maintenance efforts and health and safety
hazards.
Although chemical and physical phenomena underlying fouling deposition are extremely
complex and several details remain unknown, the understanding of the fouling process has
progressed significantly in the past 40 years. However, this knowledge has so far not been
exploited to effectively improve heat exchanger and heat exchanger network design and operation.
As a result, old methodologies that neglect the local effects and dynamics of fouling, in favour of
lumped, steady–state, heuristic models (e.g. using TEMA fouling factors) are still used.
In this thesis a novel mathematical model for pre–heat trains undergoing crude oil fouling
was developed, validated with plant data and used to propose mitigation strategies. The model is
dynamic, distributed and considers simultaneously several scales of investigation. Key phenomena
are captured at the tube level as a function of local conditions. These include the dependence
of fouling rate on temperature and velocity, the variation of physical properties, the structural
changes of the deposits over time (ageing) and the dynamics of surface roughness.
The single tube model was then extended to describe a unit–scale heat exchanger geometry.
This has been validated against plant data from four units in two refineries operated by major
oil companies. The predicted outlet temperatures over extended periods (i.e. 4-16 months) are
accurate within ±1% for the tube–side and ± 2% for the shell–side. Model simulations were then
used to assist the retrofit of one particular unit for which it was possible to save ca. 22% of the
energy losses (not including pumping power) produced by fouling over ca. a year of operation.
Finally, the interconnection of single heat exchangers in a network allowed the simulation of
the fouling behaviour of two existing pre–heat trains. To systematically assess the impact of fouling
on refinery economics, a set of key performance indicators (KPIs) was proposed. Network–level
simulations were used in conjunction with the KPIs to unveil complex interactions and propose
network retrofit arrangements that improve energy recovery over time whilst reducing fouling.
It is concluded that the model can be used with confidence to predict fouling and assist
monitoring, design and retrofit of refinery heat exchangers and heat exchanger networks. The
results shown indicate that the approach proposed can lead to substantial benefits.