Tomographic 2-D gamma scanning for industrial process troubleshooting

Abstract

Gamma scanning is a nuclear inspection technique widely used to troubleshoot industrial equipments in refineries and petrochemicals plants such as distillation columns and reactors. In this technique, a sealed radiation source and detector move along the equipment, and the intensity readouts generate the density profile of the equipment. Although many improvements have been introduced in recent years, the result of gamma scan still consists of a simple 1-D density plot. In this work, we present the tomographic gamma scanning that, using image reconstruction techniques, shows the result as a 2-D image of density distribution. Clearly, an image reveals more features of the equipment than a 1-D graph and many problems that could not be troubleshooted using the conventional technique can now be solved with this imaging technique. We use ART (Algebraic Reconstruction Technique) intercalated with total variation minimization filter. The use of total variation minimization leads to compressive sensing tomography, allowing to obtain good quality reconstruction from few irradiation data. We simulated the reconstruction of different density distributions. We applied the new technique to data obtained by irradiating with gamma rays phantoms that emulate industrial equipments. Finally, we present the result obtained by applying the innovative technique to real operating distillation column. It seems that the new technique has identified a problem in this equipment that is very difficult to detect using conventional gamma scan.