What a Deluge System Does That the Others Cannot

Why this matters

Every closed-head sprinkler system in this library shares one property: the thermal element makes the system self-limiting. Only heads that get hot open, so the water goes where the fire is and the demand is capped by how many heads can plausibly be involved at once. A deluge system throws all of that away on purpose. The heads are open, the valve is opened by something other than heat, and every nozzle in the system flows at the same instant across the whole protected area. That is an enormous capability and it is bought by giving up four specific things. If you only know what a deluge system adds, you will size a water supply that cannot feed it and you will test it in a way that ruins the space underneath.

What it actually is

Open heads or spray nozzles, no thermal element, on piping that is empty and open to atmosphere. A deluge valve holds water back at the riser. The valve is released by a detection system, by a manual release station, or by a process interlock, and once it opens, water flows from every opening in the system until somebody closes the valve by hand.

Three arrangements cover most of what a field-service shop meets:

• Water deluge, plain water from open sprinklers over an area.
• Water spray, directional nozzles aimed at a specific object rather than at a floor area: a transformer, a vessel, a pump skid, a cable tray run. The governing installation standard here is NFPA 15, in the edition the authority having jurisdiction has adopted and amended.
• Foam-water, where foam concentrate is proportioned into the water stream to blanket a flammable liquid surface. Installation is governed by NFPA 16, again in the adopted edition, and the foam concentrate side by NFPA 11, also only in the edition your authority having jurisdiction has adopted and amended.

The four things it gives up

It gives up locality. A closed-head system applies water only under the heads that got hot. A deluge system cannot do that. If the fire is in one corner of the protected area, the whole area gets wet, including the equipment, the stock and anyone standing in it. That is not a side effect to be minimised; it is the design.

It gives up the cap on water demand. This is the one that ends projects. A closed-head system's demand is density multiplied by an area of operation, and the area of operation does not grow with the building. Two warehouses on the same ordinary hazard classification, one 40,000 sq ft and one 80,000 sq ft, have the same sprinkler demand, because the design assumes a fire involving a remote area of a fixed size rather than the whole floor. Doubling that building adds nothing to the sprinkler demand. A deluge system has no area of operation at all. Its demand is density multiplied by the entire protected area, and it scales linearly and without limit. Density times area is the relationship, and the sibling coverage card owns it; what is specific here is that the second term is the whole system rather than a remote portion of it.

It gives up the head as a backup decision-maker. On a closed-head preaction system, the head is still a thermal device, so heat is still part of the chain. On a deluge system there is no thermal element anywhere. If the releasing detection does not operate, nothing happens, ever, regardless of how hot the space gets. Detection is not a convenience on a deluge system; it is the entire actuation path, and everything the sibling detection card says about a disabled zone applies here with no fallback behind it.

It gives up quiet testing. A closed-head system can be flow tested at an inspector's test connection that discharges to a drain. A deluge system's full flow trip test wets the protected area, because that is the only way to prove every nozzle is clear. NFPA 25, in the adopted edition, sets the interval and requirements, and it is a planned event with the area cleared, equipment de-energized or covered, drainage confirmed, and the discharge disposition settled in advance.

There is a fifth item that is less a trade-off than a maintenance consequence: open piping is open to the room. Nozzles collect dust, paint overspray, insects and nests, because a closed head seals its own orifice and an open nozzle does not. Blow-down and nozzle inspection are real recurring work here in a way they are not on a wet system.

What it buys that nothing else can

Simultaneous full-area application. Where a fire can involve the entire protected area faster than individual heads could open, applying water head by head is applying it behind the fire. A flammable liquid spill, a gas release ignition, a spray booth, a fuel loading rack: the fire reaches the boundaries of the spill in seconds. A deluge system is the only water-based arrangement that can meet that.

Actuation from something other than heat. Because the release is a detection decision, the system can be tripped on flame detection, on a combustible gas reading, on a process alarm, or by a person pulling a manual station. That means it can operate before a thermal signature exists at the ceiling, which is exactly what a closed head cannot do.

Exposure protection. A water spray system on a vessel or a transformer is not fighting a fire inside the object; it is keeping the object's surface cool so that a fire nearby does not fail it. Nothing that waits for local heat can do that job, because by the time the ceiling above the vessel is hot the vessel is already the problem.

The scaling arithmetic, worked

A chemical drum-filling station under an open canopy, 20 ft by 30 ft, so 600 sq ft of protected floor. The design density for a water spray system covering a flammable liquid spill area comes from NFPA 15 in the adopted edition and from the design documents; take it here as 0.25 gpm per sq ft, which is a figure the design owns rather than one this card sets.

Deluge demand equals density times the entire protected area:

0.25 gpm per sq ft times 600 sq ft equals 150 gpm, all flowing at once, from the moment the valve opens.

The owner then wants the canopy extended to cover a second filling position, taking the protected area to 1,400 sq ft. The demand goes to:

0.25 gpm per sq ft times 1,400 sq ft equals 350 gpm.

That is a factor of 2.33 on the water demand for a factor of 2.33 on the area, exactly linear, with no cap anywhere in the relationship. Run the same expansion on a closed-head system and the number does not move at all, because the area of operation is a remote portion of the building rather than the whole of it. This is the single most important number to have in your head before anyone talks about extending a deluge-protected area.

Corrections, printed.

• Raw demand basis: density times area. Correction applied: the area term on a deluge system is the entire protected area, not a remote area of operation, so the closed-head cap does not apply and is not carried over. Corrected basis: 600 sq ft and then 1,400 sq ft, both in full.
• Raw hose stream allowance: not included above. Correction applied: it is added at the point of connection on top of the system demand, it is the same addition in both the 600 and 1,400 sq ft cases, and it is excluded from this comparison by choice and named as excluded. The coverage card owns the allowance itself.
• Raw supply assumption: the existing supply fed the original canopy, so it will feed the extension. Correction applied: the existing supply was proven against a 150 gpm demand at the required nozzle pressure, and it has to be re-proven against 350 gpm, because a supply curve falls as flow rises. This is a re-basing of the supply question, not an addition to it. The new figure requires a current flow test at the point of connection, not the flow test that supported the original design.
• Raw foam assumption: none, because this example is plain water spray. If foam is proportioned in, concentrate quantity and proportioning rate become a second demand with its own duration, and NFPA 16 in the adopted edition owns it.

What would flip the answer. If the extension pushed the demand past what the supply can deliver at the required nozzle pressure, the options are a stored water supply with a fire pump, which is a machine the pump articles in this library cover, or splitting the canopy into two separately released deluge zones so only the involved zone flows. Zoning is the cheaper answer and it costs something real: a fire that starts at the boundary between zones is fought by one zone, so the zone boundary becomes a design decision the designer and the authority having jurisdiction have to sign off, not a convenience the installer picks.

Sibling-rule check on this example. Demand was computed as density times area rather than from a nozzle count, consistent with the coverage card. No thermal element was credited anywhere in the actuation path, consistent with the head card, which describes what a closed head waits for and which explicitly does not apply to open nozzles. Detection was treated as the entire actuation path with no fallback, consistent with the detection card and stronger than the preaction card's case, where a closed head still sits behind the interlock. The design density was routed to the adopted standard and the design documents rather than set here, consistent with the classification card. No control valve was closed and no releasing circuit was left disabled at the end.

Failure mode. The way this goes wrong is that somebody extends the protected area physically without anyone recomputing the demand, usually by adding a canopy bay, moving a rack, or adding nozzles to cover a new piece of equipment. Nothing looks wrong. The system tests fine, because a trip test proves the valve opens and the nozzles are clear, not that the supply can hold the required pressure at every nozzle while all of them flow. The deficiency only appears during a real discharge, as pressure at the far nozzles below what the spray pattern needs, and by then the pattern that was supposed to blanket the spill is a set of streams that do not.

Working on one without causing the event

• Never operate a manual release station to see whether it works. A manual station releases the valve, and the valve does not care why. Functional testing of the release path happens with the deluge valve isolated and the area cleared, under the procedure in the adopted edition of NFPA 25.
• Before any full flow trip test, clear the area of people, open and verify dead the supply to any energized equipment that will be wetted, confirm the drainage path handles the flow, and confirm where the discharge goes. Water landing on energized equipment is a shock path and a short, and standing water in an occupied area is a fall hazard for everyone who walks through afterwards.
• On a foam-water system, treat the concentrate as a chemical. Read the safety data sheet for the specific concentrate before handling it, use the eye and skin protection it specifies, and confirm the discharge containment before a test, because foam released to a storm drain is an environmental release with notification obligations that are not the fire code's business and will still be yours.
• Isolating the deluge valve is an impairment, with a named coordinator, a compensating measure appropriate to a total suppression outage in that area (which for a flammable liquid operation normally means stopping the operation, not adding a fire watch to it), and a documented close-out under NFPA 25 in the adopted edition.
• Do not stand under open nozzles while resetting the valve. Resetting involves restoring water to the valve, and a valve that does not seat sends the full system flow onto whoever is under it.

References

• NFPA 13, Standard for the Installation of Sprinkler Systems, in the edition adopted and amended by the authority having jurisdiction, for deluge systems within its scope
• NFPA 15, Standard for Water Spray Fixed Systems for Fire Protection, and NFPA 16, Standard for the Installation of Foam-Water Sprinkler and Foam-Water Spray Systems, both in the adopted editions
• NFPA 25, in the adopted edition, for deluge valve trip testing, nozzle inspection and the impairment program
• The safety data sheet for the specific foam concentrate, which owns handling, protection and disposal
• See related: Why Sprinkler Coverage Is a Density Problem, Not a Head Count; What a Preaction System Is Protecting Against