When President Barack Obama named Georgia Tech President G. P. “Bud” Peterson to the steering committee of the Advanced Manufacturing Partnership (AMP) in June, he was acknowledging an established fact—the Georgia Institute of Technology is a national leader in supporting American industry.

Tech joined other top universities—the Massachusetts Institute of Technology, Carnegie Mellon, Stanford, University of California-Berkeley, and University of Michigan—in the $500 million AMP push to guide investment in emerging technologies and increase the supply of high-quality manufacturing jobs and overall U.S. global competitiveness.

“We applaud this initiative, and Georgia Tech is honored to collaborate to identify ways to strengthen the manufacturing sector to help create jobs in Georgia and across the United States,” Peterson said. “Many of our challenges can be solved through innovation and fostering an entrepreneurial environment, as well as collaboration between industry, education, and government to create a healthy economic environment and an educated workforce.”

Today, the H. Milton Stewart School of Industrial and Systems Engineering (ISyE) leads the way in advanced manufacturing research and development at Georgia Tech. ISyE faculty specialize in many related disciplines, including computer-integrated systems, controls for flexible automation, manufacturing systems design, analysis and simulation, lean manufacturing strategies, and performance measurements.

Advanced manufacturing involves not only new ways to manufacture existing products, but also new products emerging from advanced technologies, observes Stephen E. Cross, Georgia Tech’s executive vice president for research. Cross, who is also a professor in ISyE, is working with President Peterson to support the AMP.

“ISyE’s competencies in manufacturing, logistics, supply chains, and methodological work in operations research, statistics, simulation, and decision support provide the intellectual core for a renaissance in advanced manufacturing,” Cross said recently. “ISyE’s track record of excellence, combined with equally stellar research throughout the rest of the Institute, has made Tech one of the leading research universities in the world.”

ISyE Professor Leon McGinnis is supporting both Peterson and Cross in their work with the AMP Steering Committee. McGinnis is being joined by Ben Wang, who in January will assume the role of executive director of the Manufacturing Research Center (MaRC) at Georgia Tech and also become a professor in ISyE.

Both educators will serve on a Georgia Tech working group that will focus on ways in which research and education can maximize the impact of emerging technologies on the U.S. manufacturing sector.

Other ISyE faculty serving the advanced manufacturing thrust includes Professor Chelsea (Chip) White III, Schneider National Chair in Transportation and Logistics, and Harvey Donaldson, associate chair of Industry and International programs. Both are involved in a workshop focusing on the Council on Competitiveness’s U.S. manufacturing competitiveness initiative. The meeting, planned for early 2012 at Georgia Tech, will focus on how the supply chain and logistics industry can best support U.S. manufacturing competitiveness.

“Advanced manufacturing can be viewed as a system of systems that involves design, processes, equipment, information, energy, materials, and the entire supply chain,” said Wang, who served as director of the High-Performance Materials Institute at Florida State University before coming to Georgia Tech. “This new kind of manufacturing relies on a highly educated workforce and on truly innovative research capable of furnishing the basis for new companies as well as supporting existing industry—and ISyE is uniquely positioned to supply both the skilled workforce and the innovative research.”

ISyE faculty members conduct some $6.5 million in sponsored research annually, in areas that support all facets of manufacturing and industrial systems– optimization, stochastic systems, logistics, simulation, statistics, natural systems, economic decision analysis, and human-integrated systems analysis.

Below are instances (in alphabetical order) of the cutting-edge work being performed by ISyE faculty in areas related to advanced manufacturing.

Jane Ammons, who is the H. Milton and Carolyn J. Stewart School Chair and a professor in ISyE, collaborates on reverse production systems with Matthew Realff, a professor in the School of Chemical & Biomolecular Engineering (ChBE) and David Wang Sr. Fellow. For more than ten years, the team has focused on two important areas: the recovery and reuse of carpet wastes and ways to reduce electronic waste (e-waste).

Ammons, Realff, and their team have developed a mathematical framework to support the growth of used-carpet collection networks. Such networks could help to recycle much of the nation’s annual carpet waste total of 4.7 billion pounds. The successful reuse of that carpet has a potential value of $2.8 billion, versus a cost of $100 million to send the waste to landfills.

In other work, the team is studying the problem of e-waste—unwanted electronic components such as televisions, monitors, and computer boards and chips. The e-waste stream includes multiple hazardous materials containing lead and other toxins, yet effective management and reuse of e-components can be profitable. Ammons and Realff have devised mathematical models that address the complexities of e-waste processing, with the goal of helping recycling companies stay economically viable.

“Working with both, companies and government, our goal is to eliminate as much product disposal in landfills as possible,” Ammons said. “By extending our work to address new operational control and infrastructure design problems, we can help to address uncertainty and variability in closed-loop supply chain flows on a global scale.”

 Associate Professor Nagi Gebraeel conducts research in the area of detecting and preventing failure in engineering systems as they degrade over time. The goal is to avoid both expensive downtime and unnecessary maintenance costs.

“We could be talking about a fleet of airlines, trucks, trains, ships—or a manufacturing system,” Gebraeel said. “In any of these cases, it’s extremely useful for a number of reasons to be able to accurately estimate the remaining useful lifetime of the system or its components.”

In one project, Gebraeel and his team worked with Rockwell Collins—a Cedar Rapid, Iowa, maker of avionics and electronics—to monitor and diagnose the performance of circuit boards that control vital aircraft communication systems.

Since the exact time of component failure is unknown, airlines are forced to anticipate when replacements are needed. Scheduled maintenance can result in replacement of parts that still have usable life. Using circuit boards until parts actually fail will result in unplanned and expensive downtime.

As Gebraeel methodically exposes an avionics component to heat and vibration, he employs a network of computers and sensors to record and analyze data on the degradation rate of the part he is testing. If he can reliably predict the failure rate of a component, he can help airlines replace parts at the most cost-effective time.

In another effort, Gebraeel has developed an adaptive prognostics system (APS), a custom research tool that allows him to investigate how quickly components degrade under vibration and other stresses. Gebraeel and his team can use APS to test a complex system—such as a gearbox—by using multiple sensors in a triangulated pattern to detect the frequency signals coming from individual components.

Gebraeel is currently in talks with a major airline to use APS to analyze critical engine components. The aim is to be able to predict engine wear rates in ways that will help optimize aircraft maintenance procedures.

“There’s a real need for information about the remaining life of components, so that users can find the economical middle ground between the cost of scheduled replacements and the cost of failure,” he said. “Think of the everyday problem of whether we really need to replace vehicle engine oil at 3,000 miles. If we replace it early, we sacrifice some useful time, but if we replace it later, we risk engine damage. It’s very useful to have detailed information about degradation in a system over time.”

Professor Leon McGinnis focuses on model-based systems engineering, an approach that uses cutting-edge computational methods to enable capture and reuse of systems knowledge among multiple stakeholders. McGinnis, his team, and other faculty collaborators are pursuing several sponsored projects in this area.

In one notable project, McGinnis and his team are working with Rockwell Collins, the Iowa-based maker of avionics and electronics. The aim is to help the corporation speed transition of new products by automating the process that simulates physical manufacturing.

In order to optimize the resources needed to make products at the required rate, McGinnis explains, Rockwell Collins creates a computerized simulation model of the manufacturing processes. Development of simulation models has traditionally been the province of experts who are skilled in using initial system designs to simulate the demands of actual production.

“This is not a trivial task—producing a simulation model requires some 100 to 200 hours per product,” said McGinnis, who holds the Eugene C. Gwaltney Chair in Manufacturing Systems. “Due to expert resource limitations, the company was only able to generate a few production models at a time, which created something of a bottleneck.”

To analyze the model-development process, an ISyE team interviewed Rockwell Collins engineers on the methods they used to develop a simulation model. The Georgia Tech investigators carefully analyzed the steps and methods that the engineers used to progress from an original system design to a simulation model.

Then the ISyE researchers turned to SysML, a language that enables the computerized modeling of complex systems. SysML lets designers delineate a new product—and multiple related factors such as people, machinery, and product flows—in a standardized way.

By describing the evolution of a given product using SysML, McGinnis and his team were able to automate the movement of that product from design to simulation. Even more importantly, the ISyE team created a domain-specific version of SysML that was customized to the Rockwell Collins environment. That achievement allowed any of the company’s new products and systems to be plugged into an SysML-based automation process.

This new way to doing things appears to reduce the time required to build simulation models by an order of magnitude McGinnis said. It also allows multiple products to be developed concurrently and encourages “what-if” studies that couldn’t be performed before.

“Essentially, this technology lets the people who own a process validate it without the middleman—the simulation expert,” he said. “There’s a two-part philosophy here—one is to articulate the system in a way that all the stakeholders can agree on, and then to automate the bringing of information and knowledge to the stakeholders without requiring mediation by experts.”

McGinnis is also working on several other projects. In one effort, he is collaborating with the School of Mechanical Engineering and the Manufacturing Research Center (MaRC) to develop semantics for manufacturing processes under a DARPA contract. In another project, he is collaborating with the Tennenbaum Institute to address the challenges of identifying and mitigating risks in global manufacturing enterprise networks. In other MaRC research, he is investigating the integration of product design and manufacturing management of flexibly automated production throughout an entire manufacturing system.

Spiridon Reveliotis, an ISyE professor, is currently involved in a project that addresses a cutting-edge approach to automation in manufacturing. This concept, known as flexible automation, involves variable-size batch production and the ability to reconfigure and rebalance the shop floor quickly to accommodate differing product mixes.

To date, Reveliotis explains, flexible automation has been most successful at the level of single manufacturing processes. To address this limitation, he is developing the analytical capability and computational tools to enable effective deployment and in the methodological areas that define the technical bases for these works.

Reveliotis is using the representation of a Resource Allocation System—an enriched version of a queuing network model—and also employing modeling and analytical capabilities derived from modern control theory, computer science, and operations research.

Using these, he is seeking to build a framework and methodology to enable rapid reconfiguration of automated production systems, with control logic capable of managing the system operation in each new configuration. One challenge, he said, involves managing the trade-offs between the quest for a high-fidelity model of the underlying shop floor dynamics and the need to keep the control logic and its deployment manageable.

In another project, Reveliotis is developing methods to help remanufacturing facilities approach component-disassembly tasks in the most efficient ways. This work, sponsored by the National Science Foundation, uses a learning-based approach comprised of efficient sampling techniques and novel machine-learning algorithms to determine the optimal disassembly plan for each product type.

Beyond addressing important practical problems in the manufacturing and remanufacturing domains, both of the above lines of work are also contributing seminal analytical results enterprise development for the aerospace industry.

Professor Jianjun (Jan) Shi’s research addresses system informatics and control. He uses his training in both mechanical and electrical engineering to integrate system data—comprised of design, manufacturing, automation, and performance information—into models that seek to reduce process variability.

Shi, who holds the Carolyn J. Stewart Chair in ISyE, is currently working on several sponsored projects. In one effort, Shi is working with nGimat, a Norcross, Georgia-based company that was a 1997 graduate of the Advanced Technology Development Center startup-company incubator at Georgia Tech.

nGimat is currently addressing the challenge of mass-producing a type of nanopowder for use in high-energy, high-density batteries for electric cars. With sponsorship from the Department of Energy (DoE), Shi is supporting nGimat as it works to increase its output of this nanopowder by several orders of magnitude.

“This nanopowder product has very good characteristics, and the task here is to scale-up production while maintaining the quality,” Shi said. “We must identify the parameters— what to monitor, what to control—to reduce any variability and do so in an environmentally friendly way.”

In work focusing on the steel industry, Shi is pursuing multiple projects including investigating sensing technologies used to monitor very high temperature environments used in steel manufacturing. With DoE support, he is working with OG technologies to develop methods that employ optical sensors capable of providing continuous high-speed images of very hot surfaces—in the area of 1,000 to 1,450 degrees Celsius.

In steel manufacturing, Shi explains, continuous casting and rolling lines can be miles long and production can take hours. Variations in the process temperature—currently difficult to detect—can lead to costly quality problems, increased labor costs, and increased carbon dioxide emissions due to wasted energy.

“We want to catch defect formation in the very early stage of manufacturing,” Shi said. “By using imaging data of the product effectively with other process data to eliminate defects, we can help optimize the casting process.”

In another representative project, Shi is investigating ways to use process measurements and online adjustments to improve quality control in the manufacturing of the ubiquitous silicon wafers used in semiconductor electronics. In work sponsored by the National Science Foundation, he is working with several manufacturers to examine the root causes of undesirable geometric defects in wafer surfaces.

Shi explains that the first step of his approach involves developing a software model capable of detecting and accurately characterizing surface characteristics on a silicon wafer. If waves are present, the model must be able to capture both their mean profile as well as detect and characterize particular types of waves.

The second step requires using this model to judge whether an actual wafer surface is of acceptable quality. If the surface is faulty, the model returns data on what must be done to improve it.

“Wafer manufacturing is another instance of a continuous process where, if you catch imperfections early, you can quickly and cost-effectively return to a previous step in the process and correct the problem,” Shi said.

Associate Professor Joel Sokol, A. Russell Chandler III Chair and Professor George Nemhauser, and Professor Shabbir Ahmed recently completed a project supporting a major float glass manufacturer. The company was automating a process where finished glass plates are removed from the production line and packed for shipment.

The company was concerned that the new machines that pick up and remove glass from the production line might fall behind, allowing valuable plates to be heavily damaged. What was critically needed was the capability to carefully schedule the sequence of production so the machines could function at maximum capacity with as little waste as possible.

The ISyE team tackled development of new software that could minimize production scheduling problems. They devised algorithms that allowed the machines to work at their maximum efficiency and enabled them to handle input data with more than 99 percent efficiency.

“The algorithms we delivered can also be used strategically to determine how many machines of each type should be installed on a production line,” Sokol said.

In another project, Sokol,  Nemhauser, and Ahmed are collaborating on a project for Korea-based Samsung. The aim is to support production throughput at a Samsung semiconductor- manufacturing facility.

The challenge involves the physical movement of semiconductors from one processing station to another throughout the factory. Because the routing of semiconductors between processing machines can differ from item to item, there’s no linear assembly- line type of procedure; instead, hundreds of automated vehicles pick up an item from one processing point and move it to its next step.

Because of the facility’s structure, these automated vehicles encounter congestion that can delay the production schedule, Nemhauser said. The ISyE team is developing ways to best route and schedule the vehicles to minimize congestion and move items between machines in ways that don’t delay production.

“This is clearly a highly complex challenge that will require development of an accurate system model,” added Ahmed. “But it’s exactly the type of problem that can be solved by devising effective software and hardware modifications.”

Valerie Thomas, Anderson Interface associate professor of Natural Systems in ISyE, is conducting research on the use of information technology, mediated by bar codes or radio frequency (RFID) tags, to improve recycling and end-of-life management for electronics and other products.

This work has been presented to the U.S. Environmental Protection Agency and the U.S. Congress and has been featured in the New York Times and the Wall Street Journal.

In another area, Thomas is collaborating with Professors Matthew Realff and Ron Chance in the School of Chemical & Biomolecular Engineering (ChBE) and with ISyE PhD students Dexin Luo and Dong Gu Choi on the design, energy efficiency, water management, and carbon footprint for facilities to produce biofuels. This work is supported by Algenol Biofuels as part of their $25 million DoE-funded pilot plant for the production of ethanol from cyanobacteria.

Associate Professor Chen Zhou, associate chair for undergraduate studies, and Professor Leon McGinnis tackled sustainability issues for Ford Motor Company in a recent project.

The issue involved shipping gearbox components from China to the United States in ways that would minimize not only cost but greenhouse gas emissions and waste.

It turned out that packaging was at the heart of the issue. The researchers had to configure component packaging so that the maximum number of components could be placed in a cargo container yet also allow for optimal recycling of the packing materials to avoid waste and unnecessary cost.

“This was definitely a complex problem,” Zhou said. “You must track every piece of packaging from its source to its final resting place, when it either goes into another product or into a landfill.”

The team created a model—a globally sourced auto parts packaging system— that optimized cargo container space. The model also enabled the use of packing materials that were fully reusable; some materials were sent back to China for use in future shipments, while the rest was recycled into plastics that became part of new vehicles.