How to Allow for Movement in an Installation

Why this matters

Every run you install is going to move. The only question you control is where. A run that has nowhere to go puts its full thermal force into the weakest thing attached to it, and the weakest thing is almost never the pipe - it is the equipment connection, the valve body, the wall penetration, or the joint you sweated last. Shops lose whole afternoons chasing a noise or a weep that is really an anchoring decision made in the first hour of the install and never written down. This is a design job you do standing on a ladder in about ten minutes, and it either happens then or it does not happen.

The steps below are ordered by what a shop loses when the step is skipped, heaviest loss first, rather than by the order your hands do the work. If you only have time to get one thing right, get the first one right.

Before you cut into an existing run

Thermal force is stored energy, and a run held between two anchors is holding it whether it looks tense or not. Cutting an anchor, a hanger, or the pipe itself releases it, and the pipe moves before you have decided it should.

• Isolate the line, then relieve pressure to zero at a gauge you can read. A closed valve is not a relieved line.
• Let the line cool to a temperature you can hold your hand on before you release any anchor on a hot run. A hot run under compression springs when you free it.
• Where a spring hanger, a counterweight, or a loaded support is in the path, restrain or de-tension it first and confirm it is de-tensioned before you cut. These devices store energy independently of the line's temperature.
• OSHA's energy control standard at 29 CFR 1910.147 requires that stored energy be relieved, disconnected, or restrained before servicing begins, and thermal and spring energy in a piping system are exactly that.

Step 1: Decide where the run is anchored, and anchor it there on purpose

What you lose by skipping it: the equipment. An unanchored run does not distribute its force evenly. It puts the load into whatever resists it most, which is usually a pump flange, a heat exchanger connection, a valve body, or a tank nozzle - the most expensive and least replaceable item on the job.

An anchor is a support that transmits no movement, only force. Everything else on the run is a guide or a hanger. Pick your anchor locations first, mark them, and make sure the structure behind them can actually take the load. Anchoring to a stud wall is not anchoring.

Anchor placement is also how you cut the movement in half without buying anything: anchoring at the midpoint of a run splits one long movement into two shorter ones, each heading in a different direction, each needing its own accommodation.

Step 2: Compute the movement before you choose hardware

What you lose by skipping it: an accommodation that is undersized, buried, and silent. An expansion device that runs out of travel does not announce itself. It reaches its limit and turns into an anchor, and now you have the step-one failure with an extra part in the line.

The calculation is one line: coefficient of expansion, times the constrained length, times the temperature swing. Use the swing between the temperature at installation and the extreme the run reaches in service, in both directions. Use the published coefficient for the actual material, and remember that plastic piping classes run several times higher than copper and steel.

Step 3: Choose the accommodation type to match the movement's direction

What you lose by skipping it: a device that binds and transmits load anyway, which is a rework of the run.

• Loops and offsets absorb movement by flexing pipe. They need room and they need the legs to be free to bend, but they have no wear parts and no travel limit to run out of.
• Bellows and slip-type expansion joints take axial movement in a short space. They are travel-limited, they need guides, and most of them will not tolerate lateral or angular movement at all unless they are specifically built for it.
• Flexible connectors at equipment absorb small movement and vibration at a nozzle. They are not expansion devices for a run.
• A deliberate gap is the right answer at a termination, a tile field, or a sleeved penetration where the movement is small and the fix is clearance rather than a device.

Match the device to the direction the pipe actually moves at that point. An axial device installed where the movement is lateral is a rigid spool with a warranty label on it.

Step 4: Guide the movement so it goes where you sized it to go

What you lose by skipping it: the run buckles sideways instead of moving lengthwise, and you get rub-through, noise, and hanger wear over a season rather than a failure you can point at.

A guide holds the pipe on its axis while letting it slide along that axis. Without guides, a long run under compression takes the easiest path, which is to bow. Expansion joint manufacturers publish guide spacing for their devices, and that spacing is not advice - a bellows that is allowed to buckle sideways fails at a fraction of its rated cycle life.

Step 5: Free the penetrations and terminations

What you lose by skipping it: a cracked wall, a broken sleeve, a patched ceiling. Local damage, real cost, but bounded.

Any place the run passes through structure is an accidental anchor unless you make it not one. Sleeve the penetration, leave annular clearance sized for the movement you calculated, and fill the gap with something that stays flexible rather than something that cures rigid. A rigid-setting compound in a penetration is an anchor you did not draw.

The same logic applies to any material field that expands: a perimeter gap at a floor, a control joint in a slab, a clearance at the end of a rail. The gap is the accommodation.

Step 6: Leave the hangers able to slide

What you lose by skipping it: noise, hanger wear, and insulation damage. Nuisance-grade, but it generates callbacks that are hard to charge for.

A hanger that grips the pipe is a partial anchor. Size the hanger to the pipe with its insulation, do not crush the insulation into a clamp, and use a shield or saddle where an insulated line rests on a support. Most of the ticking and creaking complaints in a mechanical room are a pipe stick-slipping through a support that is too tight.

Step 7: Record the anchor and device locations before you close the work up

What you lose by skipping it: the next tech re-anchors your run. That is a future problem, not a present one, which is why it is last, and it is still worth the two minutes.

Note anchor locations, device travel ratings, and the temperature the system was at when you installed it. That install temperature is the zero point for every later measurement, and without it, nobody can tell whether a device that is sitting at half travel is fine or is already in trouble.

Worked example

A copper hot-water run in a mechanical room. Straight run, 60 feet between the two structural points you can anchor to. Installed on a cool day with the system down, pipe at 55 degrees F. Operating temperature measured at the supply, 180 degrees F.

Movement. Swing is 180 minus 55, which is 125 degrees F. Copper expands about 9.3 millionths of an inch per inch per degree F. Over 60 feet, which is 720 inches, the run grows 9.3 millionths x 720 x 125, or about 0.84 inch.

Anchoring. Step 1 says anchor deliberately. Anchor at the midpoint. Each 30-foot leg now moves about 0.42 inch outward from that anchor toward its own end, and each end needs an accommodation sized for 0.42 inch rather than one point taking 0.84 inch.

Device selection. Suppose the ends are tight and a loop will not fit, so each leg gets an axial expansion joint. The candidate on the shelf is rated for 0.50 inch of axial travel.

Here is the gate, stated so it can be checked: per device, over the full installed-to-peak swing, use no more than 75 percent of the device's rated travel. If you are over, step up to the next travel rating rather than shaving the temperature estimate. The margin exists because rated travel is a limit, not a working figure, and because your installed temperature is a snapshot that will not repeat.

Run it. 0.42 inch of movement against 0.50 inch of rated travel is 84 percent of rating. That is over the 75 percent gate, so the 0.50 inch device does not go in, and the leg gets the next travel rating up.

If you had skipped step 2 and eyeballed it, the 0.50 inch device looks fine - it has more travel than the movement. It would have been installed, it would have worked through the first heating season, and it would have spent every cycle at the top of its range until a cold winter pushed it past the stop.

What the winter swing does to the same job

Change one condition and re-run it, because this is the condition most likely to be missed. The mechanical room is unheated and drops to 35 degrees F on a January night with the system off.

The swing is no longer 125 degrees F, it is 180 minus 35, which is 145 degrees F. Total movement becomes 9.3 millionths x 720 x 145, or about 0.97 inch. Split at the midpoint anchor, each leg moves about 0.49 inch.

Against the same 0.50 inch device that is 98 percent of rated travel, so the device that already failed the gate on the summer numbers is now essentially bottoming out. The point is not that the answer changed - it failed the gate either way - it is that the size of the error changed. The tech who used the operating swing rather than the true swing understated the movement by about 0.13 inch, roughly 16 percent of the 0.84 inch they computed, and on a job where the first number had passed the gate, that shortfall is the difference between a device with margin and a device with none.

The failure mode: the run stops being able to move, so it goes back to step one. The force appears at the pump flange, and the symptom is a flange leak or a cracked nozzle that looks like a gasket problem for the first two visits.

Checking you got this right

• With the system cold, confirm every expansion device is sitting near the cold end of its travel and not centred or extended. A device installed at half travel on a cold line has lost half its capacity before it ever ran.
• Confirm your anchors are attached to structure that will take the force, and that nothing between an anchor and a device is gripping the pipe hard enough to act as a second anchor.
• Bring the system up to temperature and walk the run. You are listening for stick-slip ticking and looking for a hanger that has been dragged out of plumb, both of which mean movement is happening somewhere you did not plan.
• Measure at a penetration: with the run hot, confirm there is still visible clearance in the sleeve on the side the pipe moved toward.
• Write the install temperature on the record. Without it the next person cannot evaluate anything you just did.

References

• 29 CFR 1910.147, energy control, for the requirement to relieve, disconnect, or restrain stored energy before servicing, which covers thermal and spring energy in a run
• Published coefficient of thermal expansion data from the piping or tubing manufacturer for the specific material class
• Expansion joint manufacturer documentation for rated travel, guide spacing, and permitted directions of movement
• See related: Expansion and Contraction in the Field; Thermal Expansion Mismatch Inside a Joint; Pipe Expansion and Thermal Stress Reference