Craig Stephens, right, makes a point during a tour of the Dryden Flight Loads Laboratory. From left are Dryden researcher Sunil Kukreja and two University of Michigan students with whom he worked on a recent project, Tony D'Amato and Amy Wu. (NASA Photo)
› View Larger Image Students Contribute To Dryden Researcher's Work And Gain Insight Into Engineering Careers
When Dryden researcher Sunil Kukreja assembled a team for an Innovative Partnerships Program Seed Fund effort to begin looking at health monitoring of composite material airframes, he did so with partners from the University of Michigan, Ann Arbor.
Kukreja, who holds a doctorate in nonlinear system identification, forged the partnership with his long-time University of Michigan colleague Dennis Bernstein, who has a doctorate in control engineering, and Bernstein's students. Recently, Amy Wu and Tony D'Amato, two of Bernstein's students involved in that research, visited Dryden to share the results of the collaboration.
The collaboration took the team steps closer in the theoretical mathematical formulation and development of algorithms required for on-line health monitoring of next-generation aircraft made from composite materials. The work could ultimately allow for greater passenger safety and cost savings for airliners, enabling users of the technology to learn from the aircraft when and where maintenance is required and reducing the need for long inspections currently required, Kukreja said.
The way the team was looking at achieving the goal was by development of a sensors-only, fault-detection approach using pseudo-transfer function identification, Kukreja explained. The latter refers to identifying the pseudo-transfer function – a type of mathematical representation – between two sensors in the presence of variations from the baseline operation and external factors during ambient conditions. Validation of the fault detection architecture could allow health monitoring in real time on board with the addition of the proposed algorithm and software to existing cockpit computers.
The work has been a success and Kukreja and his team learned June 22 that they were awarded an Aeronautics Research Mission Directorate Seedling Fund award for new reearch on multiple input multiple output, or MIMO, formulation and the nonlinear effects on sensor-only system identification. This award advances the team's previous work.
Wu, a second-year graduate student in aerospace engineering, worked on the system identification research part of a sensor-only fault detection. It was a positive learning experience, she said.
"It's good to work your problems with someone who gives you really good advice," she said. "When I came to Dryden to present my work, I found out that the math and theory had application not only to what I am doing, but that there are all of these unsolved problems out there that this research might make a difference in. Dryden researchers brought up things we had not thought of, and we went back and thought about it."
The research project directly related to Wu's controls studies.
"Sometimes you don't quite know how what you learn in class fits into the real world. Now I can see how they relate. I learned about some of the questions Dryden engineers asked me about in class, but I think more about it now that I've heard it from them [the researchers]. Sometimes you get lost in the research. What you are trying to achieve, getting the codes to work – it can be frustrating. But I learned it's the nature of research and you just have to deal with it and keep at it," she said.
During the visit and tour of Dryden, she had a favorite part.
"I worked on the 747-8 at Boeing and I liked seeing the Shuttle Carrier Aircraft. It had a special meaning for me. Being in the aircraft and knowing it carried the shuttle was pretty cool. I know people who have been to Dryden, but now I have seen it for myself. I really appreciated the level of questions from the engineers and their interest and what they shared with me," she said.
D'Amato's work focused on the development of structural-health monitoring techniques based on system identification and estimation methods.
"Thus far we have developed techniques for the linear case and several nonlinear models. The goal of the research is to use models obtained with identification and estimation to facilitate damage detection and localization in aerospace structures," he said.
D'Amato said he learned much from the project.
"The most important concept that I learned from this project so far is the difference between practical application and the theoretical world," he said. "Most of our ideas start in the theoretical world under a specific set of assumptions, specifically, that the world is linear. Under this assumption we have many theoretical tools at our disposal to prove ideas and demonstrate concepts.
"The problem is that the world is nonlinear and, worse yet, most of the assumptions we impose on our methods and algorithms are violated when operating in the real world. Through this project, I had the opportunity to work with real data gathered from experimental setups. This gave me the chance to reflect critically on my work so far and determine which assumptions were valid and which were unrealistic," he said.
The hard work paid off.
"The most rewarding aspect of the work so far has been the presentation of our results at national and international conferences. Meeting other researchers interested in the same types of problems, discussing the complications and possible resolution helps to give the project perspective. When I get stuck, I think back to these conversations and realize that I am not the only one struggling to solve these issues. Sometimes a simple informal discussion about the work has led to great ideas, which helped to push the project forward," D'Amato said.
Working with researchers helped focus his efforts.
"The overall goal of the project is fairly clear, namely, to detect and find damage in aerospace structures. The problem is that this is really an interdisciplinary subject, specifically, structures, dynamics and modeling with many potential routes and just as many dead ends. I have found it easy to get discouraged. However, working with experienced researchers has helped me tackle issues under the simplest conditions and then expand on these ideas until the issue is resolved. I have learned how to conduct research in a scientific fashion and then communicate my ideas and results in a clear and concise fashion to other researchers," D'Amato said.
There were challenges to overcome.
"The greatest challenge so far was recognizing the multidisciplinary nature of the work and learning how to incorporate advice and guidance from my peers and mentors. I was challenged to function as a team member to solve a complicated problem. As a senior student in our research group, I am expected to act as a mentor to incoming students. This perspective helped me realize that engineering is fundamentally a team effort and the unique experiences and backgrounds of the individual team members are what make the group more capable of solving difficult problems," he said.
D'Amato said he enjoyed the tour of Dryden facilities.
"We were offered great explanations and discussion on the various projects and experiments. I enjoyed the static aircraft displays, especially the SR-71 and the X-15. I had never seen these aircraft in person, but had read extensively about the programs. Without a doubt, it was projects like these that influenced me from early on to become an aerospace engineer. It had also been a dream to work with NASA, so getting the opportunity to visit and tour the facility was something very meaningful for me," he said.
He would recommend other students take advantage of working with Dryden.
"Working with the researchers at Dryden has helped me make the connection between the fundamental tools we learn as students in engineering and real-life problems faced by researchers advancing the state of the art," D'Amato said.
Kukreja said the students worked hard on the project and he enjoyed working with them.
In fact, Kukreja enjoys sharing his knowledge with students and peers. For example, last summer he shared his specialties in non-linear system identification as the keynote speaker at the International Conference on Non-Linear Problems in Aerospace World Congress in Brazil.
With the University of Michigan students, Kukreja said, solid work was accomplished.
"We developed good algorithms for a sensors-only approach. That means we don't need inputs to excite the system to determine the health of a structure," he said.
With the revolution under way with fiber optics, determining structural health will be possible in the future. Kukreja is already analyzing real flight data from accelerometers, as well as strain and loads data obtained from fiber optic sensors, in preparation for in-flight research to obtain critical information about aircraft health and or performance in real time, as it is being collected on an aircraft in flight. Such an algorithm could then be fully developed, paired with the mathematical estimation work and packaged for commercial use.
"The algorithm will work in a simulation environment such as Simulink in Metlab," he said, referring to the software used in testing, "but does it work in reality with interaction and complications in real-world situations? Will the flight test data validate the algorithm for structural health monitoring?"
In time, progress on the algorithms and mathematical framework for the concept could lead to a new day in safety and maintenance of composite structures. For now, researchers such as Kukreja will continue to work and collaborate with students and peers for the benefit of increasing the technology readiness level of these new concepts.