Vacuum distillation units rarely lose value in one dramatic event. More often, margin slips away as column pressure rises, heater limits tighten and wash section performance drifts between turnarounds. By the time the narrower operating window shows up in the monthly numbers, the unit has already given back yield that an earlier intervention could have protected.
A vacuum distillation unit (VDU) takes atmospheric residue from the crude unit and recovers vacuum gas oil (VGO) under reduced pressure. Lower pressure lets those heavy distillates vaporize below the temperatures at which they would crack thermally. According to the U.S. Energy Information Administration, about 80% of U.S. refineries operate a vacuum distillation unit. In those plants the VDU's gas oil yield sets how much feed the conversion units receive, which gives the unit a large influence on refinery margin.
Every tonne of atmospheric residue that stays in vacuum residue instead of being recovered as VGO is value the unit does not capture. Decisions about cut depth and steam rates carry straight into refinery operations and downstream conversion economics.
Recovering that value means balancing flash zone pressure, coil outlet temperature, wash section integrity and ejector performance at the same time. These variables interact nonlinearly, degrade at different rates between turnarounds and shift with every change in crude slate. Conventional control works from fixed linear models and fixed targets. VDU economics depend on how the constraints tighten together over the run.
How flash zone conditions set VGO yield
Flash zone temperature, flash zone pressure and stripping section performance each set part of the operating window, and every adjustment to one changes the constraints on the others. That is why VDU economics rarely respond to single-variable moves.
- Flash zone temperature. Higher temperature increases vaporization and VGO yield, but it is bounded by a hard ceiling that protects the heater coils and the residue pumparound.
- Flash zone pressure. Lower pressure also increases vaporization, but it depends on ejector and condenser capacity that may already be near its limit.
- Stripping steam. Steam lowers hydrocarbon partial pressure and allows deeper cuts, but it loads the overhead ejector system and adds to the unit's steam and energy demand.
Deep-cut operation tightens the rest of the unit. Heavy vacuum gas oil (HVGO) quality becomes harder to hold as cut points extend. Wash section integrity matters more, and the margin between productive operation and coking narrows.
At the board, these trade-offs become judgment calls every shift: how hard to push cut points when the wash zone has already been stressed, how much steam to add when the ejectors are showing fatigue, and when to pull back to protect the unit. Those calls vary by crew, by shift and by recent operating history. Yield variation between shifts is one result, and it is a meaningful source of lost margin.
Why VDU performance degrades between turnarounds
VDU performance usually degrades because several limits tighten together. Limits that would each be manageable on their own interact, and each compensating move adds load somewhere else in the unit.
From preheat fouling to wash section stress
Preheat train fouling develops after startup. As exchanger effectiveness falls, the heater has to fire harder to hold coil outlet temperature. Coke deposits raise heater tube skin temperatures, and operators may then lower firing severity or coil outlet temperature to protect tube metallurgy. Lower coil outlet temperature means less vaporization in the flash zone, so stripping steam is often increased to compensate, which puts more load on the ejector system.
The wash zone comes under stress next. Reduced coil outlet temperature and variable overflash create uneven liquid distribution across the wash bed, and dry spots can form where coking begins. Operators may raise overflash to protect HVGO quality, but that adds load to the rectification section and narrows flexibility further.
When several limits bind at once
Later in the run, vacuum tower pressure often creeps up as ejector performance weakens. At that point the unit can be pinned against coil outlet temperature, HVGO quality and column pressure at the same time. A crude blend change on already-degraded equipment tightens the window further, and the effect carries across the crude oil refining process well beyond the VDU.
Where conventional VDU control reaches its ceiling
The unit behaves nonlinearly, while conventional model predictive control uses fixed linear step-response models built during commissioning. Flash zone behavior, vapor-liquid equilibrium and wash bed response do not move linearly as feed and operating conditions shift, so the mismatch between model and process grows as the run goes on.
A conventional model does not adapt on its own as crude slates change and equipment degrades. It becomes less representative of the real process gains, timing and interactions. By mid-cycle the recommended moves can be directionally right but quantitatively off, and operators end up overriding the controller more often than the original deployment assumed.
The CDU and VDU coordination gap
The boundary between the crude distillation unit (CDU) and the VDU matters just as much. Many refineries run the two units as separate advanced process control (APC) applications, but the economic problem crosses that boundary. Shifting separation load between atmospheric and vacuum service, or balancing heater duty across both units, needs coordinated action across a shared set of constraints.
Limits of steady-state real-time optimization
Real-time optimization adds economic targets, but those targets still rest on steady-state assumptions. Targets that are valid for one steady state may not hold under the plant's actual dynamics. For a VDU that keeps moving with crude changes and equipment degradation, that limitation leaves value unrecovered.
No technology substitutes for the pattern recognition that comes from decades at the board. A VDU with several moving constraints can still exceed what any operator or linear controller can balance continuously across every variable.
How a learned model adapts as the unit changes
A model built from the unit's own operating data learns the nonlinear relationships that fixed linear models miss: how flash zone conditions, wash section behavior, ejector performance and downstream product quality move together. When crude properties shift, the model can adjust targets based on similar past transitions instead of waiting for a re-tune. Instead of moving one variable at a time and watching the response, it coordinates setpoint changes across the interacting variables inside the unit's safe operating envelope.
Coordinated decisions across the CDU and VDU
A shared model across the CDU and VDU also changes how teams work together. It can show whether shifting separation load from atmospheric to vacuum service captures more total value under current conditions, and how that decision interacts with distillation yield further downstream. Planning teams and process engineers see how cut point decisions affect downstream economics. Maintenance priorities become easier to set when degraded equipment has a visible yield cost. That visibility reduces friction between groups that have each pulled toward their own targets instead of the whole system.
Advisory mode, then closed loop
Deployments that build trust usually start in advisory mode. The model recommends setpoint changes and operators decide whether to act. Senior operators test the recommendations against their own judgment, while less experienced operators see what changed and why a particular trade-off made sense under those conditions. Planning and economics teams use the same model to run scenarios on cut depth, steam rates and crude blend changes before they reach the board, and process engineers track how exchanger and ejector performance evolve.
As confidence builds, plants move to closed loop. The controller writes setpoints within boundaries the operations team defines and can change, and operators can override it. For a VDU, where compounding degradation erodes value between turnarounds, that continuous adaptation keeps the unit near the best point current limits allow.
Frequently asked questions
Why does crude slate variability make VDU optimization harder?
The flash zone does not respond the same way to every feed, even when crude properties look similar on paper. Flash zone temperature, pressure and stripping steam interact, so a crude switch mid-run on degraded equipment can tighten the operating window further. Single-variable adjustments then lose effect, because one correction often creates a new limit somewhere else in the unit. Heavier feeds make this more pronounced.
How does ejector performance affect VGO yield?
The ejectors set the pressure the flash zone works against. When ejector performance weakens, effective tower pressure rises, feed vaporization falls and VGO recovery drops. The effect is rarely isolated. Ejector limits usually combine with heater constraints, wash section stress and HVGO quality, which is why yield recovery depends on balancing the whole unit instead of treating vacuum performance as a separate issue.
Can this work alongside existing APC on a VDU?
Yes. The DCS and APC keep handling regulatory and multivariable control, while the learned model takes on economic target-setting and coordination across interacting constraints. That matters once current crude properties, equipment condition and cross-unit economics no longer fit the assumptions built into fixed linear models. Plants usually build trust in advisory mode before moving to closed loop.
How the Imubit Platform applies
In refining, Imubit applies Map, Model and Control to the crude and vacuum units. Map traces VGO yield lost between turnarounds to the recurring decisions behind it. Model uses Deep Learning Process Control® (DLPC®) to learn a nonlinear dynamic model of the VDU and its links to the CDU from the unit's own history, guided by first principles. Control takes cut depth, steam and pressure decisions to closed loop, writing to the existing APC or DCS, inside limits set by plant engineers.
