Why a Shorter Kettle Bundle Can Need a Larger Shell: Length, Diameter and Which Area Binds

Entrainment Prediction in Kettle Reboilers, Why the Element Is Not Free to Size, and What CFD Is Actually For
11/15/2026
Reboiler Mist Eliminators in Ethylene, C3 Splitter Heat Pumps and Refrigeration Chillers
11/15/2026
Entrainment Prediction in Kettle Reboilers, Why the Element Is Not Free to Size, and What CFD Is Actually For
11/15/2026
Reboiler Mist Eliminators in Ethylene, C3 Splitter Heat Pumps and Refrigeration Chillers
11/15/2026

Compact heat exchangers are sold on being shorter, lighter and cheaper. In a propane pre-cooling chiller that is usually true for the train, but it is not uniformly true for the vessels, and the exception is the part engineers get challenged on: on some stages the diameter grows, and on one stage in the reference case it grows by a factor of more than two while the tube length falls by a quarter.

This page explains why a shorter bundle can force a larger shell, how to check which constraint is binding before the design is fixed, and what the resulting geometry costs in money, propane inventory and transport.

What the reference case actually shows

The LNG2023 pre-cooling paper compares a conventional low-fin chilling train at 3 K approach with an enhanced-tube design using in-vessel elimination, at the same duty and the same approach, for a high, medium, low and low-low pressure propane stage:

Stage Conventional: diameter / tube length Compact: diameter / tube length Effect on diameter
HP 1600 mm / 6.7 m 3900 mm / 5.0 m grows 2.4 times
MP 4000 mm / 15 m 3800 mm / 9 m falls 5%
LP 4000 mm / 16 m 3600 mm / 9.5 m falls 10%
LLP 3200 mm / 13.3 m 3500 mm / 7.1 m grows about 9%

Read the table as an outcome of one train design, not as a set of ratios to copy. The pattern that matters is structural: two of the four stages got shorter and fatter, two got shorter and slightly slimmer, and the overall train was still about 40 per cent shorter and 45 per cent lighter. Compactness of the train and slenderness of each vessel are different questions.

Two area requirements, and only one of them can win

Every kettle has to satisfy two independent area requirements, and the shell ends up sized by whichever is larger:

  • Heat transfer area. Fixed by duty and by the overall coefficient. With enhanced tubes the coefficient per unit area is higher, so the required area is smaller, and the bundle can be shorter.
  • Disengagement area. Fixed by the vapour volumetric flow and the allowable velocity of the element. It is a cross-section requirement, and it does not care how long the vessel is.

Those two numbers move in opposite directions when you change the tube type. The duty stays the same, so the mass flow of vapour stays the same, but the bundle that generates it becomes shorter. Heat is then released into the same shell volume in less length, so the vapour produced per metre of shell rises, while the cross-section needed to let it separate is unchanged. Shortening the bundle does not shrink the dome; it concentrates the same vapour into a shorter vessel.

Where the conventional vessel was long and thin, that concentration has to go somewhere. The dome was already close to its limit, so the shell has to become fatter. Where the conventional vessel was already wide, the dome had spare capacity and the shorter bundle simply leaves more room, so the diameter falls slightly.

That is the whole mechanism, and it explains the table above without needing a per-stage narrative: the compact design is not a smaller vessel, it is a vessel with a different aspect ratio.

Why the low pressure stage is the hardest to shorten

The volumetric flow of vapour rises steeply as the boiling pressure falls, because the vapour density falls with it. On the coldest propane stage the density can be a tenth of the high-pressure stage value, so for the same mass flow the volumetric flow is an order of magnitude larger, and the disengagement area grows with it.

The practical consequence is that on the low-pressure stages the dome is already the binding constraint before any compacting is attempted. Cutting the tube length there does not buy diameter, because the diameter was never set by the bundle. A proposal that claims “40 per cent shorter everywhere” is therefore either ignoring the element area or quietly accepting a higher velocity in the disengagement space.

Which is why the sizing check has to be done per stage rather than on the train average: compute the area the heat transfer needs, compute the area the vapour needs, take the larger, and only then look at the length.

The other reason the high-pressure stage gets fat

There is a second effect worth naming, because it is often missed in cost comparisons. For a fixed duty in a shorter bundle, the tube count has to rise, and the bundle envelope grows with roughly the square root of the tube count. A kettle whose conventional version was a long thin bundle of comparatively few tubes can become a short bundle with a large cross-section. The shell then contains the bundle, the element, the freeboard to the nozzle, and the space to set the level lower than a flooded kettle would allow.

Those four items add up. The result is a short, wide pressure vessel: less steel in total, because the length dominates steel mass, but a larger plan area, a larger head and a larger shell diameter to fabricate, inspect and ship.

What the fat vessel costs

Item Short and fat Long and thin
Total weight lower, typically 40 to 45% on the vaporiser reference
Propane inventory higher if the diameter grows and the level is not lowered accordingly lower per vessel
Transport and lifting diameter may exceed the road or crane limit; becomes a project constraint length usually easier to move than diameter
Foundation and plot larger footprint per vessel, shorter train overall longer train, smaller footprint each
Internals element area easier to achieve, flow distribution harder element area tight, distribution simpler

Two of these deserve emphasis. Transport limits are hard limits: a shell that cannot leave the fabrication shop by road changes the modularisation strategy of the whole project, and that is a decision made at the diameter, not at the weight. And propane inventory cuts against the safety argument that justifies compact designs in the first place, so a design that reduces weight by growing the diameter has to also reduce the level to keep the inventory claim honest.

How to check before the layout is fixed

  1. Compute the heat transfer area per stage from duty and the overall coefficient of the actual tube type, not from a catalogue ratio.
  2. Compute the vapour volumetric flow per stage at the boiling condition, using the saturated vapour density at the design pressure.
  3. Compute the disengagement area from that flow and the allowable velocity of the chosen element, at the element’s installed height and at the top of the level band, not at nominal level.
  4. Compare the two diameters, add the element, the freeboard and any deflector, and take the larger of the two requirements plus the internals allowance.
  5. Then check the transport envelope, because that is the constraint that cannot be designed around later.
  6. Repeat for 70 and 110 per cent load, since the low-pressure stages are the ones where a turndown excursion moves the vapour density most.

If that sequence is followed, the answer is defensible even when it is counter-intuitive: the stage that got 2.4 times wider is not a mistake, it is the stage where the dome was already tight and the length reduction had to be paid for in cross-section.

What this means for a specification

Three lines belong in the datasheet because they force the conversation:

  • State the allowable velocity and the basis for it (element type, orientation, K, and the density pair used), so that a change of element can be evaluated rather than argued.
  • State the level band and the element installation height above the shell centre line, so the chord-limited area is visible in the documents.
  • State the transport envelope, so the diameter has a ceiling before the thermal design starts optimising.

For the other reboiler services where the binding constraint is fouling or approach temperature rather than the bundle, see reboiler mist eliminators in ethylene, C3 splitter heat pumps and refrigeration chillers.

Reference

Diameter and tube-length data, the approach-temperature comparison and the train-level weight and length results follow Knoepfler, Buffet, El Hajal, Lang, Provost, Rambure and Redjem Saad, LNG2023, Vancouver, paper 3.2.1. The stage-by-stage mechanism above is an engineering reading of that data, offered as a method to verify per project rather than as a property of the paper.

Frequently Asked Questions

Why does a shorter tube bundle sometimes need a larger shell?

Because the vapour mass flow is set by the duty and stays the same, while the length that generates it becomes shorter. The same vapour is produced in less shell length, so it is concentrated, but the cross-section needed to let it separate is unchanged. Where the conventional vessel was long and thin and already close to its disengagement limit, the only place for that concentration is a larger diameter.

Can a chilling train be 40 per cent shorter without any vessel getting bigger?

Not on all four stages. In the reference case the high-pressure stage grew from 1600 mm to 3900 mm in diameter while its tube length fell from 6.7 to 5 m, and the low-low pressure stage grew about 9 per cent. The medium and low pressure stages did get smaller. The train as a whole was shorter and lighter, which is a different claim from every vessel being smaller.

Why are the low pressure stages the hardest to shorten?

Vapour density falls with boiling pressure, so the volumetric flow for the same mass flow rises steeply at the cold end, often by an order of magnitude. The disengagement area follows the volumetric flow, so on those stages the dome is already the binding constraint and the tube length was never what set the diameter.

What breaks first if the diameter grows too far?

Transport. A shell that exceeds the road or crane envelope cannot leave the fabrication shop as a module, which changes the modularisation strategy of the entire project. Weight is usually recoverable through design; a diameter that breaches the shipping limit is not, so the envelope should be set as a ceiling before the thermal optimisation begins.

Does a lighter vaporiser mean less refrigerant inventory?

Only if the level is also handled. Inventory follows shell volume, so a design that cuts weight by growing the diameter can increase the propane charge unless the liquid level is set lower, which is exactly what enhanced boiling surfaces make possible because the bundle needs to stay wetted rather than flooded. Weight and inventory have to be reported together.

Leave a Reply