Main break response for a small crew: isolate, fix, disinfect, and talk to customers

You find out about a main break the way most small crews do: a phone call at 2 a.m., or a soggy lawn a neighbor reports on the way to work. Either way, the clock starts before you've had coffee. The next few hours test everything you know about your system at once.

A break puts your isolation plan, dig crew, disinfection rule, and public message all in play within hours. A crew that has drilled it once knows the answers in minutes. A crew that hasn't loses hours, and loses some trust along with the water.

This guide runs a main break start to finish. Isolate it, fix it, disinfect and sample it, decide what to tell people, and close it out so the next crew learns too.

What do you close first?

You close the smallest set of valves that isolates the break, while still protecting service and fire flow nearby. A simple residential break is usually contained with 2 to 3 valves. Write down which valves, and in what order, before anyone turns a wrench.

That decision runs off one mental model: source, then storage, then pressure control, then distribution. Water moves from wells or intakes, into tanks, through boosters and PRVs, then out to the mains that reach customers. Close the wrong valve and you can starve a zone, or a facility, you never meant to touch.

City of Everytown's water department keeps that discipline on paper. Each zone gets a one-page system flow overview and a facility dependency table, built ahead of time instead of improvised at 2 a.m. The table shows what each facility feeds, its normal status, and what happens if you shut it down.

What else has to shut down besides the pipe?

Closing the isolation valves is rarely the whole job. You also have to protect pressure zones, check facility impacts, and cover fire protection. Water-quality risk and site safety round out the list before you call the isolation complete.

• Pressure zones: which other zones lose pressure once these valves close.
• Facilities: whether a well, booster, or tank needs a change in status too.
• Fire protection: which hydrants go out of service, and whether the fire district needs a call.
• Water quality: negative pressure and dead ends, and where you'll flush and sample once it's fixed.
• Site safety: who directs traffic, and who checks trench depth before anyone digs.

One rule trips up more crews than any other: a storage tank below its calling level signals its well to turn on, even if that well sits nowhere near the break. If your isolation leaves that well in AUTO, it can pump straight into a segment you just depressurized. Reason from what the tank will call for, not just the pipe in front of you.

What if a valve won't close?

You expand the isolation to the next set of valves out, rather than fighting the one that won't seat. The boundary gets bigger, but the water stays contained. You flag the bad valve for a follow-up repair instead of losing an hour arguing with it.

An expanded boundary can pull in more than you planned for, a school or a daycare inside the new outage, for instance. That changes your communication priority, even though the isolation logic hasn't changed. Make the bigger call quickly, note the valve that failed, and move on.

What changes when the break is near a well or a tank?

Isolating the pipe isn't always the whole call: you may also need to change what's happening at the facility itself. It might mean taking a booster out of AUTO, or shutting down a well or pump station. The goal is keeping that facility from firing into the area you just isolated.

The trigger is the same tank-calling logic from above. Shut off only the source tied to the affected tank or segment, not every well on the system. A well that only feeds a tank outside the break area doesn't need to be touched.

Restoration adds one more decision here. You decide when the facility goes back to normal, not just when the valves reopen.

How do you repair, disinfect, and get the water back on?

The physical repair is usually the easy part: expose the pipe, cut out the damaged section, and clamp or replace it. Trench safety and disinfection are what actually protect the next customer, and neither step is optional.

Any trench 5 feet deep or more needs a cave-in safeguard before anyone climbs in, a shored wall, sloped sides, or a trench box. Anyone working 6 feet up needs fall protection. If the repair sets a new thrust block, give it at least 5 days to cure before pressure-testing the line.

Repairing a main means disinfecting it. Standard practice doses the repaired section with at least 25 mg/L of chlorine, held for 24 hours, more than 99.99 percent inactivation of what's inside. Then it's flushed, sampled, and only opened to customers on a negative result.

If pressure in the line dropped below 20 psi and the pipe was dewatered, disinfect that way when you can. If you can't, flush at least two full pipe volumes, keep going until what comes out runs clear and the chlorine number settles where it should, then sample before reopening.

One rural system returned a repaired main to service on flushing alone, after two breaks during a hard freeze. An E. coli outbreak followed, with cases in the hundreds and several deaths. Flushing clears water; it does not disinfect a pipe.

Once the sample comes back clean, reopen the valves slowly, in the order opposite of how you closed them. Watch pressure, turbidity, and chlorine residual as you go. Flush the hydrants downstream of the repair, usually 2 to 4 of them in sequence, until the water runs clear.

When does a main break turn into a public notice?

Not every break needs one. Most get isolated, repaired, disinfected, and returned to service without a public notice ever going out.

A notice becomes likely once a sample comes back positive, or pressure drops far enough, long enough, to look like contamination instead of a plain outage. Federal rules then sort the response into tiers. The fastest covers acute health risks, like a confirmed positive sample, with a short clock to notify; slower tiers apply to violations with no direct health signal.

Which tier applies, and how to write and deliver the notice, is worth its own guide: see /guides/boil-water-notice-and-crisis-comms.

What you can decide now, before any of that: whether your crew calls the state the moment a sample comes back positive, or waits on a second sample first. Waiting costs hours you don't get back.

What makes the next break go faster?

Two habits make the difference: a short documentation record after every break, and an honest debrief while it's still fresh.

The minimum record worth keeping: the break's location, time, and classification, the valve IDs you closed, at least one photo, and any valve that didn't seat. Add a follow-up task to fix what didn't work. Skip that, and the next crew starts from zero.

After the water's back on, ask three questions while the break is still fresh: what data was missing, what you'd have wanted at 2 a.m., and what you'd change on the map tomorrow. Asked consistently, those three questions turn a bad night into a better system next time.

If your crew hasn't run a break as a drill, do one before the next storm or hard freeze. Walk a simple residential break, then one with a valve that won't close, then one near a well or a tank, and time yourselves. The valve order you argue about on paper is the order you won't have to guess at 2 a.m.

Restoration numbers and hydrant flushing sequences live in /guides/water-distribution-mains-flushing. Keeping valves in a state where they actually close when you need them lives in /guides/valve-and-hydrant-programs.