Boiler tube thickness measurement: from survey point to remaining life

Boiler tube thickness measurement is a non-destructive test, usually ultrasonic, that establishes how much steel is left in the wall of an evaporator, superheater or economiser tube. A single reading is a snapshot. Only when you measure the same points at every outage can you see how fast the wall is thinning and when the minimum wall thickness (Tmin) comes into view. That gives you a sound basis for deciding which tubes to replace, protect or survey more closely at the next outage.

Why boiler tubes lose wall thickness

A boiler tube loses metal on the fireside, on the waterside or on both. Which mechanism dominates depends on the fuel and on the location in the boiler, and that also decides where you measure: survey points belong where the mechanism acts, not spread evenly across the whole boiler.

• Chloride corrosion. Burning waste and biomass releases alkali chlorides, such as potassium chloride, and heavy-metal chlorides. They condense in the deposits on the tubes and attack the protective oxide layer, faster as the metal gets hotter. That is why many waste- and biomass-fired boilers run at lower steam temperatures than fossil-fired boilers, at the cost of electrical efficiency.
• Fly ash erosion. Ash particles scour the wall on the side facing the gas flow. Local gas velocity matters most: corners, tube rows that have moved out of line and missing gas baffles all create high-velocity spots.
• Sootblower erosion. The steam jet from a sootblower carries ash and strikes the tubes nearest the blower. Wet steam and blowing when there is nothing to clean make it worse.
• Waterside corrosion. On the inside, water chemistry sets the pace. In heat recovery steam generators (HRSGs) at gas-fired plants, flow-accelerated corrosion is a known mechanism in carbon steel economisers and low-pressure evaporators.
• Dew point corrosion. If the tube wall is colder than the acid dew point of the flue gas, acids condense on the tube. This happens at the cold end of the economiser, where the feedwater enters.

Thickness surveys are therefore a form of condition-based maintenance: you act on the measured condition rather than on a fixed calendar. How surveys fit together with the grate, the refractory and the major overhaul in a waste-to-energy plant is covered in waste-to-energy plant maintenance.

How ultrasonic thickness measurement works

In the usual contact method, the inspector places a probe on the tube. The probe sends a short sound pulse into the wall, which reflects off the inside surface. The instrument measures the travel time of the echo and uses the sound velocity in the steel to calculate the thickness. The pulse crosses the wall twice, so: thickness = sound velocity × travel time ÷ 2. The principles are set out in ISO 16809, the standard for ultrasonic thickness determination.

Three things decide whether the result is right:

• Calibration. The instrument is set up on a reference block of the same material with a known thickness, so that sound velocity and zero point are correct.
• Couplant and surface. A thin film of couplant, such as oil, glycerine or water, goes between probe and steel, because air reflects almost all of the sound. The spot must be clean: deposits, slag and loose oxide removed.
• The right probe. For thin or corroded walls, a probe with two elements (dual element) is often used: one element transmits, the other receives. That lets you measure reliably close to the surface.

EMAT and scanning

An EMAT (electromagnetic acoustic transducer) generates the sound in the metal itself, using a coil and a magnetic field, so it needs no couplant. That makes the technique suitable for automated scanning: a scanner moves over the tubes and measures a strip rather than individual points. A scan finds local wall loss that falls between the points of a grid. You then check suspect spots with a targeted contact measurement.

Have the measurements carried out by personnel certified for ultrasonic testing under ISO 9712 (adopted in Europe as EN ISO 9712), and specify in the order which procedure, probe and report format you expect. Only then can you compare readings from different outages and different crews.

A survey plan that measures the same points every outage

A trend only emerges if you measure exactly the same points at every outage. You can build a workable survey plan in six steps:

1. 1Choose the zones. Start from the mechanisms in the table and from places where leaks or heavy wall loss were found before. Also weigh what a leak at that spot would mean; an asset criticality analysis helps you rank the zones.
2. 2Fix a grid. Number the tubes from a fixed side, choose fixed measuring levels and, at each level, a few positions around the circumference, for example A in the middle of the windward side with B and C either side of it. Give each point a unique code, such as SH2-R03-T12-L2-A: superheater 2, row 3, tube 12, level 2, position A.
3. 3Make points easy to find again. Don't rely on markings on the tube itself. Measure the location from fixed features such as a header, a weld or a manway, and record it on a drawing with photographs.
4. 4Take a baseline measurement. Measure new tubes straight after installation. Actual wall thickness can differ from nominal within the manufacturing tolerance, and every later reading is compared with this first one.
5. 5Fix the measurement procedure. Probe type, calibration, surface preparation and how you record one value per point, for example the lowest of three readings. Keep it the same at every outage.
6. 6Measure the same points every outage, and more where needed. The fixed grid gives you the trend; around spots with rapid wall loss you add points or have them scanned. Record every value with survey point, date, outage and report, so you can later trace where a number came from.

Survey reports often come from an external inspection company. On delivery, check that every point was measured and that the codes match your grid; how to follow up contractor reports is covered in managing maintenance contractors.

Remaining life assessment: calculating wear rate and remaining life

A remaining life assessment estimates how long a component can stay in service safely. For boiler tubes losing wall thickness, you derive two numbers from the series of readings: the wear rate in millimetres per year and the remaining life until Tmin.

Remaining life = (current wall thickness − Tmin) ÷ wear rate

Take survey point SH2-R03-T12-L2-A from the plan above. The tube has a nominal wall of 5.0 mm and a Tmin of 3.2 mm; both are example values. The boiler has an outage every spring and the point has been measured five times:

The trend line is the straight line that best fits all five readings (least squares). It falls by 0.30 mm per year and stands at 3.9 mm in 2026. Calculate from the trend value or the latest measured value, whichever is lower; here that is the trend value. Remaining life is then (3.9 − 3.2) ÷ 0.30 = 2.3 years: if conditions stay the same, the wall reaches Tmin during 2028. The diagram below shows the readings, the trend line and the dashed projection down to Tmin.

Two checks strengthen the number. The average loss between the first and the last reading is (5.2 − 4.0) ÷ 4 = 0.30 mm per year, which confirms the trend. Had you started from the nominal 5.0 mm instead of the baseline, the wear rate would have come out at 0.25 mm per year: too low. Also use the actual dates between surveys, or operating hours if the boiler stood idle much longer in one year than in another.

Why you should not rely on the last two readings alone

If you use only the last two readings, you get (4.1 − 4.0) ÷ 1 = 0.1 mm per year and a remaining life of (4.0 − 3.2) ÷ 0.1 = 8 years. The difference comes from measurement scatter: the readings sit up to 0.1 mm off the trend line, a third of what the wall loses in a year. With two readings, every measurement error passes straight into the wear rate; a line through all outages since the last replacement largely averages that scatter out.

A projection, not a prediction

The result is a projection: the trend of your own readings extended, on the assumption that conditions stay the same. If the fuel, the load, the combustion settings or the sootblowing change, the wear rate changes with them. Every new outage updates the projection.

This calculation only looks at metal loss. In parts that work hot and under pressure, such as superheaters and headers, creep and fatigue also play a part. A thickness survey won't reveal that damage; EN 12952-4 sets out the calculation method for water-tube boilers in service. Have a qualified engineer do that part of the remaining life assessment.

Tmin: the lower limit comes from the design calculation

Tmin is the minimum wall thickness a tube needs to withstand the design pressure at the design temperature. It follows from the strength calculation of the boiler. For water-tube boilers, EN 12952-3 gives the calculation rules for pressure parts; if your boiler was designed to a different code, that code applies. The result depends on the pressure, the outside diameter, the material and the allowable stress at the metal temperature. In superheaters that temperature weighs heavily: the hotter the metal, the lower the allowable stress and the higher Tmin.

Corrosion and manufacturing allowances are not part of Tmin; together with rounding up to an available wall thickness, they make up the difference between Tmin and the nominal wall. Have Tmin set per tube type by a competent party, such as the boiler manufacturer, agree the value with your inspection body and record which value you use and why. A fixed percentage of the nominal wall is not a Tmin. You can, however, choose your own replacement limit above Tmin as a margin for measurement uncertainty and local thinning.

From measurement to decision for the next outage

Convert every reading into a class, so you can see at a glance where action is needed. You set the limits yourself; this is an example:

At 2.3 years, the point from the worked example falls in orange. The trend gives 3.6 mm at the 2027 outage and 3.3 mm at the 2028 outage: 0.1 mm above Tmin, which is within the scatter of the readings. So prepare the intervention before the 2027 outage, re-measure the point during that outage and carry out the intervention in 2028 at the latest; if that margin feels too tight, act in 2027.

Possible interventions:

• Replace tube sections, panels or whole bundles. Arrange the material well ahead of the outage. The trade-off between repairing and renewing is covered in repair or replace.
• Protect with weld overlay or a coating, such as cladding with a corrosion-resistant alloy or a thermally sprayed coating. Keep measuring afterwards: in research into superheater corrosion, the coating that performed best in the laboratory was not the best in a waste-fired boiler.
• Tube shields on tubes struck by fly ash or a sootblower.
• Tackle the cause: adjust sootblowers, repair gas baffles, improve water chemistry.
• Measure more closely: more points, a scan or an extra survey during a short intermediate stop.

Anything orange or red goes into the scope of the next outage, with its supporting data; how to build and freeze that scope is covered in shutdown planning.

A steam boiler is also pressure equipment. In the Netherlands, designated pressure equipment is inspected before it is put into service and re-inspected periodically, and the inspection body must be informed in advance of any intended repair or modification, such as replacing tube sections (Pressure Equipment Decree 2016 (Warenwetbesluit drukapparatuur 2016), article 26, in Dutch). Other countries have their own rules; there is more on this in statutory inspections.

Common pitfalls in boiler tube thickness data

Many errors in a projection sit not in the arithmetic but in the data underneath.

• Points that cannot be found again. If you measure just beside last time's point at every outage, you are measuring scatter, not wear.
• A wall that gets thicker. A difference within the scatter of your series, 0.1 mm in the worked example, is noise. If a point reads clearly thicker than at the previous outage, it was measured over deposits, beside the old point or with a different calibration, or the tube was replaced without a new series being started. Find out which of the four it is. Correct the value with a reason and keep the original, rather than quietly overwriting it.
• Calculating across a replacement. A new tube starts a new series, with its own baseline. Putting old and new readings in one line gives a wear rate that means nothing.
• Spreadsheets you cannot trace. A new file every outage with a different layout, retyped by hand: after a few years nobody knows which number came from which report.
• Measuring only the grid. Local attack around a sootblower falls between the points. Scan where the mechanism acts locally.

How to keep such a series traceable per point, with imports and corrections that carry a reason, is shown in FurnIQ.