UK industrial valve sourcing, automation packages and practical application support

ALLWAY VALVESIndustrial valve supply & selection Request a Quote

Technical article

Breather Valves and Conservation Vents: Tank Venting Design, API 2000 Sizing and Selection

  • An API 650 atmospheric tank may be designed for as little as 2.5 kPa (10 in. w.c.) of internal pressure and a fraction of that in vacuum — a correctly sized breather valve is frequently the only thing preventing a collapsed roof or ruptured shell.
  • API 2000 / ISO 28300 sizes normal venting from pump rates plus thermal breathing: the classic rule of thumb allows roughly 0.94 Nm³/h of in-breathing air per m³/h of pump-out (the old 5.6 SCFH-per-barrel rule) and up to 2.02 Nm³/h per m³/h of fill for volatile products.
  • A weight-loaded pallet typically needs 75–100% overpressure above its set point to reach rated capacity — size the vent for full flow at the maximum allowable tank pressure, not at the set pressure, or you will quietly under-vent the tank.

Breather valves — also called conservation vents or pressure vacuum relief valves — are the least glamorous devices on a tank farm and among the most important. An atmospheric storage tank built to API 650 is a thin-walled membrane structure: a 40 m diameter tank with a 6 mm roof plate will buckle under a vacuum that a domestic vacuum cleaner can generate. Every pump-out, every cold rainstorm on a sun-heated roof, every nitrogen regulator failure is a potential implosion or overpressure event. This guide walks through how breather valves work, how to size them to API 2000, and where specifications most often go wrong.

Why Atmospheric Tanks Need Breather Valves

A pressure vessel to ASME VIII shrugs off a few bar of swing in either direction. An API 650 tank does not: the standard tank is designed for essentially atmospheric service, with Annex F extending internal pressure to at most 18 kPa (2.5 psig), and the vacuum rating of an unstiffened roof is often just 0.25 to 0.6 kPa. That means the margin between normal operation and structural damage is measured in millibar, not bar.

Tank venting therefore has two jobs that pull in opposite directions. The vent must open reliably at tiny differential pressures to protect the shell, yet stay shut tightly enough during normal operation to conserve product vapour, limit VOC emissions, and keep air out of tanks storing flammable or oxygen-sensitive liquids. An open goose-neck vent does the first job perfectly and the second not at all — which is exactly why the conservation vent was invented. On gasoline service a conservation vent cuts standing evaporation losses dramatically, and on a blanketed tank it keeps the nitrogen bill sane.

How a Conservation Vent Works: Pallets, Weights and Springs

The heart of a pressure vacuum relief valve is the pallet: a disc, usually guided by a central stem or side rods, resting on a machined seat. Tank pressure acts on the pallet area; when the force exceeds the pallet weight (or spring load), the pallet lifts and vapour escapes. A second pallet lifts when atmospheric pressure exceeds tank pressure, handling the vacuum side; most breather valves combine both functions in one body.

Weight-loaded pallets

Weight loading is the default below about 60 to 70 mbar set pressure. It is beautifully simple — set pressure is adjusted by stacking calibrated weights on the pallet — and it has no spring to corrode or drift. The catch is lift characteristic: a conventional weighted pallet opens proportionally and only achieves its rated flow at something like 75–100% overpressure, so a valve set at 10 mbar may not pass full capacity until the tank sees 17.5 to 20 mbar. Modern full-lift designs using an air-cushion seat snap to full opening at around 10% overpressure, which lets you set the valve much closer to tank design pressure.

Spring-loaded pallets

Above roughly 70 mbar, the stack of weights becomes impractically tall and heavy, and spring loading takes over, typically covering set pressures up to 0.5 bar and beyond on API 620 tanks. Springs also make sense where vibration or tank movement would bounce a weighted pallet, and on pipe-away valves where orientation matters. The trade-off is that springs introduce a steeper force-versus-lift curve, so check the manufacturer’s actual flow curve rather than assuming full lift at set. Either way, insist on stainless pallets and seats as a minimum — carbon steel internals in a condensing vapour space are a corrosion claim waiting to happen.

Normal Venting: Pump Movements and Thermal Breathing

API 2000 / ISO 28300 splits venting requirements into normal and emergency cases, and the distinction drives everything downstream. Normal out-breathing is caused by liquid pumped into the tank displacing vapour, plus vapour expansion as the tank contents and vapour space warm up. Normal in-breathing is the mirror image: pump-out draws in air (or blanketing gas), and atmospheric cooling — classically a sudden rainstorm on a hot roof — contracts the vapour space with alarming speed.

Thermal in-breathing is the case that kills tanks, and it is routinely underestimated. The 7th edition of API 2000 calculates it as a function of tank volume raised to the power 0.7, multiplied by a latitude-dependent C-factor between 3 and 6.5 and a reduction factor for insulation. For a large tank, the thermal term can exceed the pump-out term several times over. If someone sizes your vacuum vent on pump rate alone, send the calculation back.

API 2000 Sizing of Breather Valves: Method and the 5.6 Rule

Generations of engineers sized breather valves with the legacy rules from the 5th and 6th editions of API 2000, and they remain a useful sanity check. In-breathing: 5.6 SCFH of free air per barrel per hour of maximum emptying rate — in SI terms, roughly 0.94 Nm³/h of air per m³/h of pump-out. Out-breathing: about 1.01 Nm³/h per m³/h of fill for products with a flash point at or above 37.8 °C, doubling to 2.02 Nm³/h per m³/h for volatile products below that flash point, to account for flashing and evaporation on top of simple displacement.

The 7th-edition general method replaces the flat thermal allowances with the volume-exponent formulas above, and for many tank sizes it demands materially more in-breathing than the old tables. My advice: run both, take the larger, and document which edition and which method you used — auditors and insurers increasingly ask. Then convert the required air flow to actual valve capacity using the manufacturer’s certified flow curves at your allowable overpressure, never a bare orifice calculation. API 2000 requires capacity to be established by test, precisely because pallet valves do not behave like ideal nozzles.

One more subtlety: capacity is quoted in air at standard conditions, but your vapour is not air. Correct for molecular weight and temperature — methanol vapour and gasoline vapour do not vent identically through the same pallet. For blanketed tanks, remember that out-breathing must also pass the maximum inflow of blanketing gas if the regulator fails wide open, which is often the governing case.

Emergency Venting: The Fire Case

Emergency venting covers external fire exposure, where heat input boils the tank contents and generates vapour far beyond anything a normal breather valve handles. API 2000 calculates heat input from the wetted surface area of the shell — with the familiar 0.82 exponent above 18.6 m² of wetted area — moderated by environmental factors for drainage, insulation and water spray. Divide the heat input by the latent heat of vaporisation and you have a required venting rate that typically dwarfs the normal case by one or two orders of magnitude.

That capacity comes from dedicated emergency vents — large weighted manway covers or hinged relief hatches — or from a frangible roof-to-shell joint designed to API 650, which counts as emergency venting in its own right on larger-diameter tanks. Do not try to make the breather valve do this job; a 250 mm conservation vent cannot pass fire-case flow at 20 mbar, full stop. And if the tank has a frangible roof, verify it genuinely qualifies — small tanks often cannot achieve a frangible joint no matter what the drawing says.

Nitrogen Blanketing and the Pressure Ladder

On blanketed tanks, the breather valve, the blanketing regulator and the tank design pressure must be arranged into a coherent ladder of setpoints. A typical arrangement: blanketing valve maintains the vapour space at, say, +3 to +5 mbar; the pressure pallet is set at +15 to +20 mbar so it stays shut in normal operation; the vacuum pallet sits at −3 to −5 mbar, below the regulator’s control band, so nitrogen — not air — satisfies routine in-breathing. Squeeze these setpoints together and the valve weeps nitrogen continuously; I have seen sites lose thousands of pounds a month through a pallet set 2 mbar above the blanketing point.

Critically, the vacuum side of the breather valve must still be sized for full pump-out plus thermal in-breathing with the nitrogen supply failed. Blanketing is a consumer of the pressure setpoint budget, never a substitute for vacuum relief. If air ingress on nitrogen failure is unacceptable, that argues for redundant nitrogen supply — never for deleting the vacuum pallet.

Set Pressure Selection: API 650 vs API 620 Tanks

Set pressure selection starts with what the tank can actually take. For a standard API 650 tank, the practical ceiling is the weight of the roof plates — often under 1 kPa — unless the tank is designed to Annex F, which permits internal pressures up to 18 kPa (2.5 psig) with anchorage. API 620 tanks are a different animal, designed for up to 103 kPa (15 psig) and, via their low-temperature annexes, for refrigerated products down to −168 °C. The vacuum rating is usually far weaker — often only 0.25 to 0.6 kPa unless the tank is stiffened for more.

Within that envelope, you want the highest set pressure the lift characteristic allows, because vapour losses fall as set pressure rises toward design pressure. With a conventional pallet needing 100% overpressure, a 20 mbar design pressure forces a set of about 10 mbar; a full-lift valve rated at 10% overpressure can be set at 18 mbar on the same tank and will breathe far less product to atmosphere over a year. Always state both the set pressure and the pressure at which rated capacity must be achieved on the datasheet — the second number is the one that protects the tank.

Flame Arresters: End-of-Line and In-Line Detonation Arresters

Where the stored product can form a flammable atmosphere at the vent, a flame arrester to EN ISO 16852 stops an external ignition — lightning, a flare, hot work — propagating back into the vapour space. The standard distinguishes sharply between deflagration arresters, which stop a subsonic flame front, and detonation arresters, which must survive a shock-driven detonation with pressures an order of magnitude higher. An end-of-line arrester under a breather valve, with an open flame path shorter than the run-up distance, only ever sees deflagration. Put any significant length of pipe between the arrester and the potential ignition source and the flame can accelerate to detonation — that is when an in-line detonation arrester, tested and certified as such, becomes mandatory.

Match the arrester’s explosion group to the product: IIA covers most hydrocarbons, IIB1 to IIB3 handles narrower maximum experimental safe gaps such as ethylene mixtures, and IIC (hydrogen, acetylene, carbon disulphide) requires the tightest elements. Also specify the endurance-burn or short-time-burning class if stabilised burning on the element is credible. Two practical warnings from the field: an arrester element adds real pressure drop, so the venting calculation must be redone with the arrester included; and a fouled element is both a flow restriction and a fire risk, so it must appear on the inspection schedule, not just the P&ID.

Seat Tightness and Leakage: API 2000 Annex A

A breather valve is not bubble-tight, and pretending otherwise leads to bad emissions accounting. API 2000 Annex A provides the standard seat-leakage test method, and a good modern pallet valve with a soft or air-cushion seat will demonstrate leakage in the region of 0.014 Nm³/h (about 0.5 SCFH) of air at 75% of set pressure. Older metal-to-metal weighted pallets can leak an order of magnitude more, particularly after a few years of seat scoring and paint overspray. For emissions inventories and nitrogen consumption estimates, use the tested leak rate at your actual operating pressure, not a hopeful zero.

Specify the leakage acceptance criterion explicitly on the datasheet and ask for the test certificate. In my experience the difference between a 0.5 SCFH valve and a 5 SCFH valve is invisible at purchase and painfully visible in the annual nitrogen invoice and the site’s fugitive VOC numbers.

Cold Climate and Cryogenic Service: Icing

Pallet valves fail in cold weather in a very specific way: water vapour in the tank atmosphere condenses on the cold seat and pallet, freezes, and either glues the pallet shut or props it open. A pallet frozen shut turns the next pump-out into a vacuum event; frozen open, the tank breathes freely and your blanketing gas disappears. Mitigations are well established — heat-jacketed or steam-traced valve bodies, non-icing coated or cushioned seats, weather hoods, and pilot-operated valves whose seats sit away from the moist gas stream.

On genuinely cryogenic tanks — LNG at −162 °C, ethylene, LPG to API 620 Annex R or Q — venting design belongs to specialists, and materials must retain toughness at design temperature: think 304/316 stainless internals and no elastomers where they can freeze. My blunt rule: if the site sees frost and the tank breathes moist vapour, an unheated bare-aluminium pallet valve will ice at some point. Plan for it in the design phase, not in the incident report.

Pipe-Away vs Vent-to-Atmosphere

A vent-to-atmosphere breather valve discharges through a weather hood straight to air, and its set point sees only atmospheric back pressure. A pipe-away valve has a flanged outlet connected to a vapour recovery unit, a flare header, or simply a remote safe location — and everything about it changes. Superimposed and built-up back pressure act on the pallet and shift the effective set point, so pipe-away valves need balanced designs or at minimum a careful back-pressure calculation. The outlet piping also becomes a potential detonation run-up length, which usually drags an in-line detonation arrester into the design.

Manifolded vent systems collecting several tanks deserve particular suspicion. One tank filling while another empties can push vapour between tanks, and a slug of condensate in a low point can dead-head the whole header. Slope the lines, drain the low points, size for the coincident worst case, and check every set point against the header’s maximum credible pressure.

Maintenance and Inspection Intervals

Breather valves are simple, exposed, and neglected — a bad combination. API 2000 and API 576 both point toward inspection at intervals set by service severity, and the sensible starting point is a visual check quarterly and a full inspection with seat-leakage or set-point verification at least annually, tightening or relaxing based on findings. Polymerising products, dusty atmospheres and marine sites all justify shorter intervals; a clean diesel tank in a temperate climate can often move to two or three years once history supports it.

The inspection itself takes twenty minutes: confirm the pallet lifts freely and reseats, check seat and diaphragm condition, verify weights match the datasheet, clean the screens, and confirm nobody has painted over the works — painters love venting equipment. Check the flame arrester element at the same visit. Record set points and leak rates so drift shows up as a trend rather than a surprise.

Common Specification Mistakes

After many years of reviewing tank venting datasheets, the same errors keep appearing. First and worst: sizing the valve at set pressure instead of at allowable overpressure, or comparing a 10%-overpressure valve against a 100%-overpressure valve as if their catalogue capacities were equivalent. Second: ignoring thermal in-breathing entirely, or using the old flat allowances on a large tank where the 7th-edition method demands substantially more.

Third: setting the pressure pallet on top of the blanketing regulator setpoint, guaranteeing chatter and chronic nitrogen loss. Fourth: adding a flame arrester late in the project without re-running the venting calculation for its pressure drop. Fifth: no vacuum case for loss of blanketing gas, and no stated leakage class, so the site inherits whatever seat the low bidder had in stock. Each of these is a five-minute fix on paper and a five-figure problem in the field.

Tank venting rewards the engineer who respects small numbers. A few millibar of set pressure, a fraction of a cubic metre per hour of leakage, one overlooked thermal case — these are the margins in which storage tanks are saved or crumpled. Size to API 2000, demand tested flow and leakage data, build a sensible setpoint ladder, and inspect the hardware as if the roof depends on it. Because it does.