X-ray Computed Tomography and Stereo-Radiographic Inspection Results of the Office of Emergency Response (NA-42) Test Object

This report has documented the worked performed in the x-ray computed tomographic and stereo-radiographic inspection of the NA-42 test object. We have described the method SRNL used to obtain high resolution (80 micron) images of the test object using PSL plates. The PSL plates are an excellent alternative to x-ray film and they eliminate the need for the wet chemistry processing and the disposal of the chemical wastes. The PSL plates were used to provide an overall panoramic view of the large test object. These images were useful in planning other inspection techniques. In addition, a customized digital radiography system with an 85-inch wide field-of-view was assembled to support the data collection for computed tomography. The trade-offs between resolution and data collection and CT reconstruction time were explained in detail. The CT projections and reconstructed slices of the test object were included in the report as static images and ''movies'' were also provided on the attached CD-ROM. The combination of the projections and the CT slices provide a thorough understanding of the internal structure of the device. The full projection CT results were also used as a ''bench mark'' for other techniques investigated during this work, such as the limited view CT and stereo-radiographic work. The limited view CT results were obtained by parsing the full data set into subsets with larger angular intervals and thus fewer projections. These subsets were then processed with the CT reconstruction software. The results of reconstructions from 720 down to 10 projections were compared. Based on these results, we concluded that 20 to 30 projections were adequate. These results were then used to predict the required data collection time for higher resolution systems. It was concluded that from a data collection time basis, limited view CT could provide the desired resolution (1 mm) within a reasonable period of time. However, there were other considerations related to actual field conditions that might preclude the use of CT. Consequently, SRNL also investigated other techniques. If the objective of the inspection is to provide the coordinates of key components of an IND, then CT is not needed. SRNL has shown that stereo-radiography using relative object-imaging system rotation or source translation are both effective in providing accurate targeting information. A detailed explanation, including examples, of these techniques was provided in this report. SRNL also believes that the accuracy of the techniques can be dramatically improved in future work, because several key parameters were measured using a tape measure. For example, a 3-D fiducial array could be used to determine key parameters without a need for direct measurement. A higher degree of accuracy in the determination of those parameters will enhance the accuracy of the overall technique. In addition, since the actual internal geometry of the test object was unknown, the CT results were used to validate the other methods. However, the calculated error in the target coordinates were on the order of the resolution of the CT results (0.109 inches). A better test of the techniques would be to construct a test object where the target points are measured with a higher degree of accuracy than the technique to be proven. A brief discussion on the usefulness of stereo-graphic vision equipment was also included. The main benefit of this method is to provide field personnel with qualitative ''3-D'' information without the need to collect a CT dataset or perform a tomographic reconstruction. Based on these results, SRNL recommends that the best approach at determining target locations is to utilize the stereo-radiography approach using source translations and fiducial markers. Lastly, the PSL plates or ASi flat panel imagers offer the most practical detector choice for a field application