An older atmospheric tank holding a flammable liquid with a flash point just under 20 °C (68 °F), toxic by inhalation and by skin absorption, nitrogen blanketed, with a lute pot as its only relief path. No flare on site. The question that actually has to be answered is not "lute pot or pressure/vacuum valve" — it is "where does the vapor go, and who stands in it." Hardware selection follows from that answer, not the other way round.
Vent Duty and Vapor Hazard Inputs
Before anything else, confirm the four numbers that fix every downstream decision. Without them the comparison below is opinion.
- Tank design pressure and design vacuum. Read them from the nameplate or the original data sheet, not from the P&ID. On an old riveted or lightly-built atmospheric tank the vacuum rating is usually the binding constraint, often only a few millibar.
- Normal venting requirement. Maximum pump-in rate plus thermal out-breathing; maximum pump-out rate plus thermal in-breathing. API 2000 gives the thermal breathing basis as a function of tank capacity and liquid vapor pressure category. A liquid flashing below 20 °C sits in the higher-volatility category — do not use the low-volatility breathing rates.
- Fire case requirement. Wetted area up to the API 2000 height limit, environment factor for insulation and drainage, latent heat at relieving conditions, converted to the air-equivalent flow the vent manufacturer's curves are published against. This number is typically 10–100x the normal breathing rate and it is what drives the vent nozzle size.
- Toxic exposure limits. Short-term and long-term permissible concentrations for the compound, from the safety data sheet and the local environmental and occupational limits. These are the pass/fail criteria for the dispersion study.
Do not move on until the fire case flow is quantified. A P/V valve sized only for normal breathing on a tank with no other emergency path is a tank that will unfold or split in a pool fire.
Venting Options Compared
| Option | Set pressure control | Fire case capability | Toxic release control | Maintenance burden |
|---|---|---|---|---|
| Lute (seal) pot | Seal depth only, P = ρgh; drifts with level, evaporation, dilution | Poor — seal blows out and the tank is left with an open pipe of fixed bore | None; discharge is at the pot, plus continuous evaporative emission from the seal liquid | High — level checks, freeze protection, algae and debris cleaning |
| P/V valve at the tank | Weighted or spring pallet, repeatable set and reseat | Only if deliberately sized for it; usually not economic in one device | Discharge at tank roof level — unacceptable for a toxic vapor unless dispersion proves it | Low — periodic pallet lift test and seat inspection |
| P/V valve piped to a remote elevated stack | Same, but backpressure from the header adds to the effective set point | Requires a parallel emergency vent; header pressure drop must be checked at fire-case flow | Good — release elevated above manned levels, dispersion-verified | Moderate — header, arrester and drain-pot upkeep |
| Emergency vent (weight-loaded or pilot-operated), separate nozzle | Set above the P/V pressure set, below tank design pressure | Yes — this is its only duty | Local release during a fire event; covered by radiation and dispersion study | Low — annual seat and gasket check |
| Thermal oxidizer / vapor destruction | Requires a back-pressure regulator and its own relief bypass | No — fire case must still go direct to atmosphere | Best for filling displacement vapors; the only route for carcinogens | High — burner, fuel gas, permits, continuous availability |
Lute Pot Failure Modes
The seal pot's set pressure is nothing more than the static head of the liquid leg, P = ρgh. Everything that changes h or ρ changes the tank's relief setting, and nothing on the tank tells the operator it has changed:
- Level drift. Evaporation of the seal liquid lowers h and drops the set point; overfilling raises it and can push the tank toward its vacuum or pressure limit. There is no reliable way to hold that level on an unattended pot.
- Seal blow-out. A rapid out-breathing event ejects the seal liquid entirely. The tank is then permanently open to atmosphere, the nitrogen blanket is lost, air enters the vapor space, and a flammable atmosphere forms inside the tank. Nothing alarms.
- Freezing and fouling. Winter freeze-up, algae growth, corrosion product and debris block the leg. A blocked lute pot is a blanked vent nozzle on both pressure and vacuum.
- Continuous emission. The seal liquid itself is an emission source, and it is in contact with the toxic vapor space — it becomes contaminated and it is a disposal problem.
- Carryover. Bubbling out-breathing gas strips seal liquid and any dissolved toxic material into the discharge.
The one legitimate reason to keep the pot is as a temporary vent path while the relief valve is off the tank for overhaul. Confirm with maintenance whether that is in their procedure. If it is not, remove it. It is not redundancy in the sense that two P/V valves are redundancy; a device whose set point is unknown and whose failure mode is "wide open" does not add a protection layer.
Recommended Arrangement
For a blanketed, flammable, toxic atmospheric tank with no flare:
- Remove the lute pot and blank the nozzle, or re-use the nozzle for the new vent.
- Install a pressure/vacuum valve sized for normal breathing, discharge piped to a remote elevated vent stack sized and located by the dispersion study.
- Install a separate emergency vent sized for the fire case on its own nozzle. Where the dispersion result permits, discharge it locally — a fire case release is short, buoyant and coincident with a fire that already dominates the hazard picture. Where it does not, pipe it to the stack and re-check the emergency vent capacity at the resulting backpressure.
- Keep the nitrogen blanket regulator on its own connection and coordinate its band with the vent set points.
- Decide the flame arrester question on the basis of the vent geometry, not by default. An end-of-line deflagration arrester on the stack protects against flashback of an ignited vent; a piped header serving several tanks needs in-line detonation arresters and a run-up length assessment. Every arrester adds pressure drop and a plugging path in a service that polymerizes or freezes — size the vent with the arrester in place and use manufacturer flow curves for the combined assembly. P/V valve and arrester manufacturers such as Protego publish those combined capacity curves.
Set the pressure ladder in this order, from lowest to highest:
| Setting | Position in ladder | Reason |
|---|---|---|
| Tank design vacuum | Lowest (most negative) | Structural limit |
| P/V vacuum pallet set | Above design vacuum, with margin for header loss | Last-resort air admission |
| Blanket regulator open | Slightly positive, above the vacuum pallet | Nitrogen makes up shrinkage before air is drawn in |
| Blanket regulator closed | Above open point by the regulator's band | Prevents continuous nitrogen bleed to the vent |
| P/V pressure pallet set | Above the blanket closing pressure | Normal out-breathing; avoids nitrogen wastage |
| Emergency vent set | Above P/V pressure set | Fire case only |
| Tank design pressure | Highest | Emergency vent must be fully open below it |
If the blanket regulator's closing pressure and the P/V pressure set overlap, the tank will breathe nitrogen to the stack continuously and the vent will be wet with toxic vapor around the clock.
Dispersion and Radiation Study
This is the step that decides stack height and location, and it is the step that closed the case here: the simulations showed negligible probability of a toxic cloud reaching persons, and in fact no significant concentration reaching grade at all. Run it as follows.
- Define the release cases separately: maximum pump-in displacement (longest duration, lowest flow), thermal out-breathing, blanket regulator failure open, and the fire case (highest flow, shortest duration).
- For each case, fix flow rate, vapor composition, temperature, exit velocity and stack diameter. Exit velocity matters — a low-velocity release from a large-bore stack falls back down the side of the tank instead of rising.
- Run the dispersion model against the site's meteorological data, including the worst-case stable, low-wind condition, not just the annual average.
- Compare predicted concentrations against short-term and long-term exposure limits at grade, at the tank top platform, at any elevated manned platform, at air intakes, and at the site boundary.
- Run a flame radiation case for the ignited vent. Check radiation levels at the same manned locations and at adjacent equipment.
- If any location fails, raise the stack, increase exit velocity by reducing the discharge bore, relocate it, or move to vapor destruction for the filling displacement stream. Elevation is normally the cheapest first move.
For a carcinogen or a compound with a very low occupational limit, dispersion may not be able to pass at any practical height. Thermal oxidation or incineration of the filling displacement vapor is then the working solution — with a direct-to-atmosphere emergency vent retained regardless, because no destruction unit is a relief device.
Why the Plant Flare Is Not the Destination
Routing an atmospheric tank vent into a general flare header is unsafe in almost every plant. The tank's design pressure is on the order of millibar; the flare header carries the superimposed backpressure of every other relieving device on it. A single relief event elsewhere in the plant puts header pressure on the tank roof, and the tank fails long before any relief valve on the header lifts. The reverse case is equally bad: header vacuum or a purge upset pulls the tank in.
A collection header dedicated exclusively to storage tanks, designed so that no significant pressure can develop in it, is workable. Even then each tank keeps an emergency vent discharging directly to atmosphere — the header is for normal breathing only. With no flare on site, the practical alternative is the elevated vent stack, and a common stack can serve several tanks provided the header hydraulics, cross-contamination chemistry and flame propagation between tanks are all assessed.
Changeover and Commissioning
- Isolate and gas-free the tank per site procedure. Confirm the vapor space is inert or free of flammables before hot work near the roof nozzles.
- Remove the lute pot and its piping. Drain and dispose of the seal liquid as contaminated waste.
- Fit the P/V valve and, on a separate nozzle, the emergency vent. Verify the nameplate set pressures against the ladder table before bolting up — factory settings are frequently not what was ordered.
- Install the vent header to the stack with a continuous fall back to a drain pot, so condensate and rain cannot form a liquid slug in the line. A slug in a vent header is a lute pot you did not design.
- Fit the flame arrester with its element accessible for cleaning without breaking the stack down.
- Set the blanket regulator open and close points, then verify against the tank pressure instrument on a slow pump-out.
Verification
- Lift-test the P/V valve pallets in place at the stated set pressures, both pressure and vacuum sides, and confirm reseat.
- Record tank pressure over a full pump-in and a full pump-out cycle at maximum rate. Confirm the pressure trace stays inside the blanket band and does not touch the emergency vent set point or either design limit.
- Confirm the nitrogen regulator does not pass gas while the tank is static — continuous nitrogen flow indicates a leaking blanket valve or overlapping set points, and it drives continuous toxic emission at the stack.
- Survey the stack discharge and the nearest manned platform with a portable detector during a maximum-rate filling operation, and confirm the readings against the concentrations the dispersion model predicted for that case.
Frequently Asked Questions
Why does a lute pot lose its seal on a nitrogen-blanketed tank?
Its set pressure is only the static head of the liquid leg, P = ρgh, so evaporation, dilution and a single fast out-breathing event all change or eject it. Once the leg is blown out the tank is permanently open to atmosphere, the blanket is lost and air enters the vapor space with no alarm.
Why does a tank need an emergency vent when a pressure/vacuum valve is already fitted?
The fire case flow calculated from wetted area and environment factor per API 2000 is typically one to two orders of magnitude above normal breathing. Sizing a single P/V valve for both duties gives an oversized, chattering device on normal service, so the standard arrangement is a normal-breathing P/V valve plus a separate emergency vent set above it and fully open below tank design pressure.
Why does routing an atmospheric tank vent to the plant flare create a hazard?
The flare header carries superimposed backpressure from every other device connected to it, and an atmospheric tank is rated in millibar. A relief event elsewhere in the plant can pressure or evacuate the tank through the header before any protection responds. Only a header used exclusively by storage tanks, with no credible pressure build-up, is acceptable — and each tank still keeps a direct-to-atmosphere emergency vent.
Why does a dispersion study decide vent stack height instead of a standard height rule?
The safe elevation depends on release flow, exit velocity, vapor density and site meteorology, and it must be checked against short-term and long-term exposure limits at grade, on the tank top, on every manned elevated platform and at air intakes. A low-velocity release from an oversized bore in stable low-wind conditions can sink back to grade from a stack that a rule of thumb would call adequate.
Why does adding a flame arrester change the vent sizing?
The arrester element adds pressure drop that counts against the margin between set pressure and tank design pressure, so the vent must be sized from the manufacturer's combined P/V-valve-plus-arrester flow curve, not the bare valve curve. It also introduces a plugging path, which is why the element has to be reachable for cleaning without dismantling the stack.