Civil & Structural

LNG tanks: single, double and full containment

What the containment categories actually mean, which parts of a full containment tank are the civil engineer's and which are not, and why the outer concrete wall is a structure that has to work twice — once every day, and once in a condition that should never happen.

IntermediateOil & GasPetrochemical

Standards referencedEN 14620ACI 376API 620NFPA 59AOISD-STD-118

An LNG tank is two tanks, and the argument about which kind of tank you are buying is really an argument about what the second one is for.

Looking steeply down into the narrow annular space of an LNG tank under construction from the top of the outer wall. The curved grey concrete wall face fills the left of the frame with starter bars projecting from its head, and the metallic inner tank shell curves away on the right. The gap between them is partly filled with pale pink-white granular perlite insulation. Far below, six workers in orange hi-vis vests and white hard hats stand on multi-level scaffold decks and access platforms spanning the annulus. Overcast daylight.
The annulus of a full containment tank. Concrete on one side, 9% nickel steel on the other, perlite going in between. Everything a civil engineer owns on this job is on the left of this photograph.

The categories, and what actually separates them

EN 14620 and API 620 describe four arrangements. The distinction between them is not size or cost but a single question: what does the outer structure contain?

Four tank cross sections side by side. Single containment shows an inner steel tank inside an outer steel shell that holds vapour only, with a remote earth bund drawn around it. Double containment shows an inner steel tank inside an open-topped concrete outer wall that holds liquid but not vapour. Full containment shows an inner nickel steel tank inside a prestressed concrete outer wall closed by a concrete dome, holding both liquid and vapour, with thermal corner protection and a secondary bottom marked at the base. Membrane shows a thin stainless steel membrane on insulation bearing directly against the concrete wall, with no free-standing inner tank. Beneath each section a row states what the outer structure holds and whether a bund is required.
The outer structure is the only thing that differs. Everything else — the base slab, the heating, the piles — is common to all four.
Type Outer structure holds Bund needed?
Single containment Vapour only Yes — a remote earth bund sized for the full contents
Double containment Liquid, not vapour No separate bund; outer wall is the bund
Full containment Liquid and vapour No
Membrane Liquid and vapour, with no free-standing inner tank No

The civil content rises sharply down that list. A single containment tank is a steel job with a large earthworks contract attached. A full containment tank is a prestressed concrete structure with a steel tank inside it.

What a full containment tank is made of

Working from the inside out:

The inner tank — an open-topped cylindrical shell of 9% nickel steel, free-standing on the base. Nine per cent nickel is used because it keeps its fracture toughness all the way down to −196 °C; it has no ductile-to-brittle transition in the service range, which ordinary carbon steel very much does. This is the primary container. It is a mechanical scope.

The annular insulation — loose perlite filling the gap between the inner shell and the outer wall, with a resilient blanket against the inner shell. The blanket matters more than it looks: the inner shell expands and contracts as the tank fills, cools and empties, and without something springy to absorb that movement the perlite would settle into the gap on each cycle and then refuse to let the shell move back, squeezing it.

The outer wall — prestressed concrete, typically 800 mm thick, post-tensioned circumferentially and vertically. This is the secondary container and it is the civil engineer’s principal structure. Post-tensioned containment walls covers how it is designed and built.

The vapour liner — a thin carbon steel plate on the inside face of the concrete wall, typically 5 mm, anchored back into the concrete. Its only job is gas-tightness. Concrete is porous and it cracks, so it cannot be relied on to hold vapour; the liner can. It is carbon steel rather than nickel steel because in normal service it sits at ambient temperature with warm vapour on both sides.

The thermal corner protection and secondary bottom — insulation at the junction of wall and base, carried over by a steel secondary bottom, so that in a leak the liquid stands on insulation rather than on cold concrete.

The suspended deck — an aluminium deck hung from the dome on rods, carrying a thick layer of glass wool. It is not structural. It holds the roof insulation in place and separates the cold vapour space below from the warm steel above.

The dome roof — a steel liner dome with structural concrete cast on top of it.

The base slab — covered separately in LNG tank foundations and base mats.

The leakage case: the whole reason the concrete exists

In normal operation the outer concrete tank does almost nothing. It carries its own weight, the roof, the internal vapour pressure, wind and seismic load. A competently designed silo would do as much.

Its reason for existing is the leakage case, also called the spill case or the major leak scenario: the inner tank has failed and its entire contents are now standing against the inside of the concrete wall.

Three things arrive at once:

  1. Hydrostatic pressure from a full depth of liquid at 450 kg/m³ against a wall that was previously carrying only vapour pressure.
  2. A thermal shock, as the inside face goes from ambient to −162 °C. The contraction strain this wants to impose is more than ten times the cracking strain of concrete — see what cold does to concrete.
  3. A loss of material properties in the reinforcement at the inside face, as it passes through its ductile-to-brittle transition.

And the wall must stay liquid-tight through all three, because the point of full containment is that nothing gets out.

This is what the post-tensioning is for. It puts the concrete section into compression so that the thermal and hydrostatic tension can be absorbed without the section ever going into net tension and cracking through.

Air raising the dome

The roof is worth a paragraph because it is the most distinctive operation on the job.

The dome steel liner is welded together as a shallow spherical cap lying on the tank floor. When it is complete and gas-tight, the tank is sealed and fans raise the internal air pressure by a few kilopascals — a fraction of an atmosphere, but acting over the area of an 80 m diameter dome it is several thousand tonnes of lift. The dome rises up the inside of the tank, guided, over a few hours, and lands on the compression ring at the head of the wall, where it is welded off.

The structural concrete is then cast on top of the liner, which serves as permanent formwork.

The alternative is falsework inside a tank 80 m across and 40 m tall. The air raise replaces weeks of scaffolding with a single morning, and it is one of the few genuinely spectacular operations in civil construction.

Where the disciplines meet

The interfaces are where integrity problems start, and on an LNG tank there are four worth knowing:

  • The compression ring at the wall head, which takes the dome thrust. Concrete, heavily reinforced, with the steel dome welded to an embedded plate.
  • The wall-to-base junction, carrying the thermal corner protection, the secondary bottom and the vertical tendon anchorages all in the same congested metre.
  • The penetrations. In a full containment tank, nothing penetrates below the liquid level — all pipework enters through the roof, and the in-tank pumps sit in columns from above. A bottom penetration would be a hole in the secondary container.
  • The base heating system, cast into the slab and inspectable for its whole life. See sub-soil heating and frost heave.

What to take away

  • The categories differ only in what the outer structure contains, and full containment holds both liquid and vapour — which is why it needs no bund and why it is now standard.
  • The inner tank is the primary container. The concrete is secondary, and in normal service it is nearly idle.
  • The leakage case is what the concrete is for, and it arrives as pressure, thermal shock and material degradation simultaneously.
  • No penetrations below liquid level. Everything comes in through the roof.

Next: post-tensioned containment walls for the wall itself, and secondary containment and bunds for what happens when the outer structure is earth rather than concrete.

Check your understanding

6 questions. Nothing is recorded — this is just for you.

1What distinguishes a full containment tank from a double containment tank?
2In a full containment tank, what is the primary container?
3Why does the inside face of the outer concrete wall carry a carbon steel liner?
4What is the function of the perlite in the annular space?
5Why is thermal corner protection fitted at the base of the outer wall?
6The concrete dome roof of a full containment tank is commonly built by which method?

#civil#lng#containment#cryogenic#storage-tanks#prestressed-concrete