SBIR-STTR Award

We will develop software called Automated Reverse Engineering of CT Scans, or ARECS. This code provides the ability to automatically find and fit geometric surfaces like planes, spheres, etc. within an unstructured point cloud. ARECS outputs a 3D solid model with surfaces parametrically defined (parasolid model). This replaces 80% of the time and effort expended today by an engineer laboriously creating these surfaces (reverse engineering). Munition components contain numerous internal surfaces and are well suited for inspection by industrial computerized tomography (CT). This proven, low risk NDT method precisely measures internal and external surfaces of many part types. Besides geometric dimensioning and tolerancing (GD&T), CT also displays minute internal cracks and corrosion to a very fine level of detail (microns). ARECS leverages commercial software (an SDK) that already performs automated surface detection and fitting using machine learning. This allows writing a plugin for Avizo, leading visualization and reverse engineering software from ThermoFisher. The plugin provides automated reverse engineering capability to Avizo with minimal risk, cost and time because no code changes are needed in Avizo. ThermoFisher is a subcontractor in our program, and our commercialization path is to license ARECS to them to expand Avizo's capabilities.

Quality Improvement of Critical AM Parts program significantly enhances quality of AM parts so a broader range of parts can be considered. Inspection improvements occur in two ways– in-process inspection (for every layer of the part during build) and final CT inspection after print completion. Flightware has implemented AM in-process inspection of the melt layer under prior programs with DOD and NASA, demonstrating high reliability. Now we will extend that work to detect flaws in the spread powder layer, which often leads to melt defects in the part. Furthermore, we will also demonstrate that powder flaws like stray spatter particles can be repaired at the layer level, like previous successful demonstration of in-process repair of melt flaws. For final inspection, we will generate higher quality CT scans for thick, steel parts such as used in Armament. We extend the successful work in Phase I to use Fan-Beam X-rays enhanced using Model Based Iterative Reconstruction (MBIR) to produce much better CT quality. This provides improved ability to detect very small, but important, internal flaws like porosity. This work is performed by a team of several firms that have collaborated successfully for several years to develop improvements in AM quality