Development of Oil Behavior Visualization Technology using X-Ray Computed Tomography

Recently, it has become possible to dynamically observe the internal oil behavior using a medical X-ray computed tomography (CT) system, which can capture internal images without disassembling the object being evaluated. The problem in applying the CT method to lubrication analysis is the artifact, like noise, which occurs during dynamic observation. In this paper, we developed a method to reduce this artifact with machine learning by generating artifacts through simulation and using them as supervised data.