Which method for exchanger tubes — eddy current, remote field, MFL or IRIS?

A tube bundle is thousands of thin-walled, small-bore, hard-to-reach components, and it usually has to be inspected in the window between pulling it and putting it back. Nobody has time to do everything properly, so the plan has to be right first time.

Two questions decide it, in this order.

First question: is the tube magnetic?

This is not a preference, it is a constraint, and it eliminates methods outright.

Conventional eddy current does not work on a ferromagnetic tube. The permeability of the material swamps the response and the small changes you are trying to measure disappear into it. This is not a sensitivity limitation to be worked around; it is why remote field testing exists.

Second question: what are you looking for?

Conventional eddy current (ECT) is the fastest thing on the list. A probe goes down the tube and comes back in seconds, and phase analysis of the signal gives both an estimate of depth and an indication of whether the flaw is on the inside or the outside. It is very good at discrete damage — pitting, ID and OD attack, cracking — and it is the standard screen for a non-magnetic bundle. Its weakness is gradual, uniform thinning over a long length, which produces a slow drift the technique is not well suited to quantifying.

Remote field testing (RFT) is what you use on carbon steel tubes. The signal travels out through the tube wall, along outside it, and back in through the wall again, so it passes through the thickness twice. That gives it a useful property: it responds to ID and OD wall loss roughly equally, which no other electromagnetic method here does. The price is resolution. RFT sees volume more than it sees sharpness, so it reports general wall loss well and small isolated pits poorly. It is also slower than ECT.

Tube MFL is the other option for ferromagnetic tubes: saturate the wall magnetically and detect the flux that leaks out at a defect. It is fast and it is sensitive to sharp localised pitting, which is exactly what RFT is weakest at. It is correspondingly poor at gradual general thinning, because gentle wall loss does not leak much flux. RFT and MFL are close to complementary, which is why some programmes run both.

IRIS — the internal rotary inspection system — is ultrasonic. A rotating mirror spins an ultrasonic beam around the inside of the tube while the probe is drawn along it, and what comes back is an actual thickness profile in millimetres, at every point, on any material, magnetic or not. It is the reference method, the one whose numbers you can put into a calculation.

The protocol most programmes converge on

Screen everything with the fast electromagnetic method the material allows, rank the tubes by indication severity, then IRIS the worst of them to turn indications into thicknesses.

1. 100% screening

   ECT for non-magnetic, RFT or MFL for carbon steel. Every tube, every time. The value of a bundle inspection is knowing which tubes are worst, and you cannot know that from a sample.

2. Rank and select

   Take the tubes with the deepest reported indications, plus a spread across the bundle geometry — inlet rows, the U-bends, the tubes nearest the shell nozzles.

3. IRIS the selection

   A modest percentage, typically in the range of five to ten per cent, gets a real thickness. This calibrates the screen and gives you numbers for a remaining-life calculation.

4. Plug or retube on the combination

   The screen tells you the extent; the IRIS tells you the severity. Neither alone supports the decision.

The blind spot every one of them shares

All of these methods degrade near the tubesheet. The roll transition, the tube end and the tubesheet itself distort the signal for the electromagnetic techniques and interrupt the geometry for IRIS, and the result is a zone at each end of every tube where sensitivity is reduced or the data is thrown away.

That zone is not a random place to be blind. The tube-to-tubesheet crevice is where deposits settle and stay, where under-deposit and crevice attack concentrate, and where residual stress from the rolling operation is highest. The most likely place for damage is the place the inspection is weakest.

There is no clever technique that removes this. What removes it is knowing about it: when a bundle leaks and the screen said the tubes were sound, the tube ends are the first place to look, and a programme that supplements the standard scan with a technique aimed specifically at the tube-end zone is making a deliberate choice rather than hoping.