In recent years, x-ray Computed Tomography (CT) imaging has become much more widely used in industrial applications. In many such applications, however, only incomplete data sets [1] are available, and image quality is degraded by the absence of complete data. In this paper, a model-based CT reconstruction technique for enhancing incomplete data CT image quality is presented. A two-dimensional registration method which will ensure proper utilization of a priori information from CAD model is introduced. Images are shown to demonstrate manipulator position variability as well as blade to blade variability. Variability is quantified by the translation and rotation factors used in geometric transformation. In addition, incomplete data CT image simulation results are presented which show the effects of manufacturing variability on flaw detectability in these images.