The EMIT lab (Engineering MaterIals for Transformative technologies) at CMU run by Sneha Narra focuses on advancing metal additive manufacturing knowledge to manufacture light-weight organic designs and utilize novel, advanced materials. At a scientific level, our group studies the fundamentals of additive manufacturing processes, investigates the resulting material microstructure and properties, and develops process design paradigms. Our group’s mission is to lower the carbon footprint, primarily in aerospace, automobile, and energy industries. This will be achieved either through reduced material and/or operational usage or through enabling transformative technologies, such as nuclear fusion and high-efficiency turbine systems.
Sneha Narra’s academic training and experience has led to the establishment of the EMIT lab, working at the intersection of mechanical and materials science and engineering and leveraging data-driven methods to expand advanced manufacturing capabilities. As an instructor, Narra’s goal is to help her students learn effectively in a comfortable environment and spark interest in them to explore outside the classroom. Narra is passionate about mentoring and participates in outreach activities, educates students about professional development opportunities, and provides opportunities to conduct research in interdisciplinary topics.
Labs and facilities
NextManufacturing Center houses various fusion-based metal additive manufacturing equipment that our lab uses in our research work.
We utilize electron microscopy and x-ray facilities at the Materials Characterization Facility (MCF), to conduct material characterization work.
We have robotic wire arc additive manufacturing (WAAM) equipment in our lab at Mill 19. In the following video, Lincoln Electric’s WAAM machine deposits weld beads in a layer-by-layer manner to build 3D structures. WAAM allows engineers to produce large 3D structures relatively fast compared to alternate additive manufacturing processes and traditional manufacturing processes. Builds on the scale in the video, around 1ft x 1ft x 0.5 ft footprint, take around 15-20 hours when using a qualified process. In the video below, three build layers are shown at an increased speed.
Critical defect size prediction
A major factor determining the fatigue life of fracture-critical parts is the effect of process-induced defects and the critical pore/defect size. Existing part qualification procedures rely heavily on time-consuming experimentation and testing procedures to determine the fatigue life. To address this gap, we are developing approaches that integrate processing and statistical modeling to systematically define data requirements and ultimately predict critical defect size in complex parts with heterogenous defect distributions.
Spatial and temporal effects on microstructure
We are collaborating with the computational modeling experts at University of Pittsburgh to develop computational process design approach that significantly reduces the need for experimentation. Toward this goal, we are developing an experimentally validated process simulation tool to predict defect regime and phase transformations in the laser powder bed fusion process. Our group at CMU is designing the experiments and conducting material characterization to understand the effects of heating/cooling and thermal fluctuations at different time scales.
AI-enabled smart drying
Heating processes account for 61% of annual manufacturing end-use energy consumption, of which the drying industry is the second-largest consumer. This project is an interdisciplinary collaborative effort integrating novel drying approaches, advanced sensing technologies, and machine learning to minimize the energy consumption in the drying process while maintaining product quality. Our group is developing physics-informed neural networks to replace time-consuming numerical modeling, handle noisy data, and integrate process conditions with sparse sensor measurements.
Microstructure evolution in wire arc AM
The high energy input in wire arc additive manufacturing leads to in-situ heating of the previously deposited layers, which in turn, can have a notable effect on the as-fabricated microstructure of age hardenable materials. Hence, a comprehensive understanding of in-situ thermal cycles and heat accumulation on the microstructure evolution and resulting properties can be a key to optimal process design. We utilize a suite of characterization and testing techniques (SEM, EBSD, XRD, hardness, and tensile testing) and in-situ monitoring data (thermocouple and IR camera), and conduction-based heat transfer models (numerical) to investigate the feasibility to tailor microstructure in age hardenable steels subjected to non-equilibrium solidification conditions and thermal gyrations.
Microstructural analysis of WC-Co binder jet samples
Material selection limits the capability of additive manufacturing. Brittle materials, specifically, are prone to microcracking due to residual stresses in the as built part reducing the fatigue life. Binder Jet technology is able to reduce residual stress due to the densification steps occurring after the green part is formed.
In this project, Binder Jet WC-Co samples provided from Kennametal are analyzed for quality and material properties. Computer vision is being used to find the relevant attributes of critical microstructures in order to assess the correlation between these attributes and the material properties.
This is funded by Manufacturing PA grant.
In-situ monitoring, dynamic modeling, and feedback control of wire-arc additive manufacturing
Wire Arc Additive Manufacturing (WAAM) allows the production of large scale parts that can economically replace forgings and castings with no need for fixed tooling. However, the process currently has weaknesses in terms of stability, defects, and final microstructure. In order to address these issues, we are working to develop techniques to monitor WAAM in real-time to predict and control the process with live feedback.
Various elements of this project are being done in collaboration with Dr. Jack Beuth, Dr. Chris Pistorius, Dr. Lu Li, and Dr. Howie Choset. Funding is provided by the Manufacturing Futures Institute.
Two-color thermal imaging method for wire-arc additive manufacturing
Wire Arc Additive Manufacturing (WAAM) is a directed energy deposition method, typically used to create components for large scale manufacturing applicable to various industries such as energy, transportation, and aerospace. Melt pool morphology, defect structures, and resulting microstructure are highly dependent on the weld pool temperature, and using the two-color method on a color camera simplifies the detection of temperature transients by neglecting the heat radiation emissivity for in-situ temperature monitoring.
This work is conducted in collaboration with Dr. Jonathan Malen in the Department of Mechanical Engineering, funded by the Manufacturing Futures Institute and The National GEM Fellowship Consortium.
AM for high-temperature alloys
Oxide dispersion strengthened (ODS) alloys contain a high number density of nano-scale oxide (e.g., yttria) particles, offering exceptional high-temperature creep strength and radiation resistance. These properties make ODS steels an attractive candidate for structural materials in next-generation nuclear fusion reactors. In this project, molecular dynamics simulations are integrated with AM experiments to investigate the mechanisms of oxygen dissolution and oxide formation, with the goal of tuning alloy composition and AM processing conditions to control oxide characteristics. This work is conducted in collaboration with Professor Alan McGaughey, an expert in molecular dynamics.
Process mapping of wire-arc additive manufacturing for feedback control and melt pool recognition
Wire arc additive manufacturing (WAAM) is desirable for printing large-scale metal parts for aerospace and automotive industries but requires significant manual intervention by the user and lacks concrete data mapping welding parameters to melt pool dimensions. In this project, a method for automating data collection for process maps is developed. Printing parameters are mapped to melt pool lengths and widths, and thresholds are determined for computer vision techniques to identify the melt pool through a filtered image camera. Through this work, predictions of defect regions and their associated melt pool dimensions can be rapidly found for new WAAM materials, with particular applications in developing closed-loop feedback control systems for the WAAM process.
This work is conducted in collaboration with Dr. Jack Beuth, Dr. Chris Pistorius, and is funded by the Manufacturing Futures Institute.
Rapid feature identification using variational autoencoders
Fatigue failure behavior in laser powder bed fusion (L-PBF) parts is a complex problem that is affected by internal defects such as gas and lack of fusion porosity. A data driven analysis is being undertaken to predict fatigue failure using the hundreds of cross-sectional images of fatigue specimens generated for the NASA ULI project. With the goal of removing presumptions about which features in the images are valuable, an variational autoencoder (VAE) is employed to learn, compress, and recreate the images – thereby learning all of the data in each image. By tailoring the size of the compressed space, occasional distributions of fatigue failure classifications can be observed in the compressed space vectors. The compressed-state vectors that produce distributions can then be fed into the VAE’s decoder to visualize the features in the images that caused the fatigue failure distribution. With this intuitive and unsupervised machine learning method, the human error associated with choosing which features to record in imaging data can be reduced, and previously untabulated features can be identified.
This work has been made possible by the NASA ULI project.
Effect of Heat Buildup on Laves Phase volume fractions in Inconel 718 produced by laser powder bed fusion
Laser powder bed fusion (L-PBF) is a layer-by-layer additive manufacturing process where successive layers of powder are fused together using a laser. Due to the layered fusion process, L-PBF parts have a tendency to build up heat during fabrication. While heat buildup is known to shift process parameters, there is a gap in knowledge about the effect of heat buildup on the as-fabricated microstructure and the resulting impact on heat treatments. Specifically, the distribution of laves phase in Inconel 718, a brittle precipitate that is formed during L-PBF fabrication, is of primary concern due to its detrimental impact to mechanical properties. To answer this question, L-PBF Inconel 718 test samples were intentionally manufactured with significant heat buildup, and are being imaged to measure the volume fraction laves phase as a function of temperature. Following analysis of as-fabricated conditions, the samples will be heat treated to study the impact of laves phase distributions on the recrystallization of strengthening phases in Inconel 718 parts.
This work is conducted in collaboration with Dr. Albert To, Shawn Hinnebusch, and Seth Strayer of the University of Pittsburgh. Funding is provided by the NASA Transformative Technology Transfer program.
Using irregular powders in laser directed energy deposition
Spherical powders are typically used in metal additive manufacturing (AM), however, standard atomization production methods are costly and low-yield, creating metal waste. Powders with irregular morphologies have lower costs, however, they require careful selection of deposition conditions to avoid defects. Our prior work has shown effective process optimization for irregular powders on powder bed fusion AM methods, while there is limited literature on the use of these powders in commercial Laser Directed Energy Deposition. To bridge this gap, we are characterizing the performance of irregular powder in the Laser DED process in order to optimize powder feed and deposition conditions.
This work is funded by the BRIDGE Program.
Gala Cassiel Solis
William Frieden Templeton
If you are interested in joining the EMIT Lab at CMU, please read the appropriate section below.
Undergraduates: Projects are advertised through the MechE newsletter as they are available. Please refer to this information for opportunities to join the lab or email Dr. Narra with the tag [Undergrad] in the subject if you are interested in whether new projects may be becoming available.
Graduate students: All graduate students (MS or Ph.D.) will only be considered through the official CMU application system. Please apply online. If you are an admitted MS student and would like to learn about available projects please refer to the MS Canvas page or email Dr. Narra with the tag [MS] in the subject.
Postdocs: No postdoc positions are available at this time.
We look forward to hearing from you!
President Biden Visit
On January 28th, our group participated in the Mill-19 facility tour and demonstrated the wire arc additive manufacturing process in our lab in collaboration with Lincoln Electric Additive Solutions. We discussed the benefits of metal additive manufacturing with the President of the United States. We are fortunate to be part of the additive manufacturing community and honored to represent our field during the President of the United States’ visit to CMU.
(Left - Right) William Templeton, Sneha Narra, President Biden. See President Biden’s tweet about the visit.
Graduate students currently advised at WPI
- Hanshen Yu, Ph.D. student, June 2020-present, co-advised with Professor Jamal Yagoobi
Biden calls for investment in American innovation
President Biden touted the importance of advanced manufacturing innovation, robotics, 3D printing, and artificial intelligence during his recent visit to Mill 19.
Metal 3D printing across scales
New faculty member Sneha Prabha Narra is interested in utilizing recent advances in in-situ monitoring, process modeling, and fundamentals of welding to advance the use of wire arc additive manufacturing for new applications and materials.
- Yao Xu, Ph.D. student, Aug 2018 – May 2022, co-advised with Professor Brajendra Mishra, currently working for Mattson Technology
- Mahya Shahabi, Aug 2019 – May 2021, currently a Ph.D. student at WPI
- Prajwal Bharadwaj, January 2021 – May 2021, currently a Ph.D. student at WPI
- Krishnan Giridharan, October 2019 – May 2021, currently an associate engineer at Rivian
- Dylan McKillip, August 2019 – May 2020, currently a test engineer at Brooks Automation
- John Trainor, aerospace engineering class of 2021 (WPI), currently a design engineer at Triton Space Technologies, LLC
- Nathaniel Rutkowski, aerospace engineering class of 2021 (WPI), currently a propulsion engineer at Firehawk Aerospace
- Kaitlin Barron, mechanical engineering class of 2022 (WPI)
- Shannon O’Connor, mechanical engineering class of 2022 (WPI), NASA LaRC summer intern (additive manufacturing research), Summer and Fall 2021
- Caitlin Kean, mechanical engineering class of 2022 (WPI)
- Nathan Maldonado, mechanical engineering class of 2022 (WPI)
- Daniel Marsh, mechanical engineering class of 2022 (WPI)
- Adrianna Yuen, mechanical engineering class of 2024 (WPI)
- Samantah Castellano, mechanical engineering class of 2024 (CMU)
- Meenakshi Sundrum, mechanical engineering, engineering and public policy, class of 2024 (CMU)
- Brenna Slomsky, mechanical engineering class of 2024 (CMU)
- Charlotte Ng, materials science and engineering, engineering and public policy, class of 2024 (CMU)