Time-of-flight estimation in acoustic pyrometry: sensitivity to pulse characteristics

Abstract

Acoustic pyrometry is a non-intrusive measurement technique that may have several applications in turbomachinery. This methodology estimates the gas temperature by measuring the time of flight of an acoustic wave moving through a medium. It can be accomplished by placing a sound source (emitter) and a set of microphones (receivers) on opposite sides of a section. The emitter generates a sound pulse, and the receivers detect it. Since the emitter-receiver distances are known and fixed, the average temperatures of the paths traversed by the acoustic pulse can be computed by estimating the time-of-flight through deconvolution techniques. However, despite the straightforward principle, an acoustic wave suffers a variation of amplitude when propagating within a medium because of energy losses and ambient noise. Hence, time-of-flight estimation becomes a critical task, especially when considering high-frequency waves or short distances between sensors. It is then fundamental to select proper acoustic waves to maximise the cross-correlation between the signals of the emitter-receiver couples, thus improving the accuracy of the time-of-flight measurements and, consequently, the estimation of the spatial temperature distribution within a specific area. This study is a preliminary investigation, based on a modelling approach, to estimate the impact of different acoustic waves on the accuracy of the time-of-flight measurement. The results of this analysis will be useful to design and setup an acoustic pyrometry application.