Why a Cold Start Is the Most Dangerous Moment

Why this matters

Steam systems fail on Monday mornings, after holiday shutdowns, and on the first cold night of the season. That is not superstition and it is not because the equipment rested. It is because a cold start is the only moment when the rate at which the system makes condensate is at its highest and the capacity of everything meant to remove it is at its lowest, and both of those extremes have the same cause. Cold metal makes condensate fast, and cold metal means there is no pressure yet to push that condensate anywhere. Every other hour of the year, one of those two is comfortable.

Three curves and where they cross

Three quantities move during a start-up and they do not move together.

Condensing rate starts at its maximum. The metal is at room temperature, the temperature difference to saturation is the largest it will ever be, and every pound of steam that touches cold steel condenses completely.

Drainage capacity starts at its minimum. A trap is a differential-pressure device, and at the instant steam is admitted the main is at atmospheric pressure, so there is almost no differential across any drip trap on the run.

Venting duty also starts at its maximum, because the main is full of air that has to leave before steam can occupy the space, and every vent on the system is being asked for its full rated flow in the first minutes.

By the time the system is hot the three have crossed to their comfortable values, which is why a system that has been running since dawn tolerates almost anything. The design margin the shop relies on for the other 99 percent of the year does not exist during the crossing.

The warm-up load is a property of the metal

The condensate produced during warm-up is not an estimate. It is the heat needed to raise the pipe metal to saturation temperature, divided by the latent heat available in a pound of steam at that pressure:

warm-up condensate, in pounds = (pipe weight in lb) x (specific heat, Btu per lb per F) x (temperature rise, F) / (latent heat at operating pressure, Btu per lb)

Carbon steel takes about 0.12 Btu per pound per degree F across this range. Hold the conditions on that figure: it is carbon steel warming from room temperature toward a few hundred degrees, it rises modestly at higher temperature, and it applies to the pipe metal only. Schedule 40 pipe weights come off a standard pipe data table, where 4 in is about 10.79 lb per foot, 6 in about 18.97, and 8 in about 28.55. Flanges, valves, strainers and fittings are not in those figures and can add on the order of ten to twenty percent to the metal on a real run, so a calculation from pipe weight alone is a floor, not a bound. Where the margin is tight, add the fitting weights from the same data table rather than assuming the floor holds.

Latent heat comes from a steam table at the operating pressure and it falls as pressure rises: around 970 Btu per pound at 0 psig, around 880 at 100 psig. That direction is worth holding on to, because it means a higher-pressure system needs more pounds of steam to deliver the same warm-up heat, not fewer.

The running load is a different calculation entirely: heat loss per foot from the insulation manufacturer's table for the actual thickness and jacket, multiplied by the length, divided by the same latent heat. The two are not related, they are not proportional to each other, and the ratio between them commonly runs from one to two orders of magnitude depending on how quickly the warm-up is done.

The rate is a decision

This is the part that changes how a shop operates rather than how it calculates.

The total warm-up condensate is fixed by the mass of metal. Nobody can reduce it without reducing the pipe. What is entirely under the operator's control is how many minutes that fixed quantity is spread over, and the rate the drainage has to handle is the total divided by the time.

Halve the warm-up time and you double the pound-per-hour rate the drip traps must pass, at a moment when the differential across them is at its lowest. That is why an unsupervised timeclock start with an aggressive schedule produces hammer on a system that a hand start over forty minutes never troubles, with no component different between the two. The procedure for doing this properly is the sibling article on warming a steam line back up without breaking it; this article's job is to show why the rate is the variable that matters.

Why every drainage device is weakest exactly now

Three things go soft at once.

Differential across the drip trap. At the moment of admission the main is near atmospheric, so the only differential available is the static head of water standing in the drip leg. Water converts at about 27.7 in per psi, so a 28 in leg supplies roughly 1 psi, and the sibling article on drip legs covers why that depth exists. Read the trap's capacity chart at 1 psi rather than at operating differential, and the number you find is a small fraction of the trap's headline rating.

Trap type. Warm-up condensate arrives cool, because the metal it came off is cool, so it does not flash when it discharges. Some trap mechanisms depend on flash across the seat to operate and are least reliable on cool water at low differential; a float-operated trap with a thermostatic air element handles the cool, low-differential start and then the hot, flashing running condition equally, which is why it is the common choice on drip service. At the other end of the range, once the main is hot the same trap is discharging condensate at saturation temperature which flashes on the way out, and that is the condition the headline capacity was published for.

Where the water is. Any low point, sag or pocket that has been collecting condensate since the last shutdown is full at the moment steam is admitted, and it is directly in the path of the steam front. That is the precondition for slug propulsion described in the sibling article on what water hammer actually is in a steam line.

The air the main is already full of

A main that has been shut down and cooled is full of air at atmospheric pressure, and that air does not simply vanish when steam arrives.

Treat the air as swept ahead of the steam front rather than mixed with it. That is close to what happens at a dead end and an approximation along the run, but it makes the consequence visible: if the air cannot leave through a vent, it is compressed into the far end of the main by the rising pressure. Air at 14.7 psia compressed to 114.7 psia occupies about an eighth of its original volume, and that eighth sits against the closed end, which is where the equipment is.

The heat consequence and the pressure-temperature consequence of that pocket belong to the sibling articles on what an air vent is doing in a steam system and on why air in steam is worse than air in water. What belongs here is the timing: the vent's entire duty happens in the first few minutes, so a vent that has failed closed produces no symptom at all during normal running and a cold zone every morning.

The artifact: a start-up worksheet for one main

Fill this in once per main and keep it with the system documentation. The entries below are one worked case: 200 ft of 6 in Schedule 40, operating at 100 psig, in a building that sits at 60 F over a weekend.

Two entries carry the article. The 21 to 1 ratio says the drip trap that comfortably passes the running load has to pass twenty-one times that, at roughly one hundredth of the differential it was rated at, at the same moment. And the last three lines say that a main whose venting has failed ends the start-up with its final 25 ft still cold, which is where the load is connected. Both of those are properties of the same twenty minutes.

The insulation figure of 90 Btu per hour per foot is illustrative and labelled as such; take yours from the manufacturer's table for the actual thickness and jacket, because that single input moves the running load and therefore the ratio directly. If the real figure were half that, the ratio would be about 42 to 1 rather than 21 to 1, which sharpens the finding rather than softening it.

What changes the picture

A system that is never fully shut down does not have this problem in the same form. A main held at low pressure overnight has warm metal, a live differential across its traps and no air in it, so the morning transition is a pressure change rather than a start. Where the load allows it, holding a main at low pressure through a shutdown is the single cheapest change available, and it converts the daily crossing into a weekly or seasonal one.

Small-mass systems behave differently by a wide margin. Steam tracing, small-bore distribution and short runs have so little metal that the warm-up load never dominates, and the arithmetic above will tell you that directly: run the pipe weight and the ratio, and if it comes back in single digits the start-up is not your risk and you should be looking elsewhere.

Supervised versus automatic start changes what the drip legs have to do rather than what the load is. A person on the floor can crack drains, watch the discharge and hold pressure back; a timeclock cannot, which is why unattended systems carry deeper legs and a slower ramp.

A long shutdown in cold weather raises the temperature rise term directly. The same main starting from 40 F instead of 60 F carries a 298 F rise rather than 278 F, about seven percent more condensate, which is not dramatic on its own but arrives alongside colder ambient air and a longer period of accumulated water in every low point.

References

• ASME Steam Tables, or an equivalent saturated-steam table, for saturation temperature and latent heat at your operating pressure
• Standard pipe data tables for Schedule 40 weights per foot and fitting weights
• Insulation manufacturer heat-loss tables for the running load at the installed thickness and jacket
• Trap manufacturer capacity charts read at the differential available at start-up, not at operating differential
• See related: articles in this library on warming a steam line back up without breaking it, on why a drip leg exists and what happens without one, and on what an air vent is doing in a steam system