Pipe Thermal Expansion Loop Sizing: PPR Numbers for Hot Climates

A pipe thermal expansion loop is the cheapest thing on a plumbing drawing and the most common thing missing from one. The joints that fail eight months after handover are rarely bad welds. They are good welds carrying a load nobody calculated, because a forty-metre straight run of PP-R grew by a quarter of a metre between commissioning and the first hot summer, and the only place that movement could go was into a fitting.

Here is the number that starts every one of those failures: monolayer PP-R has a coefficient of linear thermal expansion of 0.15 mm per metre per kelvin. Copper is 17 × 10⁻⁶ per °C, which is 0.017. The plastic moves 8.8 times as far as the metal a fitter spent twenty years installing, and nothing on the pipe tells him that.

Composite pipe production at IFAN: the reinforcing layer that changes a pipe's expansion behaviour is built in during extrusion, which is why reinforcement status is a purchase-order decision rather than a site decision. Click to play on YouTube.

How Much a Plastic Pipe Run Actually Moves

Three independent manufacturer manuals give the same figure for monolayer PP-R, which is unusual enough in this industry to be worth saying out loud. NUPIGECO's Niron manual lists αₜ = 0.15 mm/mK. Pestan's technical catalogue prints α = 0.15 mm/m°C. FV-Plast's FV AQUA manual states 0.15 for single-layer PP-R and PP-RCT. Same number, three companies, no coordination.

Put that beside copper. The Lawton Tube data sheet gives copper a coefficient of 17 × 10⁻⁶ per °C. Divide 0.15 by 0.017 and you get 8.8. You will see "ten times copper" repeated across the trade press and it is close enough for a conversation, but if you are sizing an arm from it you are overstating the movement by 13% before you start. Use 8.8, or better, use the coefficients themselves.

The only formula you need

All three manuals state the same expression, which NUPIGECO labels Formula B.1:

ΔL = α × L × ΔT — movement in mm, where L is the run in metres between two fixed points and ΔT is the temperature change in kelvin.

Work a real riser. Forty metres of DN63 monolayer PP-R, installed on a day when the pipe sat at 30 °C, later carrying 70 °C water. ΔT is 40. So 0.15 × 40 × 40 = 240 mm. You do not have to trust that arithmetic either — Pestan's Table 38 has a row for 40 m and a column for ΔT 40, and the cell reads 240 mm.

Two hundred and forty millimetres. That is the length of a large screwdriver, and it is being generated by a pipe that looks completely static. If both ends are welded solid to fittings that are themselves anchored, that movement does not disappear; it converts into stress in the pipe wall and shear at the sockets. That stress arrives on top of the pressure the pipe is already carrying, which is why it is worth reading alongside the pressure and temperature derating rules — the same hot water is reducing the pipe's allowable pressure at the same time.

That last row is deliberately empty. IFAN has published a claim about its reinforced pipe's expansion relative to standard PP-R, which is covered further down, but it has not published a measured coefficient. Filling that cell with somebody else's number would make this table look more complete and make it worth less.

The Hot-Climate Number Everyone Gets Backwards

Ask a contractor in Accra or Lagos whether a hot climate makes thermal expansion worse and you will usually get a yes. The arithmetic says something more interesting, and getting it right is worth real money on a tender.

Look again at ΔT in the formula. Every one of the three manuals defines it the same way: the difference between the temperature at installation and the temperature of the fluid the pipe will carry. Not the outdoor maximum. Not the design ambient. The installation temperature.

Why the hot line gets easier

A European installer fixing pipe in an unheated shell at 15 °C, running 70 °C water, is designing for ΔT 55. His counterpart welding the same pipe in a Lagos building at 38 °C is designing for ΔT 32. On a 40 m run that is 330 mm against 192 mm — the hot-climate job has 40% less movement to absorb on its hot lines, purely because the pipe started warm.

Where it bites instead

The exposure moves to the cold side, and this is the part that gets missed. Chilled water at 7 °C in a building where the pipe was fixed at 38 °C is a ΔT of 31 in the other direction. The pipe contracts, and contraction pulls joints apart rather than buckling the run, so it fails differently and usually later. Air-conditioning risers in tall hot-climate buildings are where this shows up.

• Worst real case: an exposed rooftop or external riser, where the pipe itself is heated by the sun to well above the shade ambient before any fluid enters it, then fed with cold water.
• What insulation does: PP-R has a thermal conductivity of 0.24 W/mK, per Pestan. Lagging an exposed run does not change α, but it does flatten the swing the pipe wall actually sees, so the working ΔT falls.
• The one to record: write the installation temperature on the commissioning sheet. It is the input the whole calculation rests on and nobody ever writes it down.

None of this means a hot market can skip compensation. It means the design case is often the chilled line and the sun-exposed run, not the hot-water riser everyone instinctively worries about. If your consultant sized every loop off a European ΔT assumption, some of your loops are in the wrong places.

Sizing the Expansion Arm: The Constant Nobody Tells You Is Theirs

Once you know the movement, you need somewhere to put it. That means a leg of pipe running perpendicular to the main route, free to flex — an arm at a change of direction, or a U-shaped loop with two of them. The published formula is the same everywhere:

Lₛ = k × √(D × ΔL) — arm length in mm, D the outside diameter in mm, ΔL the movement in mm.

Now the part that is not in any page-one search result. That k is not a constant of physics. It is a number each manufacturer publishes for its own pipe, and they disagree. FV-Plast's manual states plainly: "k material constant, k = 20 for PPR". Pestan's catalogue, for the identical expression, states the constant "for Fuidtherm pipes it is 15".

Run our 240 mm at DN63 through both. √(63 × 240) = 122.96. With k = 20 you need 2,459 mm of arm. With K = 15 you need 1,844 mm. Same pipe size, same movement, 615 mm apart — and the shorter answer is the one that fits more easily into a riser shaft, which is exactly why somebody will pick it.

What to actually do about it

Use the constant your supplier publishes for the pipe you bought. If you cannot get one, use 20 and accept a longer arm, because an oversized arm costs a metre of pipe and an undersized one costs a joint. Never take a formula off an engineering blog and apply it to a different manufacturer's product without checking which k it was written for.

Those figures are FV-Plast's own Table 17, generated with k = 20. Notice the table stops at 90 mm of movement. Our 240 mm run is far off the end of it, and that is the manual telling you something: when a single run generates that much movement, the answer is not one enormous loop. Split the run with an intermediate fixed point and give each segment its own, smaller compensator.

Loop width, and the trick that halves your arm

• Width: Lₖ = 2 × ΔL + 150 mm. The 150 mm is a safety clearance; Pestan calls it Sₐ and gives the same value. FV-Plast adds a second condition — the width must also be at least 10 × the outside diameter.
• Worked: at DN63 with 240 mm of movement, 2 × 240 + 150 = 630 mm, and 10 × 63 = 630 mm. Both conditions land on the same figure, which is a useful sanity check that you have read the diameter column correctly.
• Pre-stressing: Pestan publishes Lₛₖ = K × √(d × ΔL) / 2 — stretch the compensator at installation and you need half the arm. Useful where a shaft is tight.
• The sequencing trap: Pestan is explicit that you form the loop, weld the pipes onto it, then mount, stretch and fix it with the fixed supports, and only then continue the run. Weld it into a restrained line first and you deform the joints you just made.

Fixed Points and Sliding Clips: Where the Movement Is Allowed to Go

A compensator only works if the pipe is told where to move from. That is the job of two different clip types, and on most sites they arrive in the same box and get fitted interchangeably.

A straight run like this is where the arithmetic applies. The distance that matters is not the length of the pipe but the distance between the two points where it is anchored.

A fixed point grips the pipe so it cannot move along its own axis. It is the anchor that defines where a run begins and ends. A sliding support holds the pipe in line but lets it travel lengthwise through the clip. Get these the wrong way round and you have either a pipe with nowhere to expand, or a run with no defined ends and therefore no calculable movement at all.

The commonest field error is subtler than either: a fitting becomes an anchor by accident. Clamp a tee tightly to a wall because it was convenient, and you have created a fixed point the drawing does not show, splitting one designed run into two undesigned ones.

How far apart the clips go

Spacing is not one number per diameter. It falls as the water gets hotter, because a warm pipe sags more between supports. FV-Plast's Table 18b, for PN16 monolayer pipe on horizontal runs:

One caution about reading spacing tables across brands. Pestan's equivalent table is indexed by temperature change, not by water temperature — its DN25 row runs from 105 cm at no change down to 60 cm at a 70 K change. Those are different independent variables and the columns do not line up. Reading one table with the other's assumptions is how a 105 cm figure ends up on a line that should have had clips at 60 cm.

For vertical runs, FV-Plast multiplies its horizontal distances by 1.3. And there is one hard rule on risers that is worth more than the whole table: Pestan warns that sliding supports must not carry vertical lines, because they are not designed to bear load in that direction and can overload and pierce the pipe under the weight of the pipe plus its water. Vertical lines get fixed supports.

How far apart those anchors can sit is also published. Pestan's Table 40 gives the distance one compensator can cover between fixed points on monolayer PP-R: 8 m at DN16, 9 m at DN20, 10 m at DN25, 12 m at DN32, 14 m at DN40. FV-Plast's Table 16 lists exactly the same five figures — and then gives its reinforced pipe 24, 27, 30, 36 and 42 m for the same diameters, which is a three-fold jump and the clearest commercial argument for composite there is.

Monolayer or Glass-Fibre Composite: What the Published Numbers Really Say

This is the decision that reaches your purchase order, so it deserves the honest version rather than the sales version. Reinforced PP-R — a glass-fibre or aluminium layer sandwiched in the pipe wall — really does move less. How much less is where the published numbers stop agreeing.

The reinforcing layer sits inside the wall. From the outside a reinforced pipe and a monolayer pipe of the same diameter can look near-identical on a pallet, which is why the distinction has to be controlled on paper.

Three manufacturers, three numbers

• 0.035 mm/m·K — NUPIGECO's Formula B.1 table for PP-R with fibre glass. Pestan quotes the same value, but explicitly as a minimum.
• 0.05 mm/m·°C — FV-Plast, for multilayer PP-R and PP-RCT.
• 0.085 mm/m·°C — Pestan again, this time from a laboratory test on DN25 pipe with 10% glass fibre in the central layer.

Pestan is doing something unusually candid by publishing both. Its own note explains why they differ: the coefficient depends on temperature, run length, pipe diameter, SDR and, above all, the amount of glass fibre in the central layer. The 0.035 is what the material can do. The 0.085 is what one real tested pipe did.

The gap is a factor of 2.43, and it lands directly on your design. On that 40 m run at ΔT 40, the nominal figure predicts 56 mm of movement and the measured one predicts 136 mm. Design the arm for 56 mm and install pipe that behaves like 136 mm, and you have built a compensator two and a half times too small — which is worse than having built none, because the drawing says the problem is solved.

What IFAN publishes, and what it does not

IFAN states on its product pages that its fiberglass-reinforced PPR-GF-PPR "slashes thermal expansion to 1/4 of standard PPR". Taken against 0.15, a quarter implies roughly 0.0375 — close to the NUPIGECO nominal figure and well below Pestan's tested one. That is a ratio, not a measured coefficient, and this article is not going to convert one into the other. If you are specifying to a tight shaft, ask any supplier — IFAN included — for a measured α on the specific pipe and fibre content you are buying, in writing.

Whichever way that answer comes back, the sourcing consequence is the same. Reinforcement status is a specification, not a grade of the same product, and it has to survive the purchase order. A part-shipment where half the sizes arrive reinforced and half do not is not a like-for-like substitution — it is two different expansion coefficients in one system. Our guide to mixing pipe materials and which transitions hold covers what happens where those two meet, and the full PPR and composite pipe range shows how the lines are actually catalogued when you are consolidating an order.

• Best for composite: long exposed horizontal runs, tall risers with no room for loops, chilled-water lines in hot buildings, anywhere a compensator is physically hard to fit.
• Not worth it: short branch runs, anything chased into a wall, and small-bore final connections — you are paying for a property the installation does not use.
• On the PO: state the reinforcement type per line item, per size. "PPR PN20 DN63" does not say whether it is reinforced; "PPR-GF-PPR PN20 DN63" does.

On sourcing practicalities, so the specification is buildable: IFAN's PPR range covers sizes DN20 to DN160 in pressure classes PN12.5 to PN25, the minimum order is one container with mixed sizes accepted, and the material is 100% virgin PP-R 100 grade with batch certificates issued per shipment. Certification is DIN 8077/8078, ISO 15874, CE and SGS, with SASO, SONCAP or NOM available on request where the destination market requires them.

Two things this article deliberately does not quote are a price and a lead time. Both move with resin cost, size mix and sailing schedule, so they get quoted against an actual size list rather than published as a number that would be stale by the time you read it. And if you need a measured expansion coefficient or a sample before committing, ask for it in writing at the quotation stage. A supplier who can produce one on request is telling you something useful about how well they know their own extrusion.

See how the PPR and composite lines are catalogued

For contractors and distributors specifying a mixed order by the container: the range runs DN20–DN160 at PN12.5–PN25, with monolayer PPR and PPR-GF-PPR composite listed as separate lines so reinforcement status can be fixed per size on the order rather than discovered on site. IFAN accepts one container with mixed sizes.

View the pipe range

The Runs That Need No Compensator At All

Half of the pipe in most buildings needs none of the above, and knowing which half is what stops a specification becoming unbuildable.

Pipe chased into a wall and plastered over is the clearest case. NUPIGECO calls inside-wall installation the most recommended method for monolayer PP-R, and its reasoning is physical: the pipe is in contact with plaster, lime and cement over a large area, so the expansion spreads within the pipe rather than travelling along it. The manual is blunt about the force involved — the expansion does not carry enough force to lift the tiles or crack the plaster. Pestan says the same in different words: with a closed installation the material surrounding the pipe absorbs the elongation.

Insulated pipe in a chase behaves the same way, with the insulation compressing slightly to take up the movement. Two independent manufacturers arriving at the same conclusion is about as much confirmation as this subject offers.

Where the exemption stops

• Exposed risers and shafts — nothing is holding the pipe, so all of the movement is free to travel.
• Ceiling voids and plant rooms — long horizontal runs on brackets, the classic loop territory.
• Rooftop and external runs — the widest temperature swing on the job, and usually the least supervised.
• Anything crossing an expansion joint in the building structure, where the building moves independently of your pipe.

There is also a length below which the arithmetic answers itself. Run the formula on a 3 m branch at ΔT 40 and you get 18 mm. Pestan's own table starts at 10 m for a reason. Short branches, final connections and anything with a natural change of direction in it are usually compensating themselves through ordinary flexibility — the elbow you already have is doing the job of the arm you were about to design.

One note on standards, kept short because they are often over-claimed. ISO 15874-2:2013 is the current published standard for polypropylene pipes in hot and cold water installations, last reviewed and confirmed in 2023 and carrying amendments from 2018 and 2022. IFAN's PP-R is built to it alongside DIN 8077/8078, and our guide to PPR certifications and what each market asks for covers which marks matter where.

What that standard does not do is publish a coefficient of linear thermal expansion. Every α in this article comes from a manufacturer's technical documentation, which is the only place those numbers live. So where compliance or a warranty turns on the figure, confirm it with your supplier and, for a specific project, with the design engineer of record.

Conclusion

Thermal movement is the one load on a plastic pipe system that is fully predictable before anyone arrives on site. Four numbers settle it: the coefficient for the pipe you actually bought, the distance between fixed points, the difference between installation and operating temperature, and the arm constant your manufacturer publishes. Get those four onto the drawing and the joints stop being the weakest thing in the system.

If you are about to place an order, the useful next step is to fix the reinforcement status of every pipe line item before the quotation is signed, and to ask whoever supplies it for a measured expansion coefficient rather than a ratio. You will find out quickly how well any supplier knows their own product.

Frequently Asked Questions