BRENNEN

Industries / Industrial equipment and services / Industrial services / Heat exchanger services

Tube testing

Tube testing is checking an exchanger's tubes for thinning, pits, cracks and leaks, usually by pulling a probe through each tube from one end.

7

Milestones achieved, from 1951 to 2024.

1

To watch for: announced, not yet confirmed outside.

2

Open deliverables. Nobody has shipped them.

The road

AchievedWatch forOpen

Achieved

  1. 1951

    Reading steel tubes from inside

    William R. MacLean's US patent 2,573,799, filed in April 1946 and granted on 6 November 1951, measures the wall of a steel pipe with coils inside it. Ordinary eddy current testing, in which a coil sets up electric currents in the metal and senses how flaws disturb them, sees poorly into magnetic steel; in the remote field method a pick-up coil two to three tube diameters from the driving coil senses the field that has passed through the wall, so flaws on the inside and outside show equally. A 1996 conference paper dates the remote field technique to this patent, and cites Thomas R. Schmidt's 1984 description of it in Materials Evaluation.

    image-ppubs.uspto.govlink.springer.com

    Credit

    William R. MacLean; Thomas R. Schmidt

  2. 1975

    Check 3% of the tubes

    The US Nuclear Regulatory Commission's Regulatory Guide 1.83 of 1975 told reactor operators to test at least 3% of the tubes in each steam generator, a reactor's large heat exchanger, along their full length, with probes "sensitive enough to detect imperfections 20% or more through the tube wall", and to plug any tube with a flaw deeper than 40% of the wall. If more than 10% of the sample showed flaws, 3% more were tested, then 6% more. An IAEA review notes that reliable detection at 20% depth was beyond the methods of the day, and that by the early 1980s US plants drew at least half of each sample from areas where damage was expected.

    Lessona random sample is only as good as the probe and the guess about where damage lies.

    www-pub.iaea.org

    Credit

    US Nuclear Regulatory Commission

  3. 1982

    A loose part cuts tubes

    A steam generator tube at the Ginna nuclear plant ruptured in 1982 and leaked reactor coolant at about 390 US gallons a minute; an Idaho National Laboratory review lists the cause as loose parts and tube wear. The same review tabulates twelve tube ruptures in pressurized water reactors from 1975 to 2008, and notes that objects left on the outside of the tubes kept causing failures after water chemistry had been improved.

    Lessona tube can be worn through from outside by debris between inspections, so the space around the tubes needs checking as well as the tubes.

    osti.gov

    Credit

    Hossam H. Abdellatif and three co-authors, University of Idaho and Idaho National Laboratory (review)

  4. 1982

    Ultrasound turns inside the tube

    IRIS, the internal rotary inspection system, fills a tube with water and spins a mirror that sweeps an ultrasonic beam around the wall, so the wall's thickness is measured all the way round as the probe is drawn out. A May 1982 article credited to MatEval Ltd introduced it for air-cooled heat exchangers; the trade site Inspectioneering dates its first field use to 1978. A 2026 Brazilian study calls it the method often used to measure remaining wall in carbon steel exchanger tubes.

    api.crossref.orginspectioneering.comnew.scielo.br

    Credit

    MatEval Ltd

  5. 2000

    Remote field gets a standard

    Dave Russell formed a group in 1984 to apply the remote field method to steel tubes, and his company, Russell NDE Systems, says it sold the first commercial remote field product in 1988. ASTM's first version of E2096, in 2000, set out how to examine installed magnetic steel exchanger tubes 12.7 to 50.8 mm across by remote field testing; it leaves the pass and fail limits to the owner and the inspector.

    qualitymag.comstore.astm.org

    Credit

    Dave Russell and Russell NDE Systems; ASTM Committee E07

  6. 2013

    Sound sent through the air

    ASTM E2906, first issued in 2013, covers acoustic pulse reflectometry: a pulse of sound is sent down the air inside each tube, and its echoes show holes, pitting, wall loss and blockages. It needs only an open tube end, works on tubes 12.7 to 101.6 mm across and up to 18 m long, of metal, graphite or plastic, and may miss tight cracks.

    store.astm.org

    Credit

    ASTM Committee E07

  7. 2024

    Steel tube methods tested blind

    HOIS, an industry inspection program run through ESR Technology in the UK, ran blind trials of seven methods on plain and finned carbon steel tubes and low alloy steel tubes: IRIS, remote field, remote field array, magnetic flux leakage, near field array, partially saturated eddy current and borescope. Eight companies took part, on ex-service tubes with real damage and on tubes with machined flaws whose depths had been measured by ultrasound and by hand. The report on how each method did is confidential to HOIS members.

    esrtechnology.com

    Credit

    HOIS and ESR Technology; trial participants Baugh & Weedon, CAN, Delta Test, ETherNDE, Oceaneering, Waygate Technologies, Versa Integrity and Eddyfi

Watch for

  1. Remote field matches IRIS

    On 16 tubes from a refinery's carbon steel bundle, remote field testing found the same corrosion as IRIS, judged against the tubes' cut sections measured with a 3D scanner, and three inspectors of different experience read the remote field data alike. The flaws were first located by remote field, so flaws it missed were not sampled.

    Achieved whena larger trial, with flaws chosen without help from remote field and its results published, finds the same.

    new.scielo.br

    Claimed by

    Rafael Nardon Ferraresi and four co-authors, Federal University of Uberlândia and State University of Goiás, with Petrobras, in Soldagem & Inspeção, accepted February 2026

Open

  1. 01

    Steel tube methods, tested publicly

    A public blind trial for carbon steel exchanger tubes. At least 100 retired tubes are scanned by four or more methods (remote field, near field, magnetic flux leakage, IRIS), each run by a different team that has not seen the others' results. Every tube is then cut into short pieces and its wall mapped with a 3D scanner, and the scans, the cut-tube truth and each method's hits and misses are published.

    The brief
    Credit · unclaimed
  2. 02

    An open acoustic tube tester

    An open-hardware acoustic pulse reflectometry tester: a small loudspeaker and a microphone in an adapter held to the tube end, a USB audio interface, and Python code that sends a click, records the echoes and reports the distance to holes, wall loss and blockages. Tested only on retired tubes on a bench.

    The brief
    Credit · unclaimed

Open 01 · the brief

Steel tube methods, tested publicly

Credit · unclaimed

The first to publish is credited on this line, and the deliverable moves from Open to Achieved.

What to build

A public blind trial for carbon steel exchanger tubes. At least 100 retired tubes are scanned by four or more methods (remote field, near field, magnetic flux leakage, IRIS), each run by a different team that has not seen the others' results. Every tube is then cut into short pieces and its wall mapped with a 3D scanner, and the scans, the cut-tube truth and each method's hits and misses are published.

Why it is still open

HOIS ran such trials with eight companies, but its report is confidential to members. The 2026 Uberlândia study used 16 tubes, and cut only at spots that remote field had already found.

esrtechnology.comnew.scielo.br

Built with what exists

Retired, cleaned carbon steel tubes from a retubing shop or scrap yard; remote field, near field and IRIS instruments such as Eddyfi's Ectane 2 (used in the 2026 study), run by volunteer service companies; a band saw; a handheld 3D scanner such as Creaform's HandySCAN (also used in the 2026 study); Python for detection-rate curves.

  1. First step

    Ask a retubing shop for 20 retired carbon steel tubes from one scrapped bundle, number them, and have one remote field scan and one IRIS scan made before any tube is cut.

  2. Done looks like

    At least 100 tubes and 300 recorded flaws; for each method, the share of flaws found at 20%, 40% and 60% wall loss and the average sizing error; every scan file and the truth table posted under an open license.

Open 02 · the brief

An open acoustic tube tester

Credit · unclaimed

The first to publish is credited on this line, and the deliverable moves from Open to Achieved.

What to build

An open-hardware acoustic pulse reflectometry tester: a small loudspeaker and a microphone in an adapter held to the tube end, a USB audio interface, and Python code that sends a click, records the echoes and reports the distance to holes, wall loss and blockages. Tested only on retired tubes on a bench.

Why it is still open

The method has had an ASTM standard since 2013, but the devices found are commercial or academic, such as the PAKT-04 described by Inshakov and Kuznetsov in 2020; the low-cost open builds found are classroom echo timers that measure the speed of sound, not tube flaws. On GitHub, a 2022–2023 script simulates the echo of a pipe whose bore changes, and an April 2026 browser app compares the sound of a suspect tube with a sealed one, played from a phone and heard by a laptop; neither places a flaw along a tube or publishes results on tubes with known flaws.

iopscience.iop.orgarxiv.orggithub.comgithub.com

Built with what exists

A laptop or Raspberry Pi 5; a USB audio interface; a small compression driver and a measurement microphone; Python with NumPy and SciPy; retired copper and steel tubes 12.7 to 25.4 mm across.

  1. First step

    Drill three holes at measured distances in a 6 m copper tube and check whether echoes recorded through a laptop sound card place them correctly.

  2. Done looks like

    On 50 tubes with 150 known features (drilled holes 1 to 3 mm across, filed wall loss of at least 30%, partial blockages), the tester finds at least 90% of the holes and blockages and places them within 100 mm, every false call is published, and a second group rebuilds it from the published files and gets results within 10% of the first.