High temperature thickness correction

Sound travels more slowly in hot steel. The elastic moduli soften as temperature rises while the density barely changes, and the compression velocity falls by roughly 1 % for every 55 °C (100 °F) above the temperature at which the instrument was calibrated. A gauge calibrated cold and then applied to a hot line is still converting time to distance with the cold velocity, so every reading comes out high — the wall looks thicker than it is, which is the dangerous direction for a corrosion survey.

The correction is a simple ratio. The true thickness is the displayed thickness scaled by the ratio of hot velocity to calibration velocity, t = t_disp × (1 − k ΔT). At 350 °C the factor is about 0.94, so a 12.70 mm reading is really 11.94 mm — three-quarters of a millimetre, which on a line with a 0.15 mm/yr corrosion rate is five years of remaining life.

Temperature affects more than velocity. The couplant must survive the surface, the probe itself has a duty-cycle limit and its delay line changes velocity as it heats, and the steel has physically expanded — around 0.4 % at 350 °C — so the hot wall is genuinely thicker than the same wall cold. This calculator corrects the velocity error only; where readings are being trended against a cold minimum thickness, the thermal expansion should be considered separately.

The cleanest way to avoid the whole problem is to calibrate on a block at the operating temperature, or on the component itself against a known thickness. Where that is impossible, apply the correction and record on the report that it was applied, at what metal temperature, and with what coefficient.

ΔT = T_metal − T_calibration
correction factor  k_f = 1 − k · ΔT
t_corrected = t_displayed × k_f
k = 0.00018182 /°C for steel  (1 % per 55 °C, i.e. 1 % per 100 °F)
c at temperature = c_calibration × k_f