Enabling Next-Generation Aerospace Development
From Armor & Mobility, July/August 2019 Issue
The U.S. Air Force has stood up a third Advanced Technology and Training Center (ATTC) to promote artificial intelligence and autonomous capabilities evolution.
By Ms. Debora Naguy, ATTC Director
This past April, the U.S. Air Force officially opened the third Advanced Technology and Training Center (ATTC). The focus of the new ATTC is on advancements in robotics and artificial intelligence (AI) for the United States Air Force. The ribbon-cutting ceremony opened with remarks from Dr. William Roper, Assistant Secretary of the Air Force for Acquisition, Technology, and Logistics and Lieutenant General Robert McMurry, Commander, Air Force Life Cycle Management Center (AFLCMC) followed by a facility open house and reception.
The Advanced Technology and Training Center, located in Pittsburgh, PA, (ATTC-PITT) was selected as the location for the third ATTC stand-up because Pittsburgh is a recognized city for robotics and AI advancements. By locating the ATTC-PITT in the center of it all, the USAF can effectively leverage and expedite adoption of information age solutions. ATTC-PITT operates out of a 12,500 square-foot contracted facility space, enabling the right opportunities for collaboration between academia, industry, research, and the Air Force maintenance community.
Growing Industrial Base
Pittsburgh is advancing in the industry of robotics, AI, and machine learning. Major companies such as, Google, Apple, Amazon, Disney, Bosch, Ford, Uber, GE, and Facebook are all present in Pittsburgh to leverage the robotic and AI research and development.
There are 33 regional colleges and universities, including Carnegie Mellon University, University of Pittsburgh, Penn State, and Robert Morris providing a very high concentration of quality talent. Carnegie Mellon University is a powerhouse in the fields of Robotics and AI, offering the first Ph.D. in Robotics and undergraduate degree in AI, and consistently ranks as a top school for computer science and engineering.
Beyond Pittsburgh’s higher education organizations, the region has a motivated support network to fuel the growth of its high-tech startup community, with organizations like Innovation Works and the Richard King Mellon Foundation. Pittsburgh also has an active Technology Council, fostering a successful “incubator” environment.
Two Pittsburgh-based small businesses operate the ATTC-PITT under the guidance of Air Force Life Cycle Management Center, Product Support Engineering Division (AFLCMC/EZP). Titan Robotics, Inc. is closely tied to Carnegie Mellon University and has received an Edison Award and Department of Defense Great Idea Award for innovative automation with aircraft coating removal. PavCon, LLC is a key player in laying the foundation for the Air Force Condition Based Maintenance Plus (CBM+) program using machine learning and AI to predict component failure before affecting the mission. ATTC-PITT is equipped to deliver next- generation maintenance and sustainment solutions to the USAF. The ATTC team is dedicated to developing these technologies and deploying the solutions across the enterprise.
Building on Proven Capability
The first ATTC was established in 2017 in Dayton, OH with a focus on collaboration, innovation, and hands-on training. Due to the success of this ATTC, a second ATTC was established in Middle Georgia near Robins Air Force Base in October 2018. These ATTCs are focused on additive manufacturing, cold spray, corrosion prevention and control and laser de-paint and are strategically located near centers of engineering excellence. Both facilities perform testing, validation/verification and qualification while working in coordination with Air Force end users. The team is preparing for the stand-up of a fourth ATTC near Hill AFB, Utah that is focused on agile manufacturing and composite repair. This ATTC will be a joint effort between AFLCMC/EZP and the Air Force Sustainment Center’s RAPTOR.
The Air Force came of age in the mid-1900s as the United States transformed from the industrial age. Today, the information age is revolutionizing every facet of business and military operations. With this metamorphosis, the Air Force has become “data-rich.” The value of this data is limited without Subject Matter Expert (SME) insight. Once aggregated, the data can be transformed into information for use in Air Force decisions, ranging from senior leader dashboards, Major Command (MAJCOM) reviews, engineering review and assessment, supply forecasts, and maintainer guidance and feedback. Today, the Air Force Life Cycle Management Center, Product Support Engineering Division (AFLCMC/EZP) is executing this transformation with the implementation of Condition Based Maintenance Plus (CBM+). It is the automation of this data analysis that is supported through machine learning and Artificial Intelligence (AI).
As noted in the 2018 Department of Defense Artificial Intelligence Strategy, AI is the ability of machines to perform tasks that normally require human intelligence, such as recognizing patterns, learning from experience, drawing conclusions, making predictions, or taking action. In the age of machine learning, AI is ubiquitously reinventing the business model, and the AF is adopting AI to create a smarter, more lethal force.
Condition-Based Maintenance Plus
The U.S. Air Force is using the latest in condition-based maintenance (CBM+), a holistic methodology that utilizes data to help maintainers, logisticians, and engineers making proactive, knowledge-based decisions. Successfully implemented, CBM+ will greatly reduce and eventually eliminate unscheduled maintenance, facilitate digital flight line requirements, streamline maintenance operations into efficient human-centered processes, enhance engineering effectiveness, optimize supply chain support, and improve asset generation and fleet awareness. The overarching goal is to ensure agility and flexibility for the Warfighter. The CBM+ program is creating a cloud-based data ecosystem to provide the Air Force access to its own data.
CBM+ is not just one tool or process; it is an integration of procedural and technical tools and processes that work together to provide a unique and optimized solution for each weapon system platform. This suite of solutions includes hardware, software, and communications tools that offer big data analytics and AI in a cloud-based environment, and predictive maintenance monitoring (diagnostics and prognostics) with an interactive maintenance interface to harness the power of data.
CBM+, at its core, revolves around data, and the types of data available, dictate the CBM+ pathway(s) that can be executed. CBM+ has two distinct pathways, including predictive algorithms and enhanced Reliability Centered Maintenance (eRCM), respectively. Predictive algorithms are derived from on-board diagnostic data and/or flight data recorder files to make health-based, on-condition maintenance recommendations
Within the CBM+ cloud, data is enriched into a usable format and analyzed through cluster computing resources to identify statistically significant events. The development of these events establishes thresholds by which an “algorithm” can be formed and applied to future data sets to identify anomalies, or potential, impending failures of components. To develop a predictive algorithm for a component or sub-system, there must be accessible, on-board flight data through a network of sensors or data recorders. In-flight data may include component position, temperature, and pressure that can be analyzed for behavior trends to identify failures.
Maintenance narratives are key to understanding when and if failures occurred. Maintenance data is then cross-checked against the on-board data. A review of new flight data can reveal the need for component removal or repair, thereby preventing the potential for mission aborts or delays. eRCM utilizes maintenance and aircraft usage data to forecast remaining component life. It merges maintenance data with flight hour data to produce a failure distribution and recommended maintenance window before a part will fail. These eRCM forecasts support the schedule and performance of maintenance at the time and place when it is most advantageous to the Air Force mission. The tools used for these CBM+ pathways are part of the CBM+ Toolbox that support standardization and automation of forecast processes and predictive maintenance alert reviews.
Machine learning and AI have become key players with the automation of new data ingestion and predictive algorithms updates and eRCM forecasts. As an AI-enabled capability, CBM+ is predicting failures of critical components before they occur and suggesting maintenance before failure and feeding this information to Air Force Supply for spare part forecasting. This increases the bandwidth of the CBM+ team to on-board new algorithms and weapon systems. As the models evolve, the Air Force will realize benefits through increased mission effectiveness, increased aircraft availability, and reduced life cycle costs. CBM+ and AI will continue to expand and automate as the pathway for this foundational process is laid for the Air Force enterprise, led by AFLCMC/EZP.
Automation and Robotics
The new era of advanced computing, robotic capabilities, laser optics and sensors has allowed for a revolution in the aircraft maintenance world. The Air Force has been investigating alternate methods of coating removal for the outer mold line of aircraft for decades. The original process utilizes harsh chemical solvents and produces millions of gallons of contaminated waste water. An alternative, media blasting, has been implemented for some aircraft but still produces large amounts of hazardous waste. The robotic laser method uses a thermal degradation process to remove coating minimizing waste while also removing the technician from the hazardous environment.
For several years, Air Force Life Cycle Management Center’s Product Support Engineering Division (AFLCMC/EZP) has led the charge on reducing flow days and minimizing human exposure to hazardous waste by utilizing the latest technology in robotics and lasers. The team is working to find solutions and effective recipes to remove coatings in a clean environment.
In 2017, AFLCMC/EZP obtained airworthiness certification of the Robotic Laser Coating Removal System (RLCRS) for use on common aerospace materials. This mobile robotic system is currently in use at Hill AFB Utah where it uses a 6 kilowatt continuous wave fiber laser to de-paint F-16 aircraft. This robotic laser system automatically scans the aircraft and de-paints with little user interaction required. Two technicians are in a control booth out of harm’s way from hazardous material byproducts as those are vacuumed up by the system. The waste generated is roughly equivalent to a bag of sugar as opposed to the large fifty-five gallon drums that hold the media blast waste by-product. Flow days for the maintenance cycle of the aircraft have also been reduced as compared to our full media blast process. Through the RLCRS, the depot has been able to return aircraft back to mission-ready status for the warfighter at an increased rate. Efforts to scale-up and improve the laser to a more powerful 12 kilowatt continuous wave system are currently underway for the RLCRS. The target is being able to remove all Air Force coatings used on aircraft across all substrates.
Another area of focus for the Air Force is implementing a robotic paint solution. The advantages of implementing a robotic paint solution are very similar to the benefits of the RLCRS including reducing flow days and minimizing human exposure to hazardous waste. A robotic solution would reduce human variation in the paint process to ensure the proper amount of paint is applied. Implementing a robotic paint solution will save costs by streamlining the number of laborers required, reduce the time necessary to paint, and provide efficiencies in the paint process.
The team is working on developing a robotic laser de-paint system to remove belly tape and paint from the underbelly of the C-130. The current process of removing the protective belly tape under the C-130 weapon system is time-consuming, labor intensive, and requires the use of hazardous chemicals. By addressing a difficult area of coating removal for the depot, the system will reduce the total downtime and allow for parallel sustainment activity. Follow-on efforts will include a full C-130 robotic de-paint system.
There are many positive attributes regarding automation capabilities: improved environmental impact, increased workplace safety, and saving millions of taxpayer dollars for sustainment of our aircraft. However, the most important attribute is our ability to provide increased mission readiness. Returning our aircraft into service more quickly increases our strategic capabilities and better equips our Airmen.