Description: The objective of this activity is to develop a quantitative NDE simulation tool for computed tomography suitable for desktop work using realistic geometry descriptions of complex anisotropic geometry. The increasing demands on NDE to address inspection reliability in the area of health monitoring and remaining life assessment demand that quantitative engineering tools be available so that cost effect engineering analysis on inspectibility limits and optimal inspection protocols be done. Most NDE techniques, as they move to a more digital format, generate terabytes of data for a single scan. X-ray methods generally have high computational needs. Until recently extracting information from massive data sets was impossible due to limited computation capabilities. By applying the emerging massively parallel graphic processing cards (GPU) to a CT simulation, SimCT, we have a means to address the quantitative modeling in an important NDE method needed to characterize materials in support of health monitoring activities. The computational techniques using GPU platforms and the data analysis methods developed in the x-ray area apply to any NDE method. This R&D effort will develop a GPU implementation of the key subroutine in SimCT and demonstrate the capability to handle NDE simulation needs using complex geometry in near real time.

Benefits: There are currently no competing CT simulation products to SimCT in the market place. The extensions and modifications added to the software under this Project will increase the speed of operation by about 100 times. This allows for new categories of problems to be addressed and investigated in reasonable times or at greater levels of detail. Besides NASA applications, there are analogous uses for SimCT in both DoD and commercial aging aircraft problems. Additionally there is the potential to enter the multi-million dollar medical market and with Homeland Security CT scaning with a swifter easier to operate program. Lastly, these improvements will allow bundling SimCT with hardware systems to improve the overall operation of CT systems. Expected sales increases of the program from these applications are estimated to be in the 150 to 200 unit category or $4 million on the high end.

Developing NDE quantitative protocols, as demands on performance of materials increases, becomes increasingly critical and difficult. Quantitative simulation tools capable of assessing the impact of the multitude of inspection parameters provide a cost effective means to determine adequate procedures in safety critical inspections. Developing simulation tools addressing realistic complexity seen in new materials and real field inspections requires an ability to adequately model complex samples; which in turn requires the utilization of the latest computation platforms, namely GPU's. The massively parallel GPU platforms enable the solution a new class of computational problem at one tenth the cost of traditional parallel approaches. The computational methods involved in modeling complex geometry from a Cad format also apply to other NDE methods modeling. Finally developing more robust and polished interface for these engineering tools gives better productivity and access to more people with reduced training needs.