Corrosion of oil refinery equipment operating at high temperature is a growing challenge linked to an increase in sulfur concentration and acidity within the world supply of crude oils. Improved high temperature monitoring tools are needed to continuously measure the local minimum wall thickness of steel pipe subject to non-uniform internal corrosion. Thickness measurements must be sufficiently accurate to calculate the maximum allowable stress permissible as to avoid failure. Thickness measurements must also be sufficiently reliable and precise to correlate the rate of wall-loss with process conditions as to improve future corrosion rate predictions. Permanently installed structural health monitoring (SHM) technologies have the potential to meet these needs and an ultrasonic sensor technology based on the sol-gel ceramic fabrication process will be presented. The sol-gel thin-film sensor fabrication process is described: aerosol printing of high-temperature piezoceramic material, heat curing, polarization, electrode deposition, and wiring. The sensors are characterized with the photoelastic visualization method by first, recording the propagating elastic waves as a sequence of images, then, processing the image frames to reconstruct a beam profile through identification of the maximum optical intensity for each pixel via normalizing, filtering, and smoothing. A sol-gel transducer is shown to be similar to a traditional manual contact transducer. The systematic and environmental factors that impact ultrasonic thickness measurement accuracy, precision, and reliability are discussed. An experiment is conducted using a flat-bottom-hole calibration pipe at ambient temperature with sol-gel transducer arrays in pulse-echo and pitch-catch configuration for various time-of-flight thickness calculation methods. A weighted censored relative maximum likelihood statistical method incorporating the propagation of asymmetric uncertainty is used to report thickness measurement results with confidence limits analogous to the a90/95 terminology used in Probability-of-Detection (POD) assessments. Future work is discussed to apply the statistical analysis technique to complex back-wall surfaces at high temperature representative of naphthenic acid corrosion in oil refineries.