What Latent Heat Buys You in a Distribution System
Why this matters
Everybody repeats that steam carries a lot of heat per pound. That is true and it is not the reason steam distribution systems exist. The reason is that a condensing fluid gives up its heat without changing temperature, so the last terminal unit on a four-hundred-foot main sees essentially the same surface temperature as the first one, with no balancing valves and no oversizing of the far end. A pumped water loop physically cannot do that. When a steam system's far end goes cold, the cause is almost never a temperature gradient, because there barely is one - it is that what arrives at the far end has stopped being steam. Knowing which of those two you are looking at is the difference between re-piping a main and cleaning out a drip leg.
Stop here before opening or warming a main
Warm-up is when steam mains hurt people, not steady operation. A cold main full of air and a cold main that has just taken steam are both loaded with condensate, and a slug of that condensate accelerated by incoming steam will move a pipe and break a fitting.
Warm a main by cracking the isolation valve barely off its seat and leaving it there long enough for the pipe metal to come up, with every drip leg trap and warm-up drain open first and the discharge routed to a drain rather than the floor at your feet. Stand beside the valve stem, never in line with a drain outlet, cap or blank, and wear face protection under 29 CFR 1910.133 with sleeves down. Do not open a main quickly to "get it over with"; that is the specific act that produces condensation-induced water hammer, in which a pocket of steam trapped in cooler condensate collapses to a fraction of its volume and the surrounding water fills the void at high speed.
Before any of that, isolate and lock the supply and the return under 29 CFR 1910.147 if you will be opening anything, and confirm zero gauge pressure on a gauge you can read. If you have to disturb insulation on a main in a building that predates 1981, treat it as presumed asbestos-containing thermal system insulation under 29 CFR 1926.1101 for construction work or 29 CFR 1910.1001 in general industry, stop, and bring in a trained abatement contractor; where any limited disturbance is permitted at all it requires respiratory protection under a written program meeting 29 CFR 1910.134, because the route is inhalation and no glove or sleeve addresses it.
The property being bought
Condensation happens at constant temperature. A pound of saturated steam at 15 psig sits at roughly 250 F, releases roughly 946 Btu as it condenses, and the resulting condensate is still at roughly 250 F. Nothing about that transaction cools the steam that has not yet condensed.
Compare the two ways heat leaves a distribution main:
- In a water loop, delivered heat comes out of temperature. Every Btu you take out drops the water. That is what a design delta-T is.
- In a steam main, delivered heat comes out of mass. Every Btu taken out condenses about one nine-hundred-fiftieth of a pound of steam. The remaining steam is at the same temperature it started at, because it is still saturated at the same pressure.
Heat loss from the main itself follows the same rule. An uninsulated or badly insulated steam main does not deliver cooler steam to the far end; it delivers less steam and more water. The loss shows up in the drip legs, not on a thermometer, which is exactly why a neglected steam main can be losing a great deal and still pass a surface-temperature check.
Two numbers used below belong to the general case and are stated here so the example does not introduce them: carbon steel has a specific heat of about 0.12 Btu per pound per degree F, and 4-inch schedule 40 steel pipe weighs about 10.8 pounds per foot.
The case: two unit heaters that never keep up
A shop building, 400 feet of 4-inch main at 15 psig, six identical unit heaters tapped off it. Complaint: in cold weather the last two on the run never hold the space, and it has been that way since a re-route around a new mezzanine two winters ago. The previous contractor quoted a larger main.
First hypothesis: pressure drop along the main. Easy to test and easy to misread. A gauge at the boiler read 15 psig, a gauge at the last takeoff read 14 psig. Now convert that into the thing that matters, which is surface temperature, not pressure: saturation at 15 psig is about 250 F and at 14 psig about 248 F, so one psi of drop in this part of the curve is worth about 2 F. Against a 65 F space that changes the driving temperature difference from 185 F to 183 F, a difference of about 1 percent. A pressure drop that size cannot produce a unit heater that will not hold a space, and a bigger main would have bought that 1 percent. Hypothesis dead, and the quote with it.
Second hypothesis: the last two units are fouled or restricted. Their coils were as clean as the four that work, their fan amps matched, and swapping the control valve on one of them changed nothing. Dead.
Third hypothesis: they are undersized for their end of the building. They are identical to the four that work, and the loads they serve are no larger. If the physics of the main were doing its job, identical units on the same main should perform identically. Dead, and this is the hypothesis worth pausing on, because on a hot water system it would have been alive.
Run the water comparison and see why. Six terminal units in series on a one-pipe hot water loop, supply 180 F, 40 F total drop across the loop, 65 F space: the first unit sees water at 180 F for a driving difference of 115 F, and the last sees 140 F for a driving difference of 75 F. That is 65 percent of the first unit's driving difference, so identical hardware at the end of a series water loop is roughly a third down on output before anything is wrong with it, and the standard answer is to oversize the downstream units. A reverse-return parallel loop narrows that gap considerably, at the cost of balancing valves that stay balanced only as long as nobody changes a flow. Steam needs neither, and that is the purchase.
What was actually true. The re-route had put 120 feet of the main on a back-pitch, running slightly uphill away from the boiler, and there was no drip leg at the new low point. Condensate collected there and sat. The last two takeoffs came off a main that was partly full of water, so what they received was a wet mixture whose latent heat per pound of flow was well below the 946 the table promises, and intermittently no vapour at all.
How it was confirmed. A contact probe on a cleaned spot on top of the main at the last takeoff, taken with the lead and hand clear of the pipe, read below the saturation temperature the local gauge called for, by clearly more than the two instruments' combined uncertainty. Saturated steam cannot do that; only a partly flooded pipe or a non-condensable blanket can. Opening the drain at the new low point, standing beside it and not in front of it, produced continuous water rather than the brief discharge a healthy drip point gives.
What the fix actually was
A drip leg and trap at the new low point, and a re-hang of the back-pitched section to fall toward the boiler at a consistent grade. No larger main, no larger unit heaters.
The generalisable part: on a steam main, the far end goes cold for phase reasons, not gradient reasons. Look for water and air before you look for pressure. Non-condensable gas does the same thing by a different route - it does not condense, so it is swept to the coldest, furthest point of the run and blankets the heat transfer surface there. Both faults concentrate at the far end, which is why the far end is where a steam system tells you it is unwell.
The load that is invisible until it hurts you
Warming that main from cold is its own sizing case, and it is a much bigger instantaneous load than running.
400 feet at 10.8 pounds per foot is about 4,320 pounds of steel. At 0.12 Btu per pound per degree F, that is about 518 Btu per degree F. Bringing it from a 60 F building to 250 F is 190 degrees, so about 98,400 Btu goes into the metal. At 946 Btu per pound, that is about 104 pounds of condensate produced purely by heating the pipe.
That figure is a floor, not a total: it ignores the insulation's own mass and everything the bare and insulated surfaces lose to the room while they are warming. Compare it to running: six unit heaters at design produce condensate at a steady rate the traps handle all day. The warm-up hundred pounds arrives in the first minutes, at a rate several times the running rate, through drip legs sized for the running rate. That is why warm-up valves exist, why start-up is when hammer happens, and why the correct answer to a hammering main is usually a slower warm-up and a drip leg, not a bigger trap.
What would change the answer
A long run with a large elevation gain does introduce a real pressure gradient, and therefore a real temperature gradient, because static head is not free in a two-phase riser. On a run with meaningful lift, take the pressure at the far end rather than assuming the boiler gauge applies.
A high-pressure header changes the gradient's value, not its existence. The same 1 psi of drop at 100 psig is worth roughly 0.7 F rather than 2 F, so the argument gets stronger, not weaker, as pressure rises.
Poor steam quality moves the loss into the delivered pound. Steam carrying moisture off the boiler delivers proportionally less latent heat per pound of flow, and unlike the faults above that shortfall is present at the near end too. If the first unit heater is also slightly down, look at the boiler and the separator, not the main.
References
- ASME Steam Tables (IAPWS industrial formulation) for the saturation pressure and temperature pairs and latent heat values used above
- Published steel pipe dimension and weight data (schedule 40 nominal weights) for the warm-up mass calculation
- 29 CFR 1910.147 for isolation of pressure and thermal stored energy; 29 CFR 1910.133 for eye and face protection during warm-up and draining
- 29 CFR 1926.1101 (construction) and 29 CFR 1910.1001 (general industry) for presumed asbestos-containing thermal system insulation, with respiratory protection under 29 CFR 1910.134
- See related: What Latent Heat Explains That Sensible Heat Cannot; What a Steam System Actually Consists Of; How to Read a Steam Table Without Memorising It