Ultrasonic monitoring of pipe wall interior surface temperature

Abstract

Accurate temperature measurement is a crucial aspect of structural health monitoring and prognosis. Conventional temperature measurement devices are either incapable of measuring subsurface temperatures in solids or need to be invasively installed. This study investigates the use of an ultrasonic technique for non-invasive measurement of subsurface temperatures in steel components; the temperature of a point on an inaccessible surface is inferred using a time-of-flight measurement from a transducer placed on an opposing accessible surface. Two different inversion approaches are presented, one named the assumed distribution method and the other named the inverse thermal modelling method. The robustness and accuracy of the two ultrasonic temperature inversion methods are quantitatively assessed via simulations and controlled experiments. It was found that both the assumed distribution and inverse thermal modelling methods demonstrate short thermal response times and are able to track the temperature evolution of inaccessible surfaces. A series of experimental studies show that in the presence of a 15°C difference between the accessible and inaccessible surfaces, the inaccessible surface temperature is typically measured to within better than 2°C with respect to a resistance temperature detector reference measurement. Additionally, the article compares the measurement performance achieved using a deployable electromagnetic acoustic transducer and a permanently installed piezo-electric PZT transducer. The time-of-flight measurements taken using the electromagnetic acoustic transducer system had higher random noise than the PZT system (standard deviations of 0.42 and 0.016 ns, respectively), subsequently leading to higher random noise in the temperature estimates.