Pressure Testing Plastic Pipe: Why Creep Makes the Gauge Lie

Why a Falling Gauge on Plastic Pipe Is Not Automatically a Leak

Material selection, sizing, insulation and jointing for chilled water branch lines are covered in our pillar guide to chilled water piping. This article deals with one site procedure that fails for reasons unrelated to workmanship: the hydrostatic pressure test. On a plastic line the gauge falls even when the system is sound. The commissioning engineer who reads that fall as a leak spends days chasing a defect that does not exist.

The Plastics Pipe Institute states the mechanism plainly. Creep expansion makes the gauge fall — or forces you to keep injecting make-up water just to hold pressure. It is a property of viscoelastic HDPE that is not present in elastic metallic pipe, and it grows more noticeable in larger diameter HDPE because of the large mass of clean fill water involved. Neither the falling gauge nor the make-up water demand means a leak is present (Plastics Pipe Institute TN-46, Guidance for Field Hydrostatic Testing of HDPE Pressure Pipelines).

Iplex Pipelines Australia frames the same point as a flaw in the method. Standard sustained-pressure (make-up water) tests were designed for linearly elastic materials and are not suitable for viscoelastic materials such as polyethylene. A sealed PE pipe held at test pressure shows pressure decay over time even in a completely leak-free system, purely from strain creep and stress relaxation (Iplex POLIplex PE100 Technical Information: Testing, V1 2019). The PE100+ Association agrees from the specifier’s side: standard water-loss methods are impractical for PE pipe because the material creeps under pressure and the pressure therefore reduces without any leakage, so results can be misleading (PE100+ Association, Testing & Inspection).

Three independent bodies, three documents, one message: on plastic pipe, a pressure drop read on its own — without the clock and the shape of the curve around it — will misdiagnose a sound line. This is not a claim that no numeric drop limit exists. Several of the sources below publish one, and this article quotes them: PIPA caps the drop at 20 kPa, but only across a specified 90-minute rebound observation; Aquatherm allows 9 psi, but only over its 30-minute stage after a cyclic pre-stress, and 3 psi over the 120-minute stage that follows it; ASTM F2164 asks for pressure to hold within 5% variation, but only across the one-hour window that follows a four-hour hold. Every one of those numbers is bound to a stated stage, a stated duration and a stated preparation. Lift the number out of its window — or apply a steel-pipe habit of “no visible drop” — and it stops being a valid criterion. Everything below covers how to measure the shape of the decay curve and which published limit applies to which window.

The Physics: Why the Pipe Steals Pressure From You

A pressurised pipe expands. In steel or copper that expansion is elastic and essentially instantaneous — it happens as you pressurise, stops when you stop, and the gauge then holds. In PE, PP-R and other thermoplastics the expansion keeps going for hours under constant load. The bore grows, the enclosed volume grows, and because water is very nearly incompressible, a small volume increase produces a large pressure fall.

PPI TN-46 quantifies this best in its table of HDPE pipe-grade material modulus E_R, which falls with both temperature and test duration. That dual dependence is creep, expressed as a number:

Source: Plastics Pipe Institute TN-46, HDPE pipe-grade material modulus, imperial units. The percentages in the right-hand column are our arithmetic on PPI’s published values.

Read the corners rather than the rows. Compare 5 °C at one hour (143,600 psi) with 30 °C at three hours (87,024 psi): the modulus has fallen to roughly 61% of its cold, short-duration value. The pipe you test at noon in the Gulf is substantially softer than the one whose datasheet you read, and it softens further the longer you hold the test. That is the whole problem in one comparison.

PIPA’s metric table for PE100 behaves identically — 990 → 900 (×10³ kPa) from 1 h to 3 h at 5 °C, 750 → 680 at 20 °C, and 640 → 600 at 30 °C, with PE80B lower again at 550 → 510 at 20 °C (Plastics Industry Pipe Association of Australia TN005, September 2021, Table 1; the table assumes MDPE characteristics for PE80B and HDPE for PE100). Run the same corner comparison in metric and PE100 at 30 °C after 3 hours retains about 61% of its 5 °C one-hour modulus. The two documents agree closely despite different units and different member companies.

Two consequences the test procedure must handle

• Temperature is a primary variable, not a nuisance. A test run in cool morning conditions and one run in afternoon heat are not the same test. TN-46 requires maximum hydrostatic test pressure to be recorded at the lowest point along the pipeline and compensated for temperatures other than 73 °F (23 °C).
• Duration is not free. Holding longer does not give a more conservative result on plastic. It gives a different result, because E_R at 3 hours is not E_R at 1 hour. Hence the acceptance windows below are tied to specific clock times, and running past them is not the safe choice it appears to be.

Before You Pressurise: The Preparation That Decides the Outcome

More plastic pressure tests are lost to trapped air and thermal drift than to bad joints. Air is the worse of the two because it does not merely add noise — it acts as an accumulator and can mask a real leak. PIPA states it directly: removal of both entrapped and dissolved air is critical, because air can distort test results and mask the presence of a leak (PIPA TN005).

Fill rate

Both bodies limit fill velocity, in compatible terms:

• PPI TN-46 gives a fill-rate formula to avoid air entrainment: Q (gpm) = 0.402 × D², with D the pipe inside diameter in inches, derived from an axial filling velocity of less than 10 feet per minute.
• PIPA TN005 specifies limiting the input flowrate of test water to 0.05 m/s to minimise entrapped air, using firm foam pigs.

Worked example on our largest size. For Φ110 HDPE PN16 marked to DIN 8077/8078 (per our catalogue), the wall is 10 mm, so the bore is 110 − 2 × 10 = 90 mm = 3.54 in. Using the TN-46 formula: Q = 0.402 × 3.54² = 0.402 × 12.55 ≈ 5.0 gpm, roughly 19 litres per minute. That is a deliberately slow fill. A site hose that fills the line in minutes rather than hours puts air into the pipe that you will argue with for the rest of the day. Our arithmetic on PPI’s formula, using our catalogued 110 × 10 mm pipe dimension; substitute the true bore for your own DR — and verify it against the delivered pipe — before using the result.

Air limit and stabilisation

• PPI TN-46 Phase I requires purging air to less than 4% trapped air by volume, then allows 3 to 24 hours after filling for thermal equilibrium and for dissolved air to exit the vents.
• Iplex requires a filled PE pipeline to stabilise at its temperature for a period varying with pipe size and weather, with a minimum of 2–3 hours allowed, although it is preferable to delay testing until the following day (Iplex POLIplex, V1 2019).

“Preferably the following day” is the sentence most commissioning programmes ignore. On a chilled water branch just filled with mains water in a hot plantroom, the fluid is still equilibrating with the structure. You are not testing the pipe. You are watching a thermometer through a pressure gauge.

Instruments and gauge position

• Iplex: pressure gauges should be readable to ±10 kPa.
• Aquatherm, for PP-R building services piping: gauge accuracy must be within 0.5 psi, and the gauge must sit near the lowest point of the system where pressure will be highest, to reduce the risk of over-pressurising (Aquatherm Standard Pressure Testing Procedure, rev. 1 September 2024; Aquatherm Technical Bulletin 201802A-AQTTB).

A gauge reading in 0.5 bar increments cannot resolve the acceptance criteria in the rest of this article. PIPA’s fallback criterion is a maximum allowable pressure drop of 20 kPa over 90 minutes — you cannot adjudicate that on a coarse gauge. Instrument resolution is a pass/fail item before the test starts.

Safety: do not substitute air for water

PIPA advises that pneumatic testing should be avoided due to the substantial energy stored in compressed gas. Aquatherm limits preliminary compressed-air leak checks on PP-R to a maximum of 15 psi (1.0 bar), warning that temperature changes can significantly affect the pressure of compressed air and that exceeding 15 psi can result in a hazardous condition with risk of injury or death. The main cyclic, 30-minute and 2-hour hydrostatic tests must be performed with water only, due to its incompressibility, and SDR 17.6 should never be tested with compressed air alone. Aquatherm further requires that all air be completely purged before testing, because entrapped air can result in significant pressure surges and violent, dangerous and catastrophic rupture, and that unrestrained exposed pipe beside an exposed connection must not exceed the greater of 5 pipe diameters or 3 ft (1 m).

Setting the Test Pressure

Three sources, three formulations, all anchored to the weakest component rather than to the pipe:

Aquatherm also caps test pressure absolutely for high-rise or high-pressure systems: PP-RP (RCT) SDR 9 = 400 psi (27.6 bar); PP-RP (RCT) SDR 11 = 320 psi (22.1 bar); PP-RP (RCT) SDR 17.6 = 200 psi (13.8 bar).

What the test pressure does to the pipe wall

TN-46 gives the hoop stress the test imposes:

Estimated pipe hoop stress = (hydro-test pressure × (DR − 1)) / 2, where DR is the pipe dimension ratio.

It also shows how close a 1.5× test comes to the material’s long-term limit: PE3608 with HDS 800 psi at 73 °F × 1.50 = 1200 psi hoop stress, which is 75% of HDB; PE4710 with HDS 1000 psi × 1.25 = 1250 psi (78% of HDB); PE4710 × 1.50 = 1500 psi, which is 94% of HDB. A 1.5× test on PE4710 is not a comfortable margin. It is a deliberate excursion to within 6% of the hydrostatic design basis — defensible briefly, poor for eight hours. Another reason the acceptance windows are time-bounded.

The temperature correction, and why chilled water is the favourable case

PE pressure rating is PR = 2·HDS·f_E·f_T / (DR − 1), where PR is the pressure rating in psi, HDS the hydrostatic design stress at 73 °F, f_E the environmental design factor, f_T the operating temperature multiplier, and DR the dimension ratio (pipe outside diameter divided by minimum wall thickness). Maximum recommended operating temperature for sustained internal pressure is 140 °F (60 °C) (WL Plastics WL118, Pressure Rating, rev. 24-08).

The multiplier f_T rises below the 73 °F base: 1.28 at 0 °C/32 °F; 1.22 at 5 °C; 1.15 at 10 °C (1.15 at 50 °F); 1.12 at 13 °C; 1.03 at 20 °C; 1.00 at 23 °C/73 °F; 0.93 at 30 °C; 0.82 at 40 °C; and 0.64/0.63 at 60 °C/140 °F (WL Plastics WL118).

This produces a specific trap for chilled water commissioning, and it is an engineering observation rather than a citation: the pipe in service is stronger than the pipe under test. A line running at 6–7 °C (42–45 °F, UFC 3-430-09 Table 2-1) sits near an f_T of about 1.20. The same line tested in a 30 °C plantroom before the chillers run sits at 0.93 — a ratio of roughly 1.29 between service-condition and test-condition allowable rating, on WL Plastics’ published multipliers. Designers who set test pressure from the cold-service rating and then test warm quietly eat that entire margin. Set the test pressure from the conditions on the day of the test, not the conditions in the design brief. (Our reading of the WL118 table; the multipliers are WL Plastics’, the inference is ours. On the supply band: the Fahrenheit values and their metric equivalents are both published by UFC 3-430-09, which gives 42–45 °F as 6–7 °C. That band is a US federal design criterion, not a global default — European practice commonly works to the BSRIA BG 30/2007 6–12 °C basis, and the two should not be mixed in one calculation.)

Method 1 — The Rebound Test (PPI TN-46 and PIPA TN005)

The rebound test is the most useful method for building services work because it turns creep from a nuisance into the signal. PIPA notes it is one of two methods independent of soil support influence — the modified rebound method developed within CEN, and the Swedish VAV P78 method — in contrast to the UK IGN 4-01-03 pressure-decay approach, which PIPA describes as complex, requiring an estimate of the uniformity of soil support, and not providing data support for the curve slope coefficients used to determine pass or fail.

Phase I — Preliminary

PPI TN-46 Phase I, step by step:

• Purge air to less than 4% trapped air by volume.
• Hold the pipeline full but unpressurised for 60 minutes before test start.
• Pressurise to specified test pressure (STP) in less than 10 minutes.
• Hold STP for 30 minutes by repeatedly injecting small make-up water volumes as the HDPE creep-strains.
• At T = 30 min, valve off and record P30.
• Allow decay for 60 more minutes; record P90 at T = 90 min.
• Pass Phase I if P90 > 70% of STP.

PIPA TN005’s preliminary phase is the same test with the clock read differently: reduce pipeline pressure to atmospheric and let stand 60 minutes; raise to STP in less than 10 minutes and hold 30 minutes, pumping as needed; inspect for leaks; shut off pressure and let stand 60 minutes; measured P60 must be greater than 70% of STP.

Internalise the 70% threshold. It is the closest thing the literature offers to “how much drop is normal”, and it is generous by metal-pipe standards: a sound plastic line is permitted to shed up to 30% of test pressure in an hour of free decay. An inspector applying a steel-pipe habit of “no visible drop” will reject sound pipe every time.

Phase II — Was it air, or was it the pipe?

Once the line decays acceptably, Phase II asks whether the volume behaviour is consistent with water and pipe alone, or whether there is air in there. TN-46’s air volume assessment:

ΔV_{max allowable} = 1.2 · V · ΔP · [(1/E_w) + (D / (t · E_R))]

where 1.2 is the air allowance factor, V is the test section volume of water (gallons), ΔP the measured pressure drop (psi), t the nominal pipe wall thickness (inches), D the pipe inside diameter (inches), E_w the bulk modulus of water (psi) and E_R the HDPE pipe modulus (psi). Phase II passes when measured ΔV < ΔV_{max allowable}.

PIPA gives the metric form: ΔV_{max allowable} = 1.2 · V · ΔP · [1/E_W + D/(t_e · E_R)], with V in litres, ΔP in kPa, D the internal diameter in m, t_e the wall thickness in m, and both moduli in kPa.

Bulk modulus of water by temperature, from both documents:

Notice the relative magnitudes. Water’s bulk modulus is of order 300,000 psi; the HDPE pipe modulus is of order 100,000 psi and falling with time. The term D/(t·E_R) — the pipe’s contribution — dominates the bracket for any thin-walled pipe. The pipe, not the water, is where your injected volume goes. The formula tells you what the gauge is already telling you.

Phase III — The rebound itself

This is the diagnostic step, and it is elegant: drop the pressure and watch it come back up.

Per TN-46, after dropping pressure 10–15%, the internal pressure should stabilize and remain within ±5% of the Phase III reduced test pressure. The HDPE molecules were stretched during the expansion phase; on lowering the pressure they elastically revert, compressing the test water and causing pressure to rebound and rise slightly. A continuously falling rebound plot means the pipeline is leaking and the test fails. The standard Phase III observation period is 30 minutes.

PIPA’s version: after reducing pressure by a ΔP of 10–15% of STP in under 5 minutes and recording the bled volume ΔV, observe the pressure rise for 30 minutes. The pipeline passes if there are no leaks, no components break, and the internal pressure rises or remains static over 30 minutes. If in doubt, leave it for 90 minutes — the maximum allowable pressure drop in 90 minutes is 20 kPa.

The trap at 90 minutes

Here is the most counter-intuitive item in this article, explicit in TN-46: if the Phase III observation is extended to 90 minutes, the initial pressure rise may be followed by a flat stable pressure and then a slight drop as viscoelastic creep resumes later in the extended period. An over-long observation window can make a sound pipe look like it is failing.

So the diligent engineer who thinks “I’ll watch it a bit longer to be sure” is not being conservative. He is walking into a second creep regime and generating a false failure. Note also the tension between the two documents. TN-46 warns that the 90-minute window can produce a spurious drop; PIPA offers 90 minutes as the tie-breaker with a 20 kPa cap. These reconcile — PIPA’s 20 kPa allowance is precisely the room needed for the resumed creep TN-46 describes — but only if you apply PIPA’s numeric allowance rather than a “must not fall at all” reading. Decide which document governs your contract before you are standing at the gauge.

Method 2 — Log-Log Slope Analysis (WRc Three-Point Method)

The rebound test gives a yes/no. Slope analysis tells you which failure mode you have — leak or air — which makes the extra arithmetic worthwhile when a line has already failed once.

The foundation, from Iplex: the PE pressure-decay curve is non-linear in linear axes, but plots as a straight line on log-log axes, and the slope of that line is the diagnostic that separates a sound pipe from a leaking one. A steeper slope than the sound-pipe band indicates probable leakage; a shallower or decreasing slope indicates entrapped air acting like an accumulator.

The procedure

WRc three-point slope analysis, per Iplex. Take P1 at t1 = t_L (the pressure loading time), P2 at t2 ≈ 7 × t_L, and P3 at t3 ≥ 15 × t_L. Because the pipeline begins relaxing during pressurisation, each time is corrected by adding 0.4 t_L. Then:

• n1 = (log P1 − log P2) / (log t2c − log t1c)
• n2 = (log P2 − log P3) / (log t3c − log t2c)

Acceptance bands

Iplex gives the acceptance bands for n1 and n2 in a sound, leak-free PE main. They depend on restraint:

Two interpretation rules come with the table. Values significantly below the minimum indicate too much air in the main. A linear slope falling between the two bands indicates poor backfill compaction but is not a failed test — worth quoting verbatim to any inspector minded to reject on it.

For chilled water work inside a building, note which band applies. Exposed plantroom and riser pipework is unconstrained, so you work to 0.08–0.10, roughly double the decay rate permitted for a buried main. Exposed plastic decays faster because nothing outside the pipe resists its expansion.

Worked example

Take an exposed PE branch, loading time t_L = 15 minutes, test pressure 12.00 bar.

Time points: t1 = 15 min, t2 = 7 × 15 = 105 min, t3 = 15 × 15 = 225 min. Correction = 0.4 × 15 = 6 min, so the corrected times are t1c = 21 min, t2c = 111 min, t3c = 231 min. The log spans are log(111) − log(21) = 0.7231 and log(231) − log(111) = 0.3183.

Our arithmetic, applying Iplex’s WRc formulae and bands to an illustrative t_L and test pressure; the formulae, the 0.4 t_L correction and the acceptance bands are Iplex’s. These are constructed examples, not measurements from an installed system.

Dwell on the third row. The “trapped air” line held 11.62 bar of 12.00 — it lost barely 3% and looks, to anyone reading the drop on its own, like the best result of the three. It is the only one that must be rejected outright and re-done. The flattest curve is the most suspicious curve. That inversion is impossible to reach intuitively, and it is why a bare “maximum permissible drop” clause written for steel — one with no stage, no duration and no slope attached — does real damage when pasted into a plastics specification. The published plastics limits quoted in this article are not that kind of clause: each is tied to a defined window, and PIPA’s and Aquatherm’s are applied alongside a “steadily decreasing pressure” test that catches exactly the case a lone number misses.

The predictive form

Iplex also gives a closed form for the expected pressure at any time during a PE decay test:

P = P_L [2.5 (t/t_L) + 1]⁻ⁿ

where P is the predicted pressure at time t, P_L the test pressure at the moment test pressure is first reached, t the time from reaching test pressure, t_L the loading time, and n = 0.04 for pipes installed in compacted soil or n = 0.10 for pipes installed without support. Recorded pressure below predicted indicates probable leakage; above predicted indicates air entrapment.

Normalising P_L to 1.000 with t_L = 15 min, the predicted decay is:

Our arithmetic on Iplex’s formula, at t_L = 15 min. Recompute for your own loading time — the curve shape depends on it.

This table answers “how much drop should I expect?”, and it explains the 70% rebound criterion. An unsupported line at n = 0.10 is predicted to sit at 0.787 of test pressure at one hour and 0.738 at two hours, so a P90 threshold of 70% is a floor placed just under the predicted sound-pipe curve, not an arbitrary round number.

Method 3 — Sustained Pressure and Make-Up Water, Where It Still Applies

ASTM F2164, as summarised by PPI TN-46: fill and thermally stabilize the pipeline with no air entrapment, pressurize at test pressure for 4 hours, slightly reduce the pressure, then observe the pressure for one hour to remain essentially constant, within 5% variation, for an acceptable test. The four-hour hold does the same job as the rebound test’s 30-minute hold — working the creep out before the measurement window opens. Note that F2164’s one-hour window is the observation period of this procedure, reached after a four-hour hold. It is not an extension of the rebound test’s Phase III window, and the two should not be swapped: a one-hour Phase III observation without F2164’s four-hour hold in front of it runs into the resumed-creep effect described above.

Where make-up water volume is the contractual criterion, PIPA gives an allowable rate for PE pipe where joints are not accessible for inspection, over a minimum test period of 2 hours:

Q < 0.14 · L · D · H

where Q is allowable make-up water in litres/hour, D the pipe internal diameter in m, L the length of pipeline under test in km, and H the average test head over the length under test in m. Where joints are accessible and no leaks are evident after at least 15 minutes, the section may be deemed to pass — worth remembering for exposed plantroom pipework, where the cheapest valid test is often a competent visual inspection.

For referee purposes and to quantify a leak rate, PIPA describes the Reference Test Method (VAV P78, the Janson method): raise to STP and close off, allow the pipeline to settle at least 12 hours (the gauge will show a drop), inspect for leaks; then raise to STP and maintain for 5 hours, recording make-up water volume between 2 and 3 hours and between 4 and 5 hours. The pass criterion is:

ΔV(5h − 4h) ≤ 0.55 · ΔV(3h − 2h) + V_{leak 1h}

The method has been used in Sweden since 1989, in diameters to DN800 and lengths to 3,000 m (PIPA TN005). See the logic: it does not ask how much water went in, it asks whether the rate of water demand decayed by the factor a creeping-but-sound pipe should produce. A sound pipe’s appetite for make-up water falls off; a leaking pipe’s does not.

PP-R and Building Services: The Aquatherm Three-Stage Procedure

Branch-level chilled water piping in buildings is frequently PP-R, and the PP-R procedure handles creep differently. Instead of measuring it, it deliberately exhausts it first.

Stage 1 — Cyclic pressure test

Bring to test pressure for 2 minutes, reduce to 15 psi for 2 minutes, release — repeated three times — then bring to test pressure for 5 minutes and reduce to 15 psi for 5 minutes. Aquatherm states this test is explicitly intended to expand and stress the system and joints, so additional pump pressure may be necessary to maintain test pressure initially (Aquatherm Standard Pressure Testing Procedure, rev. 1 September 2024).

Stage 2 — 30-minute test

Bring the system to test pressure; it will expand slightly once up to pressure, so additional pressure may be required to help it stabilize. Once stabilized, observe 30 minutes. The loss of more than 9 psi (0.62 bar) — or 6 psi (0.41 bar) for SDR 17.6 systems — or steadily decreasing pressure, is indicative of a leak.

Stage 3 — 2-hour test

If the system lost pressure during the 30-minute test, bring it back to test pressure, then observe for 120 minutes. The loss of more than 3 psi (0.21 bar) or steadily decreasing pressure is indicative of a leak; the test pressure must have less than 3 psi loss and must have stabilized at a value of less than 3 psi loss during the test.

Why the allowance gets tighter as the test gets longer

This detail gives the whole procedure away. The allowance goes from 9 psi over 30 minutes to 3 psi over 120 minutes — as an allowable rate, 0.30 psi/min tightening to 0.025 psi/min, a factor of twelve (our arithmetic on Aquatherm’s published limits). No leak gets smaller with time. The allowance tightens because creep has largely been worked out by the cyclic and 30-minute stages, so by stage 3 the remaining pressure loss should be almost entirely leakage.

Every stage is a creep-management strategy: cycle it to force the fast strain, hold it 30 minutes with a loose allowance to absorb the rest, then measure properly against a tight allowance. Skipping stage 1 to save twenty minutes carries unrelieved creep into stage 3, where the allowance is twelve times tighter. That is exactly how a sound PP-R system fails its own test.

The Rule With No Metal-Pipe Equivalent: Recovery Before Re-Test

If a plastic pressure test fails, you cannot simply fix the suspect joint and re-pressurise that afternoon. Iplex: if a PE pressure test must be repeated, it should not be re-attempted until the pipeline has recovered from the previously imposed conditions, and a recovery period equivalent to 5 times the previous total test period may be taken as a guide.

The arithmetic bites hard. A failed 4-hour test buys a 20-hour recovery before the re-test may begin — a next-day activity, minimum. A full Janson reference test at 12 hours settling plus 5 hours running is a 17-hour test period, implying an 85-hour recovery, most of a working week.

The reason is viscoelastic. The pipe that just failed is not the pipe you started with. It has been creep-strained and its modulus has fallen, per the E_R tables above. Re-testing it immediately means testing a softened pipe against acceptance criteria derived for a relaxed one, which tends to produce a second failure and a third round of joint-hunting. This has no counterpart in metal pipe testing, and in our experience it is the single most commonly violated requirement on site, because nothing about a plastic pipe looks tired.

Programme implication, stated as an opinion: commissioning schedules for plastic pipework should carry an explicit re-test contingency of at least one full day per test section. A programme that assumes first-time pass on every section has no slack for a rule published in the manufacturer’s own literature.

Reading the Result: A Diagnostic Table

The One Failure You Must Not Sit and Watch

Everything above tells you to be patient with a falling gauge. There is a single exception, and it is a safety matter rather than an interpretation matter.

Aquatherm states that heat fusion is essentially an all-or-nothing joining process: an improper fusion will hold very little pressure and will fail prematurely when subjected to pressure, while a proper fusion becomes at least as strong as the pipe itself and can easily pass a system pressure test. Leakage at a fusion joint indicates a possible poor joint with imminent potential for complete separation — installers should move away immediately and depressurize (Aquatherm Technical Bulletin 201802A-AQTTB).

The consequence is that a fusion joint has no partial credit. A steel weld can leak a little and be a poor-but-holding weld. A weeping fusion joint is not 90% made — it is on its way to opening completely, on a line at up to 1.5× design pressure. That is the one signal on the whole test sheet that overrides “wait and observe”.

A Note on Drainage Pipework: EN 1610

Chilled water plantrooms carry condensate drainage, tested to an entirely different regime. EN 1610’s water test (method W) sets the test pressure as that resulting from filling the test section to ground level of the upstream or downstream manhole, with a maximum of 50 kPa and a minimum of 10 kPa measured at the pipe invert, pressure maintained within 1 kPa of test pressure, for a duration of (30 ± 1) minutes. Allowable water addition is 0.15 l/m² for pipelines, 0.20 l/m² for pipelines including manholes, and 0.40 l/m² for manholes and inspection chambers, where m² refers to wetted internal surface (via PlugCo, Pipeline Leak Test as per EN 1610).

The air test (method L) uses an initial pressure approximately 10% in excess of the required test pressure p₀, held for approximately 5 minutes before adjusting to test pressure, with pressure drop measurement within 10% accuracy and time measurement within ±2.5 s. The four air test methods are LA = 10 mbar, LB = 50 mbar, LC = 100 mbar and LD = 200 mbar test pressure (via Esders GmbH, DIN EN 1610 leak test of wastewater pipes).

The per-method, per-diameter permissible pressure drops and durations are given in Table 3 of EN 1610 itself, a paywalled CEN document. We have not reproduced those values because we could not verify them at source — obtain the standard. Coming soon if we are able to license and publish a compliant summary. Note the essential difference from everything above: these are low-pressure tightness tests on non-pressure pipe, where creep is not the governing effect. Do not carry the EN 1610 mindset onto a pressure line, or the plastic-pipe mindset onto a drain.

What This Means for IFANNova Pipe on Your Site

IFANNova is a French brand; our pipe and fittings are manufactured by Zhuji Fengfan Piping in Zhuji, Zhejiang — over 30 years in production, 1000+ employees, shipping to 118+ countries from a 120,000 m² facility. Everything in this article applies to our plastic lines exactly as it applies to any other manufacturer’s, because the governing physics is the material’s, not the brand’s.

Where our range sits, and the honest limits (all figures per our catalogue):

• HDPE PN16 marked to DIN 8077/8078 (per our catalogue), Φ20–110, with 603/604 weld-free compression fittings. The PE-specific methods above — rebound, log-log slope, the Iplex prediction formula, the 5× recovery rule — apply most directly to this line. With compression rather than fusion joints, PIPA’s “joints accessible, 15 minutes visual” allowance is often the pragmatic route on exposed work.
• PPR PN20, where the Aquatherm-style three-stage philosophy is the relevant model. Our 1103 pipe range is only 20, 25 and 32 mm, with 75 fitting items in the 1138 series.
• UPVC 806 PN16, Φ20–110, with 203 fitting items in the 1806 series. Solvent-cement joints, and no fusion plant required.
• PVC 902 drainage, 902 pipe Φ32–110 with 1902 fittings Φ32–160 — non-pressure drainage, excluded from chilled water pressure duty; the EN 1610 regime, not the pressure regime.
• PEX 2114/2121 (16–32 mm) and brass 2405.

The limit, stated plainly: our pressure-pipe ceiling is Φ110 — UPVC 806 PN16 and HDPE PN16, both Φ20–Φ110; PPR PN20 stops at 32 mm. Our PVC 902 line reaches Φ160 in the 1902 fittings only — the 902 pipe stops at Φ110 — and 902 is non-pressure drainage and is excluded from chilled water pressure duty (per our catalogue). We cannot supply DN150–400 chilled water mains. If your project needs main-header pipe, we are not your supplier for that scope, and no amount of test methodology changes it. We supply branch and terminal-level piping, and we would rather you know that before the submittal stage.

Certifications on record: SKZ, CE, WRAS, DVGW, SGS, ISO 9001 and ISO 14001. Certificate numbers: Coming soon. If your commissioning documentation requires certificate numbers against a pressure test record, request them from us directly rather than assuming they can be cited from this page. Insulation can be supplied alongside the pipe, but published specifications are Coming soon.

We publish no test data of our own in this article. Every numeric acceptance criterion above belongs to PPI, PIPA, Iplex, Aquatherm, WL Plastics or EN 1610, and is attributed accordingly. Where a figure could not be verified at source — EN 1610 Table 3 in particular — it is marked Coming soon rather than estimated.

Pressure Testing Plastic Pipe: Frequently Asked Questions

How much pressure drop is normal on a plastic pipe test? There is no single generic percentage, and any source offering one should be treated with suspicion. The published criteria are indirect: P90 must exceed 70% of STP in the PPI TN-46 rebound test; the log-log decay slope must sit in 0.08–0.10 unconstrained or 0.04–0.05 in compacted backfill per Iplex/WRc; and the expected value at any moment is given by P = P_L[2.5(t/t_L)+1]⁻ⁿ. A sound unconstrained line is predicted to sit near 0.79 of test pressure after an hour at n = 0.10.

Why does my plastic pipe keep needing make-up water? Because it is still creep-straining and the bore is still growing. PPI TN-46 Phase I explicitly instructs you to hold STP for 30 minutes by repeatedly injecting small make-up water volumes as the HDPE creep-strains. The demand for make-up water is the expected behaviour of a sound pipe, not evidence of a leak.

Can I test plastic pipe with compressed air? PIPA advises avoiding pneumatic testing due to the substantial energy stored in compressed gas. Aquatherm permits a preliminary leak check on PP-R to a maximum of 15 psi (1.0 bar) only, warns that exceeding it can create a hazardous condition with risk of injury or death, and requires the main tests to be performed with water. SDR 17.6 should never be tested with compressed air alone.

My pressure barely dropped at all — is that a good result? Probably not. A slope below the acceptance band, or a recorded pressure above the predicted curve, indicates entrapped air acting as an accumulator (Iplex). Air can mask a real leak (PIPA TN005). A suspiciously flat curve on plastic warrants re-venting and re-testing.

How soon can I re-test after a failure? Not until the pipeline has recovered — Iplex gives a guide of 5 times the previous total test period. A failed 4-hour test implies a 20-hour wait.

Does testing in a hot plantroom matter? Yes, in two ways. The pipe modulus falls with temperature, from 143,600 psi at 5 °C/1 h to 87,024 psi at 30 °C/3 h — about 61% (PPI TN-46) — so a warm pipe creeps more. And the temperature multiplier f_T falls from 1.00 at 23 °C to 0.93 at 30 °C (WL Plastics WL118), reducing the allowable rating. TN-46 requires test pressure to be temperature-compensated from 73 °F (23 °C).

Should I watch the rebound longer than 30 minutes to be safe? No — that is the counter-intuitive one. PPI TN-46 notes that at 90 minutes the initial rise may be followed by a plateau and then a slight drop as viscoelastic creep resumes, so an over-long observation window can make a sound pipe look like it is failing. Judge at 30 minutes, or apply PIPA’s 20 kPa/90-minute cap.

What if a fusion joint is weeping slightly? Treat it as urgent, not marginal. Per Aquatherm, leakage at a fusion joint indicates a possible poor joint with imminent potential for complete separation; move away immediately and depressurize.

Request Technical Support for Your Test Procedure

If you are specifying or commissioning branch-level chilled water piping within our range — HDPE, PPR, UPVC or PEX, to a pressure-pipe ceiling of Φ110 (per our catalogue) — contact us with your line sizes, SDR/PN, test pressure and site temperature conditions. We will confirm what our published data supports and where you will need the source standards. For the underlying material and sizing decisions that precede any of this, start with our chilled water piping guide.