Noncontact measurement of internal temperature distribution in a solid material using ultrasonic computed tomography

Numerical simulations and experiments have been carried out for a noncontact measurement of the internal temperature distribution in a solid material using ultrasonic computed tomography (CT). The method is based on the fact that the sound propagation velocity in a material depends on its temperature as well as its density and structure. From the numerical simulations, the convolution method is found to be an effective algorithm for the reconstruction of the sound velocity distribution. To obtain an accurate temperature distribution, it is found to be necessary to measure the sound propagation time with a resolution of I ns. In the experiments, the temperature distributions are measured in an agar-gel cylinder of 40 mm in diameter, along the center axis of which a platinum wire with 0.1 mm in diameter is located. By comparing the experimental results with the theoretical ones, the temperature distribution inside the agar-gel can be reconstructed with an error of 0.1 K, except for the region close to the platinum heater wire where temperature gradient is high. Further, the effects of an obstacle to the sound propagation, such as an acrylic resin cylinder inside the agar-gel, are investigated. Although the obstacles causes a part of projection to be missed, by using a linear-interpolation method to compensate for the incomplete projection, the temperature distribution can be reconstructed well but with a little larger error of 0.2 K, except for the regions close to the platinum heater wire and obstacle. (C) 2001 Elsevier Science Inc. All rights reserved.