Changing Crude Diets Opens the Risk of Fouling in Crude Preheat Exchanger Trains

Several Eastern European refineries have continued to process Russian export blend crude transported to them through the Druzhba pipeline, via an exception to EU sanctions imposed following the invasion of Ukraine in February 2022.

From June 5^th 2025, these refiners had to switch over, by EU law, to stop processing Russian crude for the first time in their history. Most of these refineries were designed and configured to use this crude in the most optimum way. Planning to diversify away to process alternative crudes is a major effort, requiring new logistics configurations/schedules, realignment of operations with the new yield profile of the alternative crudes, possibly with process reconfiguration and the adjustment to downstream operations and blending to account for changing product qualities.

‍

‍

Although there is margin for improvement, refineries exposed to the international markets have established processes to evaluate new crudes with tools, like linear program models, configured to carry out this process in the most time and cost-effective way. The Eastern European refineries have had to learn this process, rebuild their tools to provide the capability to make these crude evaluations and have had to train their staff and build new skills.Within this complex transition of operations, two risks remain and could be overlooked. These are:

• The risk of fouling in the crude unit preheat exchangers
• Corrosion in the crude overhead (chloride) and in the hot end of the main crude and vacuum towers (acid).

Focusing on fouling, it is not always easy to understand what is going on in a heat exchanger network. Often there is inadequate (or inaccurate) data to do a complete assessment exchanger-by-exchanger, meaning that assumptions have to be made to fill in the gaps. The fluid thermal properties can also change, especially for crude units where different crude slates (or bio-/synthetic crudes) are often processed.

These assumptions and measurement errors accumulate into the overall heat transfer coefficient or the calculated fouling resistance, making the calculated values change erratically.

The fouling behaviour is a complex issue – it can vary with feed rate, the feed mixture and different areas of the network can be affected at different times: the predominant fouling mechanism at the cold end is often particulate deposition; at the hot end it is often due to chemical reaction fouling, causing asphaltene precipitation. The evaluation of fouling and heat exchanger performance using traditional simulation-type technology is time consuming for the unit engineers, with the results that it becomes aperiodic exercise, at best monthly, more likely two or three times per year –it is therefore probable that specific fouling events relating to certain new crudes or crude mixtures or operating conditions will only be detected after the event and too late to enable corrective action to be taken.

Since fouling causes under-performance of exchangers it leads to loss of energy efficiency and higher pressure drop. Crude units can be or become thermally limited (furnace operating at maximum duty, draught or tube metal temperatures) or hydraulically limited (maximum booster pump discharge pressure, motor limits) – sometimes both; sometimes the situation can transition from thermal to hydraulic limit as fouling progresses.

Once facing these issues due to fouling, operators have several choices:

1. Fire the furnace harder to compensate for the thermal inefficiencies, incurring higher fuel costs and producing more airborne emissions, but also placing more thermal stress on the mechanical structure.
2. Reduce the feed rate to handle the higher pressure drop across the network.
3. Clean the exchangers, with (depending on the status of the network) a possible reduction in throughput and loss of energy efficiency for the duration of the cleaning activity.

Cleaning exchangers is also not a trivial exercise – the maintenance work needs to be planned, cranes ordered to pull the bundle and contractors organised to do the work. The earlier this can be planned, the more effectively it can be executed and the shorter the duration of the slowdown for cleaning.

But what if you could maintain close to the start-of-run capacity and energy efficiency / emissions across the cycle between major turnarounds and not worry if you have chosen a specific crude or crude mixture that has a higher fouling tendency?

That is the problem that we are solving at Hexxcell. The Hexxcell Studio™ platform consists of a hybrid physical/AI digital twin model of the heat exchanger network – this is built from the actual mechanical configuration of the various heat exchangers. The Hybrid-AI is trained on the performance of the network over the past cycle and accounting for the configuration of the various exchangers, it identifies bad data and fills the data gaps (i.e.reconciles or imputes as necessary) to match key performance parameters (i.e. temperatures, pressures and flow rates) of the overall network system. And the model does this in autonomous way near real-time, linked to live data from the plant historian with a big saving of engineering time and effort. This enables a much closer tracking of where and when fouling is occurring, as it is happening, providing alerts for mitigating actions to be taken and enabling the linkage of the fouling event to specific feed mixtures or operations.

The intelligence gained from the digital twin can then be fed back into resetting operating limits, crude scheduling and crude blending limits and into the important crude selection and purchasing process to build up a reliable history of the fouling tendencies of these new crudes.

Hexxcell Studio™’s functionality does not end there. Some fouling is going to happen irrespective – the key is to understand at what point does the cost of extra fuel and emissions plus any loss of throughput justify a cleaning event, and which exchangers should be cleaned and when. Hexxcell Studio™ predicts the optimum cleaning schedule for the heat exchanger network based on the future production plan (crude mixes, rates) to ensure that exchangers are cleaned before the financial loss from fouling becomes too high and enabling maintenance to plan the cleaning events well in advance.  

With the digital twin model it is also possible to retrospectively assess the impact that network changes (new exchangers, change of exchanger type, exchanger tube inserts) would have had on overall performance over the last cycle, using real data, to provide a more reliable business case for revamping preheat trains, to enable them to be better aligned with the changing yield profile of the new crude slate.

Thus, Hexxcell Studio™ helps reduce the risks of fouling for any refiners, but particularly these Eastern European refiners who are looking to diversify crude slate or process ‘new’ crudes, by:

1. Avoiding higher energy costs, higher airborne emissions and additional stress on the mechanical structure of the furnace.
2. Avoiding loss of throughput due to higher heat exchanger network pressure drop.
3. Identifying specific crude mixtures / ratios that have a higher fouling tendency to build that intelligence into future crude purchasing decisions and crude operations scheduling.
4. Evaluating the business case for changes to the configuration of the heat exchangers to more optimally process the new crudes.

Would you like to know how our platform can help you diversify crude slate while mitigating fouling? Contact us to schedule a demo and discover how Hexxcell Studio™ can optimise your operations, reduce energy losses, and increase profitability.