A double wall tank is one vessel sitting inside another, with a gap between them. That is the whole of most catalogue descriptions, and it is the least useful part of the definition, because the gap is not a spare wall. The gap is a space with a job, and the job changes completely between one family of tanks and the next. In one family it is a sealed chamber you put a vacuum gauge on. In the other it is an open space you deliberately let liquid into. Both are sold as double wall tanks. They share a name, a drawing convention and almost no engineering.
That distinction is worth settling before anything else, because every question that follows depends on it: whether the outer shell is a jacket or a pressure boundary, whether “leak detection” means a gauge reading or a liquid level, and whether the containment percentage on the datasheet describes the gap or the assembly. A buyer who does not know which family he is looking at cannot tell a good proposal from a bad one, and the two proposals will use the same words.
This page works from the inside out. It starts with the two annulus philosophies and why they cannot be combined, then goes to the question almost nobody asks — what actually holds the inner vessel up — then to what the outer shell is asked to do on the day the inner shell fails, with the arithmetic, then to the sizing of the gap, then to air and rain, then to the detection that a double wall is supposed to make possible, and finally to when a single wall vessel with a bund is the better answer. The containment volumes your jurisdiction requires are a separate question with a separate page.
Key Takeaways
- A double wall tank comes in two incompatible types. A sealed annulus is a monitored chamber: a vacuum gauge watches it and any change means a leak. An open annulus is an equalising space: liquid enters it, the level falls, and the contents stay usable. Sealing an open annulus or venting a sealed one destroys the thing it was built for.
- The annulus holds nothing in normal service, which is why the outer shell is usually drawn thinner than the inner. On the day the inner fails, that shell has to take a liquid column — and 40 CFR 264.193(c)(1) already requires a containment structure to have strength for pressure gradients including static head.
- Because the gap is thin, equalisation barely lowers the level. In a 1.8 m vessel filled to 2.0 m, the level falls to 1.62 m and the gap ends up holding 19% of the working volume. The gap cannot contain the inner volume on its own: that would need an 8.5 m column in a 2 m tank.
- Worked at 1.8 specific gravity and 12 mm walls, the outer shell at that moment carries 28.6 kN/m against the inner’s 31.8 kN/m in normal service. The outer shell does about nine tenths of the inner shell’s normal duty, once, after a life spent dry.
- What holds the inner vessel up decides whether the annulus is one connected space. A continuous, seal-welded support ring divides it into compartments, and a sensor outside the ring will never see a leak from the inner base.
- A double wall tank does not contain an overfill. Liquid that leaves through the filling opening runs down the outside, past the annulus entirely. That is why an overfill sensor is part of the containment design, not an accessory.
Two Shells and a Gap — and Two Different Ideas About What the Gap Is For
The cleanest definition of a double wall tank in print is not a commercial one. It is in the hazardous waste tank rules, at 40 CFR 264.193(e)(3)(i), which asks for an inner tank completely enveloped within an outer shell, so that any release from the inner tank is contained by the outer shell. Read that carefully and a second conclusion follows: the same section, at (d)(3), lists a double-walled tank as one of the devices that satisfies secondary containment, alongside an external liner, a vault and an equivalent device approved by the Regional Administrator. A double wall tank is a bund that was built into the vessel instead of being poured around it.
That is a useful way to hold the object in your head, because it explains both what the tank is good at and what it cannot do. Everything a bund does, the gap does inside a smaller footprint. Everything a bund cannot do — keep the weather out, keep the contents recoverable, move with the tank — the gap can do because it is closed at the top. But a bund and a gap fail in the same place: at whatever crosses them.
What a Double Wall Tank Actually Is
Three words cover the whole product family. Double wall tank is the general term. Double walled tanks is the same object with a different adjective. Dual wall tanks is the American trade spelling, and it appears in the same search results because it is the same thing. None of the three tells you which annulus philosophy you are buying, and that is the only thing that matters.
The mechanical description is straightforward. There is an inner vessel, which is the one that holds the chemical and is therefore the one designed for the duty, the temperature and the specific gravity. There is an outer vessel, which is normally built to the same standard because it may one day hold the same chemical. There is a gap between them, and the gap has a width, a volume, a top closure and a low point. Everything this page is about lives in those four attributes of the gap.
The Sealed Annulus: Monitoring a Vacuum
The first family comes out of the fuel tank world, and its logic is that an empty space can be watched. The annulus is closed at the top and either evacuated or held under a slight pressure, and a gauge sits on it. If the inner shell cracks, the vacuum is lost and the gauge moves. If the outer shell is breached, the same thing happens. The gap is not a container; it is a sensor chamber, and the reading is the product.
This is a genuinely elegant arrangement, and it is why steel double wall tanks for fuel are usually described by their monitoring rather than by their capacity. It also has a consequence that buyers miss: because the gap is sealed and monitored, nothing can ever equalise through it. The inner vessel stays full, the outer vessel stays empty, and the tank must be taken out of service when the gauge moves. The contents are not recoverable in place, because they are still in the inner vessel and the inner vessel is the thing that is suspect.
The Open Annulus: Letting the Level Equalise
The second family comes out of chemical storage, and its logic is the opposite: an open space can be filled. The annulus is at atmospheric pressure, the top is shielded from rain by the inner vessel’s dome overhanging the outer shell’s wall, and the gap drains to a low point. If the inner shell cracks, liquid moves into the gap and the level falls until the two spaces agree. For an acid storage tank, that is the difference between a spill that reaches the floor and one that stays inside the unit.
Moulded polyethylene ranges describe this openly. One manufacturer’s answer to “how does the secondary tank contain the capacity of the primary tank” is that the contents leak into the secondary tank, the primary tank is still used as part of the containment basin, and filling stops once the liquid level equalises between the two. Another states that the two tanks equalise, and that the chemical can keep being used until the system is drained and inspected.
That is a real operational advantage and it should be stated plainly, because it is the strongest argument for the open family. A leak becomes a degradation rather than a shutdown. But it comes with the subject of the next two sections: an outer shell that was designed as a jacket now has to be a containing wall.
Why You Cannot Mix the Two
The two philosophies are not two ways of doing the same thing. They are answers to incompatible questions, and each one breaks the other.
Seal an open annulus, intending to keep the rain out, and you trap air in it. When liquid enters, the air compresses, the level stops well short of where it would otherwise have gone, and the containment volume you paid for is no longer there. Worse, the trapped air now pushes back on the outer shell, which was never designed to act as a pressure vessel. Vent a sealed annulus, intending to let it breathe, and the gauge that was the entire detection system stops meaning anything.
This is also why the familiar question — can you tell a double wall tank from a single wall tank by looking at it — has no reliable answer. A second rim, a slightly larger outer shell and a gauge are all hints. On a steel fuel tank the gauge is close to proof. On a moulded chemical tank the proof is a dome that overhangs the outer wall, and even that is a detail of one family, not a rule.
What Holds the Inner Vessel Up
A single wall tank never has to answer this question, because its base sits on the floor. A double wall tank has an inner vessel whose base is somewhere above the outer vessel’s base, and something has to carry the contents from one to the other. That something is never shown on a datasheet, and it decides three properties of the gap at once: whether the outer base is loaded in normal service, whether the annulus is one connected hydraulic space, and where its true low point is.
There are three ways to do it, and they behave differently enough that a double wall storage tank bought on capacity alone may not be the one you thought you were buying.
One-Piece Moulded: the Base Is Part of the Vessel
In a rotational moulding, the inner and outer shells can be produced as one integrated article, with the inner vessel’s base formed continuous with the outer shell’s base. The gap then exists around the sides and, depending on the mould, across the bottom as a shallow space bounded by the two bases.
This is the construction behind most moulded double wall poly tank ranges, and it produces a very clean inner vessel: no separate support, no ring to weld, no seating detail to seal. The question it leaves open is where the inner vessel actually bears. Both of the large moulded ranges describe an inner dome that overhangs the outer sidewall, and neither describes the seat. Where the seat sits matters, because a seat that is continuous around the inner base will trap anything that leaves the inner vessel through its floor in the pocket inside the seat, away from the low point you instrumented on the outside.
Fabricated: the Inner Shell Sits on a Support Ring
A welded double wall plastic tank cannot be moulded, so it is built as two shells. The inner shell is an ordinary welded vessel — the same vessel described in our polypropylene tank guide — and it has to be carried. The usual arrangement is a support ring: an annular plate welded to the inner shell’s base perimeter, standing on the outer shell’s floor, sometimes stiffened with gussets.
The ring is a short cylinder in compression carrying the full weight of the contents plus the liquid’s own load on the base. It is also, potentially, a dam. If the ring is continuous and seal-welded to both floors, the annulus stops being one space and becomes two: a pocket under the inner base, inside the ring, and a ring-shaped channel outside it. Those two pockets are hydraulically separate. A crack in the inner base fills the inner pocket; the channel outside stays dry; a level device in the channel reads nothing.
The fix is easy and has to be specified: interrupt the ring, or notch it, or provide a deliberate drain path, so that the whole annulus is one connected space with a single low point. Our fabricated double wall tanks are built with interrupted support rings for exactly this reason, and the phrase to put on a drawing is that the annulus shall form a single connected space, drainable to one point. For an open top tank design, where the inner vessel has to be entered from above, the same ring carries the same load and the same requirement applies.
Support-ring detail is fixed at the drawing stage, and the wider build order it sits in is described in fabricated construction sequence.
Nested: Two Separate Vessels, One Inside the Other
The third route abandons the integrated idea entirely. You take a complete inner vessel and stand it inside a complete, slightly larger outer vessel that has its own closed bottom and its own top. This is sometimes called a tank in a tub, and it has the cleanest load path of the three, because the inner vessel stands on the outer vessel’s floor on its own feet or stand, and the outer shell carries nothing but its own weight.
The annulus is continuous and open from the top, which makes inspection easy: you can see the whole channel, you can reach the low point, and you can add a device later. The compromise is the top. Two independent vessel tops mean the rain shield is an accessory bolted or moulded on afterwards, rather than a property of the inner dome, and a nested pair usually needs a purpose-made cover to reach the same weather protection a moulded design gets for free.
The Support Ring That Quietly Dams the Annulus
Put the three routes side by side and one line runs through them: the support detail is the difference between a containment space and two compartments that happen to share a wall.
This is not a detail a buyer is expected to draw, but it is a question a buyer can ask, and the answer is checkable. Ask where the inner vessel bears. Ask whether the annulus is one connected space below the liquid line you care about. Ask where the lowest point is and how liquid from the inner base gets to it. A supplier who can answer those three questions has thought about the gap. A supplier who answers with a containment percentage has answered a different question.
The Day the Inner Shell Fails: What the Outer Shell Has to Take
In normal service the annulus holds nothing. That is not a margin; it is a description of a structure at rest. The outer shell of a double wall tank, for its whole working life, carries its own weight, whatever wind and snow the site imposes, and nothing from the contents. It is, mechanically, a shell with no internal liquid load at all. Then, once, it may be asked to hold a column of the chemical it was built to contain.
The transition is the point of the whole product, and it is the thing a datasheet never quantifies. A double wall tank is bought for that one day, and on that day the outer shell stops being a jacket and becomes a containing wall for a liquid with a specific gravity, a temperature and a corrosion history it has never actually met.
The Annulus Carries No Head — Until It Does
Two states, one structure. In the first state the hoop tension in the outer wall is essentially zero: there is liquid on one side and air on the other, and the pressure difference is a few centimetres of nothing. In the second state there is liquid on the inside face all the way up to the equalised level, and the outer wall carries the same kind of hoop load that the inner wall carried, computed at a slightly different radius and a slightly lower height.
The rules anticipated this, which is worth knowing before a supplier tells you the outer shell does not need designing. 40 CFR 264.193(c)(1) requires a secondary containment system to be constructed of materials with sufficient strength and thickness to prevent failure owing to pressure gradients, and it names the pressure gradient explicitly: including static head and external hydrological forces. Since (d)(3) makes the double-walled tank one of the devices that constitutes secondary containment, that strength requirement attaches to the outer shell. The outer shell is not exempt from static head because it is normally empty.
Equalisation: How Far the Level Actually Falls
When the inner shell fails, the two spaces become one, and the level settles at a height that conserves volume. Call the inner cross-section Ai, the annular gap area Aa, and the working volume V. The settled level is V divided by the sum of the two areas. Because the gap is thin, the sum of the two areas is only slightly larger than the inner area — and that is the whole result.
Take a 1.8 m inner diameter vessel filled to 2.0 m, with a 100 mm annular gap. The inner area is 2.545 m², the gap area is 0.597 m², and the working volume is 5.089 m³. The settled level is 5.089 divided by 3.142, which is 1.620 m. The level has fallen 380 mm — 19% of the working height — and in that state the gap holds 0.967 m³, which is 19.0% of the working volume.
Read those two numbers together and the design rule falls out. The liquid that left the inner vessel is distributed across an area only 23% larger, so it does not go far down. And the volume sitting in the gap is a fifth of what was in the tank, not the whole of it. The containment in an open-annulus double wall tank is mostly the inner vessel, still full of the chemical, with the gap taking the difference.
Why the Outer Shell Has to Reach the Working Level
The settled level is 1.620 m and the inner vessel’s working level was 2.000 m. The difference is 380 mm, and that difference is the freeboard the outer shell has to have above the inner’s normal working level. It is not much, but it is not optional either.
If the outer shell’s top edge sits below the settled level, the liquid does not stop there. It runs over the top of the outer wall and down the outside of the tank, and what you have built is a bund with a weir in it. This is why the moulded designs put the inner vessel’s dome above the outer shell’s rim and let it overhang: the outer wall’s top has to be at or near the inner’s working level for the arithmetic to work, so the weather shield cannot be a separate lid sitting inside that height. It has to be the inner vessel’s own dome, drawn down over the outside.
Hoop Stress in the Outer Shell at the Moment of Failure
Put numbers on both states and the comparison is uncomfortable. Hoop tension in a vertical cylindrical shell is the liquid’s unit weight times the depth times the radius, T = ρgHR, evaluated at the base where the depth is greatest. Membrane stress is that tension divided by the wall thickness. Whether that figure is carried by a standing vessel or a lying one is a separate decision, and what the standing version asks of its base and access is in the vertical tank orientation behind it.
| State | Liquid depth (m) | Radius (m) | Hoop tension T (kN/m) | Membrane stress at 12 mm (MPa) |
|---|---|---|---|---|
| Inner shell, normal service, filled to working level | 2.00 | 0.90 | 31.78 | 2.65 |
| Outer shell, after full equalisation | 1.62 | 1.00 | 28.61 | 2.38 |
| Outer shell, if the gap fills to the inner’s working level | 2.00 | 1.00 | 35.32 | 2.94 |
At specific gravity 1.8, a wall thickness of 12 mm and the geometry above, the outer shell at the settled level carries 28.61 kN/m against the inner shell’s 31.78 kN/m in normal service. That is a ratio of 0.90. The outer shell is asked to do nine tenths of the inner shell’s ordinary duty, once, after a working life in which it has held nothing at all.
Two conclusions follow, and both of them change what a specification should say. The first is that an outer shell sized as a thin jacket on the reasoning that it never holds anything is sized on the wrong state. The second is about testing. A hydrostatic test of the assembly fills the inner vessel; our own practice tests at 1.5 times design head. For a 2.0 m working level that means a 3.0 m test column in the inner vessel — and it tells you nothing whatsoever about the outer shell, which is empty while it happens. The outer shell’s own design head is the settled level, 1.62 m, so its test is 2.43 m, and that test has to be run separately, in the annulus, at build.
Sizing the Gap — the Arithmetic Behind the Percentage
Every double wall tank proposal carries a percentage. The number is real, but it is almost always describing something other than what the buyer thinks it describes, so it is worth doing the arithmetic once and seeing where the percentage comes from.
Why a Thin Annulus Cannot Hold the Full Volume
Start with the strictest possible reading: the gap alone must contain everything in the inner vessel. The height required is the working volume divided by the gap area. For the 1.8 m vessel above, that is 5.089 m³ over 0.597 m², which is 8.53 m of liquid in a 2 m tall tank.
That is not a design choice; it is an impossibility. A thin annular gap cannot contain the inner volume without being four times taller than the vessel, and widening the gap enough to change that answer stops it being a gap. The sealed, vacuum-monitored family dodges the problem differently: its outer vessel is a genuine second vessel of comparable size, and the containment is the outer vessel’s own capacity, not a gap’s.
For the open-annulus family the conclusion is simply that equalisation is not an enhancement. It is the only way a thin gap can work at all, and every property that follows from it — the level fall, the outer shell’s load, the rain shield that must not seal — is downstream of that one geometric fact.
Worked Example: Three Vessel Sizes
Three vessels, each with a proportionally similar gap, worked through the same formulae. Inner diameters 1.2, 1.8 and 2.4 m; working levels 1.5, 2.0 and 2.5 m; annular gaps 75, 100 and 125 mm. The last column is the strict reading from the previous section, and it is there to show how far from it every real design sits.
| Inner diameter (mm) | Working level (mm) | Annular gap (mm) | Gap area ÷ inner area | Settled level after full inner loss (mm) | Liquid in the gap at that moment ÷ working volume | Joined capacity to the working level ÷ working volume | Gap alone would need (mm) |
|---|---|---|---|---|---|---|---|
| 1200 | 1500 | 75 | 26.6% | 1185 | 21.0% | 126.6% | 5647 |
| 1800 | 2000 | 100 | 23.5% | 1620 | 19.0% | 123.5% | 8526 |
| 2400 | 2500 | 125 | 21.9% | 2050 | 18.0% | 121.9% | 11406 |
Two things in that table are worth carrying away. The settled level is always 18 to 21% of the working height below the working level, which is small enough that the outer shell’s top detail has to be drawn with it in mind and large enough that it cannot be ignored. And the gap’s share of the working volume is always under a quarter, while the joined capacity above it is always in the 120s of a percent.
What 110 Percent by Equalisation Actually Requires
Now the percentage on the datasheet can be read properly. A figure like 110% is a statement about the joined system’s capacity up to a stated level — normally the outer shell’s rim — expressed against the inner vessel’s working volume. It is not a statement about the gap. In the table above, with the outer rim at the inner’s working level, the joined capacity is 126.6%, 123.5% and 121.9% for the three sizes. Every one of them clears 110%, and none of them has a gap that holds anything like 110% on its own.
Which is why the useful question to a supplier is not “is it 110%” but two questions with checkable answers. Show me the volume calculation to the stated top level. And show me the drawing that proves the gap is one connected space up to that level. There are three ways the second answer fails: trapped air in a sealed or partly sealed gap, which removes volume silently; a support ring that divides the gap into compartments; and a bottom profile that leaves a dry heel the liquid cannot drain out of. Each of them removes volume from the numerator while leaving the percentage claim untouched.
Whether 110% is the number your site needs is a different question, and not one this page can answer. Tank containment requirements come from more than one rule and they do not use one number; the two containment rule sets are set out in full there, with where each of them applies.
Freeboard and the Level That Must Never Be Reached
There is one failure the annulus does not catch, and it is the most common one. A double wall tank does not contain an overfill.
Follow the liquid. Filling happens through an opening in the inner vessel’s top. If the fill continues past the working level, the liquid rises in the inner vessel, reaches the filling opening and leaves through it — onto the inner vessel’s dome, and then down the outside of the tank. The outer shell’s top edge is below the dome, so the annulus is not in the path at all. The gap catches what leaves through the inner shell’s wall; it does not catch what leaves through the inner shell’s opening.
That is why the same manufacturers who sell the sealed annulus also sell high-level sensors and purpose-made overflow assemblies, and why some buyers still want a bund around one. The freeboard above the working level is not a safety factor for the annulus; it is the space you have to protect with instrumentation. Set the high-level trip below the filling opening’s invert, take any overflow connection through the wall rather than over the top, and treat “the tank will never be overfilled” as an assumption to be designed away rather than relied on.
Air, Rain, and an Annulus That Has to Breathe
An open annulus has to satisfy two requirements that pull against each other. It has to stay dry, because a gap full of rainwater has no containment capacity left and no way to tell a leak from the weather. And it has to let air in and out, because a gap that cannot breathe cannot equalise, and a gap that cannot equalise is not an open-annulus double wall tank at all.
Both requirements are met by one detail, and it is the detail that makes the moulded type recognisable from across a yard.
The Overlap Dome: Rain Shielding Without Sealing
On the moulded designs the inner vessel’s dome is drawn wider than the outer vessel’s sidewall and turned down over it, forming a lip or skirt. Rain running off the dome falls outside the outer wall. Snow cannot settle into the gap. Debris that lands on the dome stays on the dome. Both of the large moulded ranges describe this in almost the same words, which is a fair indication that it is the settled answer rather than one company’s preference.
The important word is shield. The skirt is not gasketed to the outer wall, and it should not be. It is an umbrella, and an umbrella with a gap under it lets air through. That gap is why the same component does both jobs: it sheds water because it overhangs, and it lets the annulus breathe because it does not touch.
Why an Airtight Annulus Cannot Equalise
Suppose the gap were sealed. Liquid enters from a crack in the inner shell; the air trapped in the gap has nowhere to go; it compresses; the level in the gap rises only until the compressed air balances the head above it. The gap is then holding a fraction of what its geometry says it holds, and it is holding it under pressure.
Neither consequence is acceptable. The first silently voids the containment volume the tank was bought for. The second loads the outer shell with an internal pressure that appears nowhere in its design case — a shell specified for static head on one face now has static head outside and compressed air inside, and 40 CFR 264.193(c)(1) does not contemplate that combination for a welded thermoplastic wall. A well-intentioned attempt to keep the rain out has made the tank worse at the only job it had.
Condensate, Thermal Cycling, and the Vent You Cannot See
An annulus is a narrow gap between two plastic walls, and it is not thermally inert. Day and night, the outer wall heats and cools faster than the inner, and the air in the gap moves in and out through the same clearance that lets the rain out. In humid conditions, air entering a cooler gap deposits water on the inside of the outer wall. Over a season, that water collects at the low point.
Condensate matters for two reasons. It is indistinguishable from a small leak if the only instrument is a level switch at the bottom, and a switch that trips on condensate every wet season is a switch that gets bypassed. And standing water in a gap that is supposed to be dry is a corrosion and biological load on two surfaces nobody can reach. So the low point needs to be inspectable and drainable as a routine, not only as a response to an alarm.
Heated tanks add a second case. A tank maintained above ambient pushes heat into the gap, and the air in a shielded gap expands. The mechanism that lets it out is the same overlap clearance, which is another argument for keeping the skirt loose: on a heated vessel a sealed gap becomes a pressure cycle, and a pressure cycle on a welded thermoplastic seam is a fatigue load.
When the Annulus Is Genuinely Closed
There is a legitimate design in which the annulus is closed, and it belongs to the steel fuel tank family. There the gap is sealed and monitored, the reading is the detection, and the closure is the point of the product. But a closed steel annulus and a closed thermoplastic annulus are not the same idea in different materials. The steel arrangement depends on holding a stable vacuum against a welded metal shell; a welded polypropylene shell does not offer that with the same confidence, and the rule under which a chemical containment arrangement is usually judged asks for detection of a release, not for a gauge reading.
So the practical split holds. For a chemical double wall tank, the normal arrangement is an open annulus, shielded at the top, drained and inspectable at the bottom. A closed annulus on a chemical tank is a question worth asking about rather than a feature worth paying for.
Making the Leak Visible: 24 Hours, and What Has to Be Drawn
The strongest statement in the containment rules is not about capacity. It is about time. 40 CFR 264.193(e)(3)(iii) requires a double-walled tank to be provided with a built-in continuous leak detection system capable of detecting a release within 24 hours, with a narrow escape only where the owner or operator can demonstrate to the Regional Administrator that existing detection technology or site conditions would not allow it. Paragraph (c)(3) says the same thing in operational terms: the detection system must be designed and operated so that it detects the failure of either the primary or the secondary containment structure, or the presence of any release, within 24 hours.
That converts detection from an adjective into a design deliverable. A double wall tank that could hold a leak but could not reveal it within a day does not satisfy the clause, however good its shell. And the things that make the difference are drawn at the bottom of the annulus, not specified in the sensor’s datasheet.
The One Penetration That Decides Everything
A vessel that has to be drained from below has to have a hole in its wall, and a hole in the inner wall of a double wall tank goes through the gap. The wall is cut to take a nozzle, the nozzle passes through the annulus, and the annulus is bridged by the one component that gets the most vibration, the most handling and the most bending in service.
The answer is to carry the containment through the penetration rather than to interrupt it. A contained transition — a double-walled fitting, a shrouded outlet, or a flexible bellows element that encloses the pipe where it crosses the gap — keeps the inner and outer walls joined at the penetration and routes anything that escapes at the joint back into the annulus. One manufacturer describes the failure mode of a plain fitting precisely: if the seal fails, the release is not contained by the outer shell, and the assembly behaves as a single wall tank.
The alternative is to avoid the crossing. A top outlet with a dip tube crosses only the dome, where the annulus does not exist. It is a legitimate answer for smaller vessels and for duties where a top entry is acceptable.
Why One Side-Wall Fitting, Not Two
There is a physical limit on how many low connections a double wall tank can have, and it comes from how the fittings work. A double-wall side fitting has to pull the inner wall and the outer wall together and seal them against each other, usually by a threaded or bolted body that clamps both. Tightening it puts the inner wall in local compression and bending, and the seal depends on both walls staying where the fitting put them.
One fitting is fine. Two fittings on the same vessel, at different azimuths, each pull the inner wall toward a different point on the outer wall. The inner shell is a thin cylinder, and pulling it in two directions at once sets up a bending pattern across it that neither fitting was designed for; one of the two seals will yield first. The published limit from one moulded manufacturer is exactly this: one lower side-wall fitting per tank, because a second one at 90 or 180 degrees would pull the tank in two directions and cause one or both fittings to fail prematurely.
The design consequence is to plan the whole outlet arrangement around a single lower penetration, and to put everything else — fill, vent, level, instrumentation, spares — in the dome. Where a process genuinely needs two low connections, the answer is two top outlets with dip tubes, not a second hole.
Low Point, Sight Tube, Float, or Sensor
Whatever the device, the requirement is the same: there must be a path from the most likely leak to the place where the device sits, and that path must go downhill without passing a barrier. Design it in this order.
The low point. Where is the lowest point of the connected annulus, and is it a real low point or a flat floor that will hold a film? A pocket is better than a flat, and a pocket with a drain boss is better than a pocket.
The observation. A vertical transparent sight tube from the low point to above the rim is the arrangement that needs no power and cannot be defeated by a flat battery, and it makes the annulus level readable in the same way a gauge glass reads a vessel. It requires the annulus to communicate with the tube, which is the same connectivity requirement as everything else on this page. A dip port with a dipstick is the low-technology version.
The device. A float switch or a conductance probe in the low point turns the same liquid into a signal, and a float or conductive sensor is the usual way to satisfy a 24-hour detection clause without relying on somebody walking past a sight tube. It needs a mounting boss on the outer shell — a penetration on the outside, where the liquid is on the other side of the wall in the gap and the fitting is dry in normal service, which is a much easier problem than the outlet.
The reading of the contents. A level instrument that reads the inner vessel without the chemical being outside the containment is worth naming, because it removes the temptation to put a conventional gauge on the inner wall and cross the gap to do it. A reverse-reading float or an ultrasonic instrument mounted through the dome achieves the same result without a side penetration.
What 40 CFR 264.193(iii) Actually Asks For
Read the clause word by word, because each word narrows the field.
Built-in. The detection is part of the tank system, not a procedure performed on it. A monthly dipstick check is not a built-in system, and it does not sit within the 24-hour limit either.
Continuous. The detection operates continuously, not on a schedule. This is where a sight tube scores: it is continuously reading, whether or not anyone is watching, provided somebody looks at it every shift.
Of either the primary or the secondary containment structure. A failure of the outer shell with a sound inner shell also has to be detectable. On an open annulus that is naturally satisfied, because a cracked outer wall shows as a falling gap level or as wetness at the outside face; on a sealed annulus it is the same gauge that reports both.
Within 24 hours. This is the number that governs the layout. Twenty-four hours is long enough for a slow seep to be visible in a pocket and short enough that the pocket must be reachable. It is also the same period that paragraph (c)(4) allows for the removal of what has been collected: spilled or leaked waste and accumulated precipitation must be removed from the secondary containment system within 24 hours, or as timely as is possible to prevent harm where a demonstration has been made.
The practical reading of all four together is that the load-bearing parts of the design are the pocket, the connectivity and the access. A sensor is the simplest item on that list and the easiest to substitute. A gap that is divided by a support ring into two compartments, one of which has no instrument, fails the clause on the day the wrong compartment fills.
What a Double Wall Does Not Tell You
Two answers from outside the tank trade are worth having, because both of them correct an assumption that is very widely held.
The first is about the risk management programme. A double wall tank is not passive mitigation. The Environmental Protection Agency states the position directly: passive mitigation is defined at 40 CFR 68.3 as equipment, devices or technologies that function without human, mechanical or other energy input, and while building enclosures, dikes and containment walls are passive mitigation, the double wall of a tank is not, because passive mitigation minimises the effects after the loss of containment whereas a double wall prevents the loss in the first place. The practical consequence is that a double wall cannot be used to shorten a worst-case distance in a hazard assessment.
The second is about protection from impact. In a 2002 interpretation of the construction standard at 29 CFR 1926.152(e)(4), which requires dispensing units to be protected against collision damage, the Occupational Safety and Health Administration concluded that the fact that a tank is double-walled would not by itself be sufficient to meet the requirement, because the equipment attached to the tank can be equally vulnerable and the hazard addressed is the collision rather than the leak. That is a construction-industry standard and should be read for its principle rather than applied generally, but the principle is sound and it generalises: a double wall tank contains what leaves the inner shell, and nothing else.
Choosing Between a Double Wall Tank, a Single Wall, and a Bund
By this point the decision has a shape. A double wall containment tank moves the containment inside the vessel’s own footprint, closes the top, and makes the leaked contents recoverable. A bund moves the containment outside the footprint, leaves the top open, and usually loses the contents. Neither is safer in the abstract; they fail differently and the question is which failure the site can live with. The same choice arises for water containment tanks as for chemical ones — the duty changes, the logic does not.
When a Single Wall Vertical Tank With a Bund Is the Better Answer
The bund’s great advantage is that its capacity is independent of the tank’s height. A bund can be sized to whatever volume the applicable rule asks for without changing the vessel, extended later, or subdivided when a second tank arrives. That is why large vertical single shell tanks — the term means single-wall upright vessels here, and not the nuclear-waste tanks the same phrase mostly retrieves — are almost always contained by a structure rather than by a second shell.
The bund’s second advantage is access. You can see into it, walk to it, clean it, patch it, sample what is in it. Its inner shell is invisible in service; the only thing you can inspect is what reaches the gap.
The bund’s third advantage is that it tolerates change. A tank inside a bund can be replaced, resited, upsized or joined by another, and the containment keeps working. It is a single object and its containment is sized once.
What a bund cannot do is keep the rain out. That is not a minor point: a bund with standing water in it has no capacity left, and the rules treat rainwater as a real load rather than a nuisance. 40 CFR 264.193(e)(1)(ii) and (e)(2)(ii), for external liners and vaults respectively, require the system to prevent run-on or infiltration unless the collection system has enough excess capacity for the precipitation from a 25-year, 24-hour rainfall event. The double wall answers the same question differently, by closing the top with the inner vessel’s dome. Two legitimate solutions to the same problem, and it is worth knowing that the rules give the open solution a specific rainfall allowance.
Footprint, Inspection, and the Leak You Cannot See
Set the two against each other on the four things that actually decide installations.
Space. A double wall containment tank occupies the footprint of one vessel. Both shells can be rectangular as well as cylindrical, and in a box the outer wall has to be stiffened as hard as the inner one; that structural side is in the rectangular plastic tank case. A bund occupies the footprint of the vessel plus the working space plus the wall. On an existing concrete pad with a fixed envelope, this often decides the question before any technical comparison starts, and it is the practical argument for double wall tank secondary containment: the containment takes no extra ground beyond the vessel.
Recoverability. What reaches a bund is usually contaminated by the bund — dirt, debris, rainwater, and any previous spill — and is normally disposed of. What reaches a closed annulus has been in a sealed, dry space and is, in the ordinary case, still usable. That is a real operational difference and both moulded manufacturers lead with it.
Movability. A double wall storage tank can be moved, and a bund cannot reasonably be. For a pilot plant, a leased building, or a process that is expected to change, that matters more than either of the first two.
Inspectability. The bund wins, decisively, and it is the factor most often left out of the comparison. It hides its inner shell permanently. The inspection you can actually perform is confined to the gap and to whatever instrumentation you put in it, which is the reason the previous section exists.
Retrofitting: What a Single Wall Vessel Would Need
Adding an outer shell to a single wall vessel is not a matter of building a jacket around it. The inner vessel would have to be lifted out of its support, re-supported inside the new outer shell, have its top joint remade to the new rim, and have every penetration rerouted with a contained transition — and the gap would have to be made inspectable, which usually means a removable section or a handhole that was not in the original design. In practice this is a new vessel.
What does work, and is worth knowing about, is the middle option: a single wall vessel placed inside a new outer vessel that is itself a complete tank. That is the nested arrangement described earlier, and it is how most retrofitted containment tanks are built. It converts an existing vessel into a contained one without touching its walls. It solves the containment question for a tank already in service, in exchange for a gap that is open at the top that needs its own cover.
Where a vessel is already installed and undamaged, the realistic choice is usually between a bund under it and a replacement with a double wall tank, and the deciding factor is normally whether the process can be shut down for the replacement.
What This Page Does Not Cover
Four things sit outside this page on purpose, and each of them is treated where it belongs.
The volume your jurisdiction requires is not a tank question. Containment volume comes from more than one rule, the rules do not use one number, and which one applies depends on what is stored and where; the two rule sets are set out there with their boundaries.
Material selection is a separate discipline. Polypropylene is the subject of its own guide, including how it compares with high density polyethylene for a given chemical at a given temperature.
Vessel geometry has two more entries. The cone bottom tank has its own angle, outlet and support problem, and the open top tank has the rim stiffness problem that comes from having no lid.
Anything that crosses the gap — outlet piping, fill lines, instrument pockets — is ancillary equipment in the language of the rules, and 40 CFR 264.193(f) gives it its own containment requirement. A double wall tank with a single wall outlet line is still a single wall tank with a second skin; the line is the hole.
Frequently Asked Questions
Is a double wall tank the same as a tank in a bund?
No, and the difference is the word “within”. A bund surrounds a tank at a distance, is open to the sky, and can be defeated by a leak high on the tank’s sidewall that clears the wall. A double wall tank encloses the inner vessel within an outer shell, so the outer is closed at the top by the inner vessel’s dome and rain can be kept out of the containment space. That is why what reaches an annulus is usually still usable and what reaches a bund usually is not.
Does a double wall tank still need secondary containment?
The double wall is the secondary containment. 40 CFR 264.193(d)(3) lists a double-walled tank among the devices that satisfy the containment requirement, alongside an external liner, a vault and an approved equivalent. What the double wall does not do is contain everything else that can release: an overfill leaves through the filling opening, and ancillary equipment such as piping is dealt with separately by 40 CFR 264.193(f). Whether a given tank satisfies the rule for a given site depends on the rule and on what else is connected to the tank.
How much volume should the annulus hold?
There is no single number, and the arithmetic on this page shows why. A thin gap cannot hold the inner vessel’s volume on its own — that needs a column several times the vessel’s height — so what a real design delivers is a capacity of the joined assembly up to a stated level. Ask for the volume calculation to that level and for the drawing that proves the gap is one connected space up to it. Which percentage applies to your site is a containment-rules question rather than a tank question.
Can you tell a double wall tank from a single wall tank by looking at it?
Not reliably, and the honest answer is usually no. A slightly larger outer shell, a second rim, a monitoring gauge or a dome that overhangs the outer wall are all hints and none of them is proof; a plain single wall tank with a weather cover looks like a double wall tank from a distance. That is why dual wall tanks and single wall tanks are routinely confused on site and why a survey of what a yard actually contains usually turns up at least one of each in the wrong column. The reliable answer is the drawing, which also tells you the thing you actually need to know — whether the gap is a monitored chamber or an equalising space.
Is a double wall tank passive mitigation under the RMP rule?
No. The Environmental Protection Agency states that passive mitigation is equipment, devices or technologies that function without human, mechanical or other energy input, and that while dikes, containment walls and building enclosures qualify, a double wall of a tank does not — because passive mitigation reduces the effects after a loss of containment whereas a double wall prevents the loss in the first place. A double wall therefore does not shorten a worst-case distance in a hazard assessment.
Can a double wall tank be installed outdoors?
Yes, and the moulded ranges are designed for it: the inner vessel’s dome overhangs the outer wall so that rain and snow run off outside the containment space. Two conditions go with that. The shield has to stay a shield, meaning the air path under the skirt must not be gasketed shut. And whatever does get in has to reach a low point where it can be seen and removed, otherwise a wet season and a small leak look the same.
Get a Double Wall Tank With an Annulus You Can Inspect
Three inputs settle most of a double wall tank. The first is the duty: the chemical, its specific gravity, its temperature, and whether the vessel is heated or outdoors. The second is the volume the gap has to hold and the level it is measured to, because those two together set the outer shell’s height and the freeboard above the inner vessel’s working level. The third is where the outlet goes, because a single lower side-wall penetration and a top outlet with a dip tube lead to two different vessels.
The vessels we build for this duty are PP chemical storage tanks with a fabricated or moulded annulus. With those three inputs we can set the support detail so the annulus is one connected space, place the low point and the instrumentation to match, size the outer shell for the settled level rather than for the impression that it never holds anything, and test the gap separately from the vessel. Send the duty, the chemical and the annulus volume and we will come back with a drawing, the settled level, and the inspection arrangement that goes with it.
We have been fabricating industrial plastic equipment since 2009, which is 17 years. Our vessels use 5 mm to 30 mm polypropylene sheet, hydrostatic testing at 1.5 times design head, ultrasonic inspection of thick-section welds and a working tolerance of ±2 mm. We operate more than 40 production bases and offices, and hold ISO 9001, ISO 14001, CE, SGS and RoHS. Bodies and welds carry a 12-month warranty.
Note on the figures — The areas, settled levels, hoop tensions and membrane stresses on this page are our own hand calculations from the formulae shown, on the dimensions stated in each table, and any reader can repeat them: inner area πD²/4, annular area π(D+2g)²/4 − πD²/4, settled level V/(Ai+Aa), hoop tension T = ρgHR with g taken as 9.81 m/s² and ρ taken as 1,800 kg/m³ for a 1.8 specific gravity liquid, membrane stress σ = T/t for the wall. They are order-of-magnitude illustrations of where the load goes, not design checks, and they carry no allowance for temperature, creep, forming, impact, out-of-roundness or the local effects of penetrations. No published standard specifies an annulus width, a settled level or a containment percentage for a double wall tank; the percentages in the table are geometric outputs of the stated dimensions, not requirements, and the regulatory quotations are reproduced as published in the sources below. Wall thickness, test pressure, tolerance and warranty are our manufacturing figures.
Sources
- U.S. Electronic Code of Federal Regulations — 40 CFR 264.193 — Containment and detection of releases (paragraph (c)(1): secondary containment systems must be constructed of or lined with materials compatible with the wastes and must have sufficient strength and thickness to prevent failure owing to pressure gradients including static head and external hydrological forces, physical contact with the waste, climatic conditions, and the stress of daily operation including stresses from nearby vehicular traffic; (c)(3): a leak-detection system designed and operated to detect the failure of either the primary or secondary containment structure, or the presence of any release of hazardous waste or accumulated liquid in the secondary containment system, within 24 hours or at the earliest practicable time on a demonstration to the Regional Administrator; (c)(4): spilled or leaked waste and accumulated precipitation must be removed from the secondary containment system within 24 hours, or as timely as possible where a demonstration has been made; (d): secondary containment for tanks must include one or more of a liner external to the tank, a vault, a double-walled tank, or an equivalent device approved by the Regional Administrator; (e)(1)(ii) and (e)(2)(ii): external liner and vault systems must prevent run-on or infiltration of precipitation unless the collection system has excess capacity sufficient to contain precipitation from a 25-year, 24-hour rainfall event; (e)(3): double-walled tanks must be designed as an integral structure, that is an inner tank completely enveloped within an outer shell, so that any release from the inner tank is contained by the outer shell; protected, if constructed of metal, from both corrosion of the primary tank interior and of the external surface of the outer shell; and provided with a built-in continuous leak detection system capable of detecting a release within 24 hours, or at the earliest practicable time on a demonstration to the Regional Administrator; (f): ancillary equipment must be provided with secondary containment meeting the requirements of paragraphs (b) and (c), with stated exceptions).
- U.S. Environmental Protection Agency — Double wall tanks and “passive mitigation” (passive mitigation is defined at 40 CFR 68.3 as equipment, devices, or technologies that function without human, mechanical, or other energy input; passive mitigation systems include building enclosures, dikes and containment walls, but a double wall of a tank is not considered to be passive mitigation, because passive mitigation serves to minimise potential adverse effects after the loss of containment whereas a precaution such as a double wall serves to prevent the loss of containment in the first place).
- U.S. Electronic Code of Federal Regulations — 40 CFR 68.3 — Definitions (passive mitigation means equipment, devices, or technologies that function without human, mechanical, or other energy input; active mitigation means equipment, devices, or technologies that need human, mechanical, or other energy input to function).
- U.S. Occupational Safety and Health Administration — Standard interpretation, 31 December 2002 — Use of double-walled tanks as the only form of protection from collision damage (29 CFR 1926.152(e)(4)) (while double walls reduce the chance of a tank rupture from a collision, they do not necessarily eliminate the combustion or explosion hazard from collisions; the location and design of equipment attached to the tank may render the system equally vulnerable irrespective of whether the tank is single or double-walled; in general the fact that a tank is double-walled would not by itself be sufficient to meet the requirement, and typically would be insufficient by itself, with the extent of other measures depending on the proximity of the tank to high volume or high-speed traffic and on whether vehicles back up to the tank; the interpretation applies only to states covered by Federal OSHA).
