SBIR-STTR Award

This project aims to implement novel techniques for feedforward and feedback control that will allow better control, validation, and documentation of Selective Laser Melting (SLM) additive manufacturing (AM). Three complimentary key innovations will be realized in this project (two in Phase I and a third in Phase II) by combining and improving two current technologies. The first is the integration of Fringe Pattern Projection Profilometry (FPPP) into the SLM process. FPPP is the first profilometry technique that can capture high resolution dimensional measurements of the entire SLM build platform, in situ and nearly instantaneously. This facilitates direct dimensional measurement and validation of every single layer, and post-process 3D models (built from the measurements) for the digital twin. By capturing all dimensional information (including residual stress induced distortion) the FPPP sensor will provide a unique set of data for calibration of AM modelling software, which is the second key innovation. The FPPP data will identify defects in layer morphologies that can be used to train unique integrated computational adaptive additive manufacturing (iCAAM) feedforward modeling tools (distortion is predicted and compensated for with the build strategy before the build starts). In most simulators, the layer thickness is assumed to be constant and perfect, but it is not. FPPP data will quantify the true variability present in layer thickness as the part is built. Access to this information will allow more accurate calibration of the model so final part distortion can be virtually eliminated. In Phase II the model will also be inverted and turned into a fast-feedback lookup table for further tuning the build process to compensate for suboptimal layer morphologies that may arise, which is the third key innovation. The result will be a combination of hardware and software tools that eliminate distortion and capture critical information for the digital twin. Potential NASA Applications The technology is applicable to the Space Launch System (SLS), which is currently building components for the RS-25 rocket engine. Feedforward control can also avoid waste of time and cost associated with failed builds for several other rocket nozzles currently built with NASA SLM systems. Other applications include CubeSats and small deep space engine components that need to be distortion free. Potential Non-NASA Applications DoD need improved SLM for flight critical aerospace components. The medical device industry needs better validation for SLM processes in order to pass FDA scrutiny before parts can be used commercially. The feedforward technology could be readily implemented by any commercial SLM supplier that builds critical components which require validation.

In Phase I the research team demonstrated a superior in situ profilometry sensor, based on fringe pattern projection, which quickly measures the whole build plate. In this data, significant process phenomena are accurately measured and easily identified, such as spreading defects, rogue particles that have been sintered to the part’s surface, distortion, surface roughness variation, and virtually any geometric feature. Of particular importance is the measurement of powder layer condensation and uniformity. This data serves as input to a model that generates feedforward information to adjust process parameters, resulting in better prediction and control of key material properties such as residual stress and density. In Phase II the team will further improve the sensor and test the feedforward model. After fine-tuning the modelling capability for stress and distortion, mechanical testing will be conducted to validate model performance and determine the effect of defects (measured with the profilometry) on mechanical performance. The result will be real-time determination of part quality by a modelling tool that integrates profilometry-detected defects into the performance predictions. This novel data will then be used to feed and validate a fast-feedback look-up table (generated by inverting the feedforward model), for layer-to-layer laser parameter adjustment during builds. Next, a new design of the profilometry sensor will be completed to make it very compact (a few inches) so it can easily be added to OEM AM machines. Then the research team will implement a new sensing technique (with the same hardware) to record video-rate, measurements, at nanometer precision, of thermal expansion and shrinking during the melting process, thereby facilitating novel and powerful analysis of residual stress and/or delamination formation. Finally, the research team will demonstrate the whole sensor/modelling package on a NASA geometry of interest. Potential NASA Applications (Limit 1500 characters, approximately 150 words) Applications include any system that wishes to use AM parts in critical areas, including: Rocket Engines: The SLS program heavily utilizes AM. These components can be very large and require long build time, experiencing failed builds is painful. Deep Space Exploration: Research on Stirling engines is heavily interested in AM and engine components must be reliable Material development: High quality in situ data, like this profilometry, may be useful for investigating process phenomena during the development stages of new AM materials. Potential Non-NASA Applications (Limit 1500 characters, approximately 150 words) Applications include any system that wishes to use AM parts in critical areas, including: Department of Defense supply chain: DoD suppliers aim to build an ever-growing list of critical parts that must have adequate process validation and documentation for the digital twin. Medical Device: AM is experiencing strong pull in medical devices. Anything that goes in the human body must be qualified.