Jul-2015

Managing fouling in refinery networks

Modelling heat exchanger and other networks enables the use of dynamic fouling models in performance assessment and predictive studies

SIMON PUGH and EDWARD ISHIYAMA
IHS Energy

Article Summary

Fouling of crude and product streams in oil refineries is an inescapable problem that causes very significant financial, environmental and safety problems. For example, refinery preheat train (PHT) networks nearly always suffer from fouling, resulting in reduced throughput, increased energy consumption and more frequent shutdowns. The financial, environmental and safety impact of fouling on processing 73 million barrels of crude oil per day worldwide is enormous. It is estimated that crude fouling in PHTs costs 0.25% BOE of all refined crude, or 66 million barrels per year.

Prediction of fouling is an increasingly challenging task, with laboratory fouling studies able to assess blend compatibility but often little more than a first approximation of real plant performance. Refinery networks such as PHTs often exhibit fouling that builds up relatively slowly and is often managed by exchanger cleaning or improved exchanger design, as well as rapid fouling events. It is imperative for effective refinery operation that rapid events are flagged up quickly and this can only be done where the fouling inside individual shells can be inferred correctly from plant measurements. These measurements are often inadequate for spreadsheet-type analysis, where the geometries of the individual exchangers are in any case over-simplified.
 
Pragmatic framework
A comprehensive research programme into crude oil fouling was completed in 2010 in the UK. This programme was facilitated by IHS and guided by the majority of the super-major oil companies plus several technology providers. The group identified that a pragmatic framework to apply the research knowledge to networks such as PHTs was needed. SmartPM (Smart Predictive Maintenance) is the software tool that fulfils this purpose, and is now assisting refinery operators, engineers and managers to make decisions on how to alleviate fouling in existing and new plant.

In addition to greatly improved reconciliation of monitoring data that can flag up rapid fouling episodes, the detailed heat exchanger modelling in SmartPM allows dynamic models for crude blend and product stream chronic fouling to be inferred from the reconciled data. These models allow operators to compare capex-heavy solutions (revamps and so on) with opex routes (such as regular cleaning). Using detailed (rigorous) thermohydraulic exchanger modelling and dynamic fouling models, SmartPM models correctly the behaviour of networks of heat exchangers. The wide range of generic and proprietary heat exchanger configurations, including arrangements such as helical baffles and tube inserts, is included.

Within a state-of-the-art, simple-to-use graphical interface, SmartPM is a turnkey application that connects operation, maintenance, engineering and fouling research, with each group accessing the appropriate facilities of the software for the same refinery model. For example, whilst operation and maintenance run data reconciliation and cleaning scheduling, engineering costs the benefits of revamp options, researchers review the fouling propensity of various crude slates, perhaps in a range of refineries. All share the same goal, which is maximising profit through increased efficiency and reduced maintenance.

SmartPM is a heat exchanger network simulation software aimed at tackling fouling related problems encountered in oil refining. The software couples thermohydraulic network analysis with dynamic fouling simulations, enabling the user to reliably predict plant operational limitations, economic penalties and environmental impacts resulting from heat exchanger fouling. The software was designed (but not necessarily limited) to model heat exchanger networks in crude oil refineries.

Missed opportunities
Many refineries use spreadsheet based models or other approximate models to evaluate plant operational data and assess heat transfer efficiency. These rudimentary models are often used for candidate selection in exchanger cleaning. There are many missed opportunities when using these types of model as they fail to resolve key issues:
• Many heat exchanger networks are poorly instrumented, with important temperatures and even flow rates missing, or at least temporarily unavailable.
• Intermediate temperatures between shells in series are rarely measured yet temperature fields in these shells are very different; hence the fouling in the shells differs significantly.
• An increasing number of refineries are now using efficient baffle types such as Helixchangers and EMbaffles; tube inserts such as Turbotal and Spirelf are in frequent use.
• Online exchangers compensate for loss of overall heat duty when other exchangers are taken offline for cleaning.
• Cleaning hottest shells can lead to a loss of preheat in upstream shells as the temperature of streams (for example, streams such as vacuum residue are often bridged between exchangers downstream and upstream of the desalter, where temperature variation propagates throughput the network when associated units are cleaned).
• Increase in pressure drop with fouling can result in throughput reduction.
• Hot end exchangers can experience  crude-side boiling, which can lead to throughput reductions.

SmartPM resolves these and other issues by using full network simulation, detailed exchanger modelling and dynamic fouling models in all modes of operation from data reconciliation to cleaning scheduling.

Application to preheat trains
A key application of the software is in the modelling of the crude preheating system. Fractional distillation of crude oil is an energy intensive process. The process requires the crude to be heated from ambient temperature to around 700°F (370°C). The required heat is provided through a network of heat exchangers called preheat trains (PHTs) and furnaces. PHTs recover heat from the product and pump-around streams of the distillation column. A simple example of a PHT based on an Argentinian refinery is shown in Figure 1.1 Crude is pumped through a set of heat exchangers and enters a desalter. The desalter washes the crude with water to remove the inorganic, water soluble impurities in the crude. The vapour component of the crude is removed through a flash column. Crude is further heated downstream of the flash through a set of heat exchangers where it enters the furnace at the furnace inlet temperature (FIT). The crude is then heated in the furnace to the furnace outlet temperature (FOT), prior to entering the fractional distillation column.

Crude oil is a complex mixture of petrochemicals and impurities and is prone to fouling.^1,2,3 US refineries are reporting new fouling problems that are a direct consequence of the increasing amounts of domestic tight oils in the crude slate. These problems arise from the changing nature of the slate (one refinery for example is seeing swings upwards of 5-10 API points on a day-to-day or week-to-week basis), additives in the crudes and crude incompatibility issues.

Many tight oils have added corrosion inhibitors (usually amine based) and drag reducing agents. This can lead to operating issues in the preheat exchangers and furnaces, and in the atmospheric distillation towers where the amine salts are deposited, inhibiting heat transfer and fouling the tower. Increased deposition of sand and particulates are observed in heat exchangers upstream of the desalter. In usual PHT operation, chemical reaction fouling dominates the fouling mechanism downstream of the desalter. Operational irregularities such as inefficient desalting can lead to inorganic deposition. Such problems are exacerbated as tight oils stabilise emulsions in the desalter and increase heat exchanger fouling in downstream units Desalting efficiency can be further challenged by wax precipitation in the cold train exchangers.⁴