PITA Fiscal Year 2018 Projects


Study of the Spreadability and Printability of Hydride-Dehydride (HDH) Ti-6Al-4V Powders in Electron Beam Melting Process
PI: Anthony Rollett
Co-PI(s): Sneha Prabha, Jack Beuth

Additive Manufacturing (AM), also known as 3D printing has attracted interest from industry and academia in the past few years. Among direct metal 3D printing processes, powder based processes are an important category that utilizes powder as a feedstock material. Metal powder is one of the contributors to the final cost of the component. Hydride-Dehydride (HDH) powders are approximately 50% cheaper than the regular gas atomized powders that are spherical in shape. However, HDH powders are non-spherical and require the development of spreading and deposition parameters that are different from the parameters developed by the machine manufacturer for atomized powders.

The goal of the proposed project is to develop methods to use the HDH powders in electron beam melting (EBM) process. Specifically, spreading and deposition parameters will be developed followed by the study of as-built porosity and surface roughness. In addition, powder flow characteristics will also be studied and the part quality will be compared against the parts built using spherical powders. Though, the methods will be developed for Titanium powders (Ti-6Al-4V) in EBM, the general technique can be extended to other materials and other beam based powder bed process such as selective laser melting.

This project utilizes effective collaboration between CMU and Ametek Powders to enable the use of HDH powders in powder bed additive manufacturing processes. This capability will eventually result in reducing the cost of feedstock powders by approximately 50% which is beneficial to both the powder companies and the end users of powder bed processes. Along with Ametek, other powder manufacturing companies in PA such as ATI Powder Metals Research, Carpenter Powder Products, and Arconic will also benefit by manufacturing cheaper powders that can be used in AM. Graduate students involved in the project will gain AM expertise that is valuable to the industries mentioned above.


Computational Methods for Enterprise-wide Optimization under Uncertainty
PI: Ignacio Grossman

Our main vision has been to develop advanced computational models and solution methods for Enterprise-wide Optimization (EWO) for process industries. A major challenge that is involved in EWO is the integrated and coordinated decision-making across the various functions in a company (purchasing, manufacturing, distribution, sales), across various geographically distributed organizations, and across various levels of time scales (strategic, tactical and operational). A major focus of the proposed PITA project, which is a collaboration with EQT, is the optimization of infrastructure investment, operations and water management of shale gas supply chains accounting for uncertainty in the decline curves, and gas demand and prices.

The funding requested for 2018 PITA, $25,000, is for the period January 1, 2018-December, 31, 2018. Funding is requested for the Ph.D. student Can Li who will be developing stochastic programming methods to anticipate the effect of uncertainties in decline curves, and gas demand and prices. Can will work closely with Dr. Markus Drouven from EQT to gather data to develop a case study that will be used to test the proposed ptimization models in the Marcellus Shale play.

The special interest group EWO has been created with the membership of 12 companies: ABB, Air Liquide, Aurubis, Braskem, Dow Chemical, EQT, ExxonMobil, P&G, Petrobras, Praxair, SK-Innovation and Total. Since these companies are members of the Center for Advanced Process Decision-making (CAPD http://capd.cheme.cmu.edu, in addition to the basic annual membership of $20,000, they pay a fee of $13,500 per year for the EWO project (http://egon.cheme.cmu.edu/ewo/). With Dow we have set up special projects that provide full support for graduate students. We conduct semi-annual EWO meetings in which the progress of the case studies are reported, and where industrial representatives give presentations. We are currently undertaking 13 case studies related to shale gas, petroleum processing, and electric power.

Soft Magnetic Materials Development for Energy Applications
PI: Michael McHenry
Co-PI(s): Paul Ohodnicki

Michael E. McHenry-PI, and Paul R. Ohodnicki-co-I, propose work leveraging funding from Carpenter Technology (PA), the National Energy Technology Lab (NETL) (PA) and Carnegie Mellon Univ. (CMU) to benchmark soft magnetic magnetic materials (SMMs) in rotating machinery with increased energy efficiency. Co- and FeNi-based metal amorphous nanocomposite (MANC) alloys are being studied with DOE SunLamp funding (CMU subcontractor to NETL). A CMU led DOE Advanced Manufacturing Office (AMO) program targets Co- and FeNi-based MANCs for high-speed motors (HSMs). Frequency limitations with silicon-steels, Carpenter FeCo Hiperco, and other conventionally cast and rolled alloys, are typically < 400 Hz. MANCs enable higher speed motors (magnetic switching frequencies) and higher power densities. Worldwide studies cite > 50% of the world’s power to pass through motors, making increased motor efficiency attractive for sustainable energy demands. Three project aims are: (1) with PITA funding for a CMU MSE grad student, to benchmark MANCs against crystalline materials in finite element analysis (FEA) of motor designs; (2) with Carpenter support of a 2nd CMU MSE grad student to explore MANC glass former modifications to improve mechanical properties and formability in shapes needed in novel motor designs; (3) with Year 3 SunLamp funding for CMU Ph.d student, Natan Aronhime, to benchmark economics of Co-based vs. FeNi-based MANCs for which CMU has IP on both. CMU will interact with Carpenter through Sam Kernion, Manager Alloy Design & R&D, a 2012 McHenry group Ph.d. Yuval Krimer, who will receive a 2018 CMU M.S. degree in MSE, will intern the summer of 2018 at Carpenter and if accepted return to complete a Ph.d beginning Fall 2018. The second PITA-funded CMU student will be identified to also begin in Fall 2018. CMU will interface with NETL through Paul Ohodnicki, Staff Scientist and CMU Adjunct Professor, a 2008 McHenry group Ph.d.

Electrolyte Design through Physics-Driven Machine Learning
PI: Venkat Viswanathan
Co-PI(s): Jay Whitacre

The aim of the proposed project is to enable a rational design approach for electrolytes via a combination of physics-driven models that are coupled with large datasets and machine learning. The design of an electrolyte for properties such as conductivity, voltage stability depends on a large number of properties. While physics-driven models can identify simple descriptors, these tend to be inadequate for actual material selection. The advent of big data and machine learning allows the opportunity to couple physics-driven descriptor selection for machine learning. This effort will leverage our current capabilities around SEED, System for Electrolyte Exploration and Discovery, which contains an exhaustive dataset on liquid electrolytes. In partnership with Citrine Informatics, we aim to leverage the Citrination platform to carry out advanced electrolyte discovery.

Mitigating Cracking of Steel Slabs
PI: Bryan Webler

The 3rd Generation of Advanced High Strength Steels (AHSS) are currently under development for advanced lightweight vehicle applications. Before they can see wider adoption, a cracking problem in semi-finished slabs must be solved. This cracking problem results in constrained operations at steel production facilities and high scrap rates, i.e. wasted money and energy. Initial industrial observations suggest that steel microstructure produced after casting determines slab cracking susceptibility. Microstructure evolution in these steels is complicated by their high levels of alloying elements. In this project, we will perform heat treatments on controlled-composition steels to examine conditions under which potentially susceptible microstructures develop. We will then test the mechanical performance of steels with various microstructures via Charpy impact testing. Completing these studies will link steel composition, processing, and properties in the as-cast condition. Developing these links will enable strategies to mitigate detrimental microstructures and reduce cracking of AHSS slabs.


Chemotopographic Control of Adhesion in Complex Fluids
PI: Christopher Bettinger

High-performance adhesives that are effective in fluids have utility as materials for medicine, consumer projects, and industrial applications. To date, the design and implementation of materials that provide robust adhesion in aqueous or oily environments has been elusive. Various elaborations of bioinspired approaches have been implemented to overcome this formidable technical barrier. One such approach is to functionalize materials with catechol groups, chemical motifs in proteins that are used by organisms to adhere to inorganic surfaces in marine environments, for example. Catechols in these specialized natural proteins increase the interfacial adhesion to many substrate materials by forming coordination bonds, hydrogen bonds, and aromatic interactions. PROBLEM: To date, the design of synthetic catechol-bearing adhesives has been moderately successful, but many technology gaps remain. Three key problems with existing catechol-bearing adhesives: 1) the areal density of adhesive catechols is low (<10 nmol/cm^2); 2) the adhesive polymers delaminate from substrates (when used as a thin film) or undergo cohesive failure (when used as a bulk material); 3) interfacial chemical bonding is reduced in rough surfaces. INNOVATION: We have recently discovered a novel technique to synthesize and transfer print catechol-bearing nanomembrane adhesives to virtually any polymer substrate. These adhesives, termed polydopamine nanomembranes, contain catechols at areal densities of ~26 nmol/cm^2. Furthermore, polydopamine nanomembranes can be covalently bonded to many industrial polymers, therefore reducing the risk of delamination of the functional adhesive film from bulk substrates. APPROACH: In this project, we propose a partnership with nanoGriptech, a Pittsburgh-based company founded by a CMU professor that fabricates microstructured adhesive polymers for use in consumer products and industrial applications. Microstructured adhesives will be combined with polydopamine nanomembranes to address challenges with surface roughness. Specifically, we will integrate polydopamine nanomembranes with microstructured substrates to create a new class of high-performance adhesive that is effective in both aqueous and oily environments. HAZARD MITIGATION AND DISASTER RECOVERY Automated Aerial Sensor Data Analysis for Detection of Abandoned Oil Extraction Wells PI: Burcu Akinci Co-PI(s): Silvio Maeta The Pennsylvania oil boom launched the american commercial oil industry when the first well was drilled in 1859 near Oil Creek,Venango County. Pennsylvania was one of the first states to have oil exploration fields in the country, starting around 1860’s until the early 1900’s. Now, most of those oil exploration fields are covered by dense forests and the abandoned wells present an environmental and societal hazard since they can leak gas and/or they can collapse, causing damage to property and endangering people. To eliminate such hazards, it is important to identify all of the wells in the region and seal them. However, this is a challenging task given the absence of records that show where those wells were drilled and the absence of visual marks that indicate the location of those wells. Any indication of man-made structures are mostly gone after one century. To address this problem, we propose to team up with a NETL team and utilize data from aerial lidar surveys to develop an automated processing approach to detect abandoned wells. Other sources of information (e.g. magnetic field variations caused by buried metal pipes, old pictures of Oil Creek, GPS position of already identified wells) will be combined to improve detection accuracy. This project will evaluate different automated machine learning and deep learning approaches to assess their performance and feasibility in detecting abandoned wells. We will conduct a detailed field case study with NETL in which we will process 3D imagery data and generate a list of candidate locations for abandoned wells. We will verify if the detected well candidates are valid against existing ground truth data and measure the confidence level. Our team is well equipped to perform this research bringing knowledge on aerial robots, 3D imaging and data analytics applied to civil engineering problems. Teaming up with NETL will provide an unprecedented access to data and resources to perform a detailed case study at Oil Creek.


Cost-Competitive Mass Customization for Non-assembly Manufacturing Environments in the Greater Pittsburgh Region
PI: Katie Whitefoot
Co-PI(s): Erica Fuchs

Manufacturers are currently facing a challenge of balancing the potential of mass-customization to cater to diverse customers with the costs of increased product diversity in their production lines. While research on mass-customization in assembly-based manufacturing is prevalent, very few methods exist for non-assembly fabrication (e.g., forming, casting, drawing, additive manufacturing, etc.), which dominates many manufacturing activities in PA. The proposed research will generate new knowledge and methods to support non-assembly-based mass customization, and to characterize the impacts of product variety on production costs in these environments. The project will be conducted in collaboration with Kennametal, a company headquartered in PA that provides innovative custom and a large variety of standard metal cutting tools. The project will focus on a Kennametal plant located in the Greater Pittsburgh Region as a test-bed before expanding out to additional fabrication manufacturing facilities in PA.


Minimalist Electrodes for Suppressing Brain Tsunamis Through Noninvasive Neurostimulation
PI: Pulkit Grover
Co-PI(s): Marlene Behrmann, Michael Tarr, Shawn Kelly

The goal of the proposed work is to design and develop novel minimalist electrodes and electrode waveforms to noninvasively stimulate neurons in the human brain. In particular, our focus is on precise, reliable, and steerable patterns of noninvasive neurostimulation that can be used to suppress “Brain Tsunamis,” i.e., Cortical Spreading Depolarizations (CSDs). These waves of neural silencing occur during migraine attacks, as well as after brain injuries, and are known to cause secondary brain injuries. These disorders affect millions of Americans every year. Suppressing these waves is therefore a problem of immense societal importance.

The PIs have already developed the first automated algorithm for CSD detection, and presented it at the World Congress on Brain Injury. The proposed work takes it a step further and aims at suppressing these CSDs in an automated and noninvasive fashion. Our focus is on electrode design, while complementary work (from a cost-sharing partnership with UPMC enterprises) obtains algorithms for detection and suppression. Our goal here is to have electrodes that a) require minimal setup time (a few seconds, instead of the usual hour or more); b) have the smallest electrode count while having current waveforms that precisely stimulate only the target regions, and not elsewhere.

To accomplish this, the PIs will advance on and integrate three of their recent innovations: (i) Conductive sponge-based electrodes; (ii) “STIMULUS” technique for deep, focused noninvasive neurostimulation without stimulating shallow neurons; (iii) Precise and dynamic current stimulation-based CSD suppression. The obtained techniques and electrodes will be validated through ex-vivo and human experiments, demonstrating improvements in noninvasive stimulation accuracy over existing modalities. Precise, noninvasive, and long-term-use neurostimulation will find many applications beyond CSD suppression.

Non-invasive intracranial pressure monitoring in traumatic brain injury
PI: Jana Kainerstorfer
Co-PI(s): Pulkit Grover

There is a critical clinical need for non-invasive intracranial pressure (ICP) monitoring. Current methods are invasive and are only applicable to severe traumatic brain injury. Changes in cerebral autoregulation, which is the brain’s mechanism to maintain blood flow despite pressure changes, after brain injury may contribute to cerebral ischemia, elevated ICP, and/or cerebral hyperemia and may worsen patient outcome. In order to quantify autoregulation and manage patient’s health based on autoregulation and perfusion, ICP needs to be measured. Using a combination of optical methods, such as near-infrared spectroscopy (NIRS) and diffuse correlation spectroscopy (DCS), in acute measurements of increased ICP in non-human primates, we found that hemodynamic changes (as measured with NIRS and DCS) can be used to monitor ICP as well as autoregulation non-invasively. The animal studies have been funded by AHA as well as NIH and are ongoing. Preliminary results demonstrated that cerebral hemodynamic changes as measured with NIRS in combination with a transfer function analysis approach can yield ICP traces and therefore feedback about the autoregulatory state of the brain. Translation of the methods into a clinical setting at the Children’s Hospital in Pittsburgh was proposed in our CMLH submission, which is used as matching funding and is the core of the proposed work. The PITA proposal expands the proposed study by developing a combined NIRS/DCS system, which is an important step towards ease of use in the clinic.


Test Chip Design For Maximal Yield Learning
PI: Ronald Blanton

The primary anticipated result of this proposed work is a silicon-based validated methodology for uncovering systematic defect mechanisms through a comprehensive methodology for the design, test and diagnosis of logic characterization vehicles (LCVs). Ideally, an LCV is a test chip composed of interconnected standard cells that is fabricated and tested in volume to validate the capability of a new technology to yield working, reliable logic circuits in actual customer products. Conventional approaches for LCV design do not ensure however that the resulting vehicle is both highly testable/diagnosable and simultaneously reflective of actual product designs. At present, the Carnegie Mellon methodology produces vehicles that are transparent to single, static failures, and reflect logic characteristics (i.e., cell-instance demographics) of customer designs. In this task, we will extend design reflection to include physical (layout) characteristics from cells to back-end interconnect, and transparency to multiple inter- and intra-block failures, and dynamic/parametric failures. Most importantly, we will work with industry to demonstrate the effectiveness of the methodologies with data measured and analyzed from manufactured CM-LCV designs. We are planning to fabricate and test a significant number of CM-LCV test chips using the most state-of-the-art technologies. Over 60 of our designs have already been fabricated in volume in state-of-the-art factories located around the world. Tester data measured from these chips will be analyzed using custom diagnosis software developed in this work. In order to gauge the capabilities of this new software we plant to compare the results (diagnostic accuracy, resolution, analysis time, etc.) with the capabilities of existing commercial tools. Our main industrial partner, PDF Solutions, has a significant footprint in Pennsylvania and success of the proposed work has the potential of growing the number of employees in the state and establishing other, PA-based companies in the semiconductor space.


Data-driven Design of Frequent Transit Service Network in Allegheny County
PI: Sean Qian

Public transportation offers shared transportation service and plays a pivotal role in regional economic develop. Efficient, reliable and sustainable public transportation service would help boost local business, re-shape cities with environmentally friendly and dense land-use, and ultimately support smart and connected communities. The Port Authority of Allegheny County provides public transportation services in Allegheny County and the City of Pittsburgh via 97 bus routes, 2 light rail lines, and 2 inclined planes. Most of these routes are operated as a hub and spoke model, with the majority of the transit services connecting near downtown Pittsburgh. Recent decline in total ridership in the Allegheny County has negative impact on financing and planning transit services. This research will help the Port Authority of Allegheny County understand: 1) how future investment or reorganization of services towards routes with frequent service could affect its ridership; and 2) the spatial distribution of jobs, population and local business that transit riders can access through a Frequent Transit Service Network (FTSN). This PITA project will use high-resolution automatic passenger counters (APC) and autonomic vehicle location (AVL) technology, coupled with traffic data on roadway and Census data, to design a Frequent Transit Service Network (FTSN) for Allegheny County given a limited budget. The total estimated service cost, social benefits and financial benefits will be computed and provided for each frequent route selected for the FTSN. In addition, additional service routes will be suggested to connect those pairs of traffic analysis zones which do not currently provide a direct service route while being potentially critical for increasing total ridership. A prototype web application will be developed to visualize the recommended FTS in details.