The Boeing 787 Dreamliner is a huge innovation. The first new aircraft launched in more than a decade, Boeing uses incredibly advanced manufacturing technology to build a lighter-weight carbon composite plane for improved fuel-efficiency. In addition, the planes include a number of state-of-the-art design features to increase passenger comfort on long-haul flights.
At first, the planes appeared to be a huge success and flew thousands of times without incident. I even flew one from Boston to Tokyo last summer. However, two Dreamliners had failures in the lithium-ion battery system in January, causing the fleet to be grounded and Boeing to delay future deliveries.
Why did this happen? Some people say that all new technology is risky. But that’s not really the issue here. Your life isn’t at stake when you use a new mobile phone technology. Boeing’s battery failure derives from the fact that this new technology is embedded in a highly complex system.
In complex systems, you can have thousands of interconnections of parts, which will create unanticipated and unintended interactions. It is likely that one or more of those interactions is causing this failure in the Dreamliner’s battery system.
Complex, system-integration failures can not only be very difficult to anticipate, but also difficult to solve. This is analogous to taking your car, which is indeed a fairly complex system, to the mechanic and describing an intermittent problem that is not currently acting up. There is some guesswork involved until the technician can get to the bottom of the problem.
What Boeing is experiencing is just as perplexing because the batteries didn’t cause a problem on the vast majority of flights, and it is difficult to replicate under test conditions.
So how do we reduce the risk of failure in complex systems? While we can’t completely eliminate failures, the answer lies in systems engineering. This involves a process of careful design and architecture of the system itself in terms of the subsystems, connections and components as well as a staged integration of the entire system, and extensive qualification, verification and validation testing.
At MIT, we offer an interdisciplinary graduate program in System Design and Management (SDM) for experienced engineering professionals who will lead the design and management of complex products, organizations and systems. The program is jointly offered by MIT School of Engineering and MIT Sloan School of Management. In the SDM program, we teach the principles of how to design and manage complex systems such as those in aerospace, transportation, health care, mechatronics and capital equipment industries.
Boeing is actually one of the best practitioners in the world of these principles. Indeed, the aerospace industry is a leader in systems engineering. Some of the best and brightest people in this field work at Boeing. Still, given the complexity of the 787 Dreamliner, it’s actually quite remarkable that there is only the one failure. That is evidence of just how good Boeing’s engineers are.
Boeing’s battery problem is a good reminder that designing a safe and reliable complex system is tremendously difficult. This challenge applies to all kinds of complex systems from automobiles to offshore oil rigs to aerospace systems. As we all remember, Toyota, BP and NASA have all recently experienced system failures.
This also partly explains why we haven’t seen many new automobile and airplane companies succeed in decades. It’s hard to pick up a capability for systems engineering when you start from scratch. That kind of capability takes a long time to develop through a combination of systems engineering training and learning from experience with the actual systems in the field.
Boeing will eventually get to the bottom of its 787 battery problem and fix it, just as Toyota got to the bottom of its accelerator pedal problem. The Dreamliners will be safely flying again and it will be the result of a deep, forensic systems engineering analysis and even more extensive testing.
Prof. Steven Eppinger is codirector of MIT’s System Design and Management Program. His research is focused on improving product design and complex system development practices. His recent book, Design Structure Matrix Methods and Applications, is published by MIT Press.