Of all the through-insulation techniques, pulsed eddy current is the one that has changed CUI inspection most. A probe pressed against the jacketing gives a remaining-wall reading in seconds, on a hot line, with no scaffolding for insulation removal. This note explains what the instrument is measuring, where it is reliable, and the limits that decide when a reading needs a follow-up.
The principle
A PEC probe contains a coil. The instrument switches a current on, which builds a magnetic field that passes through the jacketing and insulation and into the steel. When the current is switched off, the collapsing field induces eddy currents in the pipe wall, and those currents decay. In a ferromagnetic material the decay proceeds in a way that depends on how much steel is there: a thick wall sustains the currents longer, a thin wall lets them die away faster. The instrument records the decay curve and converts a characteristic time on it into an average wall thickness under the probe.
The reading is reported relative to a reference: either a nominal thickness entered by the operator or a calibration on a known sound spot of the same pipe. That is why PEC results are usually quoted as a percentage of nominal wall rather than as millimetres, and why the reference matters as much as the probe.
What it sees through
- Insulation of any non-conductive type: mineral wool, calcium silicate, cellular glass, aerogel, foam. Thickness is limited by probe size; larger probes reach further at the cost of a larger footprint.
- Aluminum and stainless jacketing, which are non-magnetic and only mildly attenuate the signal.
- Galvanized steel jacketing, which is magnetic and does interfere. Modern systems can often compensate, but it costs sensitivity and needs care in setup.
- Coatings, corrosion product and fireproofing, which is why the method also serves for corrosion under fireproofing and for scale-covered surfaces.
- Heat. Nothing touches the metal, so lines in service at high temperatures can be screened, within the probe's own rating.
Footprint averaging
The single most important thing to know about a PEC reading is that it is an average. The magnetic field spreads as it travels through the insulation, so the wall the instrument "sees" is a patch whose diameter grows with the distance between probe and steel. A deep, small pit inside a wide footprint of sound wall reads as a modest average loss; a broad area of general thinning reads accurately. Vendors publish the footprint for each probe and lift-off, and procedures set the reporting threshold accordingly, often flagging anything below a set percentage of nominal for follow-up.
Consequences for CUI work: PEC is a screening tool that finds areas worth opening, and a low reading is a reason to strip and measure, not a measurement in itself. Where pitting is the expected form of damage, at supports for instance, the procedure should say how a pit-dominated loss is treated.
Where it goes blind
- Stainless steel
- The method relies on ferromagnetism. Austenitic stainless piping, the grades that suffer chloride cracking under insulation, cannot be screened with conventional PEC. Cracking needs a different approach, usually insulation removal and surface methods.
- Geometry
- Nozzles, flanges, welds, elbows and the edge of a vessel head distort the field. Readings within roughly a footprint of such a feature need interpretation or a different technique.
- Supports
- A steel support shoe or saddle sits inside the field. Specialized procedures and probes exist for corrosion under pipe supports, but the reading will reflect both the pipe and the support unless the setup accounts for it.
- Nearby steel
- Adjacent pipes on a tight rack, structural steel and cladding stiffeners can influence the decay. Operators check for this with a lift-off sweep.
- Very thick or very thin walls
- Each probe has a working thickness range. Outside it, the decay is either too slow to resolve or too fast to time reliably.
- Resolution
- Beyond the footprint issue, the method cannot distinguish internal from external loss. On a line with an internal corrosion mechanism, a low reading needs UT to say which side it is on.
Array probes and coverage
Single-point PEC is quick per reading but slow per metre. Array probes carry several coils and are drawn along the pipe, producing a thickness map at walking pace. That changes the economics of coverage: instead of a few readings per circuit at chosen points, a crew can scan whole runs and let the map suggest where to open windows. The averaging and geometry limits still apply, and calibration on a known spot remains the first step.
What a good procedure covers
- Reference and calibration: where the nominal comes from, and how a sound spot is confirmed.
- Probe selection for the insulation thickness and the pipe diameter, with the footprint stated.
- Lift-off check at each location and the treatment of galvanized jacketing.
- Reporting thresholds and the follow-up action for each band, for example a strip-and-UT window below a stated percentage.
- Exclusion zones near features and supports, and the alternative method used there.
- Recording: location referenced to the circuit and a fixed datum, so the next campaign can return to the same point.
Personnel are typically trained and certified on the specific equipment and work to the owner's written procedure; there is no single code that replaces that.
Where the probe goes first
A crew can read a few hundred points in a campaign, whether by single probe or array. The readings are quick; the order is the plan. Circuits in the warm, wet temperature band with old coatings, absorbent insulation, many supports and a marine or cooling-tower environment deserve the first day, and a ranked list built from those attributes, checked against past findings, is how to decide that before the truck arrives. That ranking is the subject of risk-based inspection for CUI.
The reading simulation is a clay-model rack of 48 insulated circuits with a reading point at every support, low point and dead leg. Run the sweep and a probe works through the circuits in workbench order, turning each point from unread to a wall-thickness result, so you can watch where the ranked order finds the damage and how early. The anatomy section shows the decay curve for a nominal and a thinned wall.
Open the CUI reading simulationReferences
- API RP 583, Corrosion Under Insulation and Fireproofing: applicability of NDE methods, including pulsed eddy current, to insulated equipment.
- Equipment manufacturers' technique sheets for footprint, lift-off range and jacketing compatibility of specific probes.
- CUI inspection methods for how PEC sits alongside radiography, guided wave and insulation removal.