Aerospace Bets on Big Data
Dec 11, 2015
by Chitra Sethi Managing Editor, ASME.org
Big Data is getting bigger and so are its wide-ranging benefits. From IBM’s Watson analyzing Big Data to help shoppers identify the trendiest holiday gifts to Boeing predicting part failures in airplanes.
The aerospace industry is capitalizing on the enormous amount of data, transmitted via sensors embedded on airplanes, by using it to improve business processes. Sensors are generating so much data that airlines often face the challenge of interpreting it all.
“Big Data by itself doesn’t produce any value,” says Andrew S. Bicos, PhD, director, platform performance and systems engineering and chief engineer, aeromechanics at Boeing Research & Technology, the Boeing Company. “The value really comes from being able to look at the data through algorithms, machine learning, and data mining that help you pull out the information and then analyze the results and transform that data into right decisions,” he adds.
ASME.org spoke with Dr. Bicos at ASME’s Industry Advisory Board Fall 2015 meeting, where Board members discussed the wide-ranging impact and opportunities that the nexus of manufacturing, design, and Big Data presents to the field of engineering. His responses below highlight the impact of Big Data on the aerospace industry and how it will influence the future of flying.
Q. How is your team using Big Data to improve the design, construction, and operation of an aircraft?
Bicos: Using the correct analytics can help us design more efficient airplanes as well as to operate and maintain them more efficiently. For example, if you have a plane in flight, Big Data can allow you to fly it more efficiently by adjusting the fuel consumption rates for the particular flying conditions that it’s going through. You can also adjust the flight path to maximize efficiency as well as reduce turbulence. Typically there are thousands of flight parameters that are being recorded and transmitted during the flight. This data can be used to diagnose potential issues before they occur. If there are issues with any of the systems or structure, you can collate that data with flight operations and do prognostic/predictive analysis to take the appropriate action before it happens.
Q. What data points are important for the aerospace industry?
Bicos: It really depends on the point of the lifecycle of the product and what data you are looking for. In production, the quality of the parts can be monitored: Where the parts are in the supply chain, where they are in assembly line position, and what’s the degree of completeness versus the planned one? These data points can help make decisions in the process of putting a product together. In flight, it could be things like the fuel consumption rate, stress levels, temperature, and power usage. By monitoring those things we can make sure everything is working as it should and if we detect any anomalous behavior, then we can use that to predict what might happen. For example, if the part needs to be replaced when it lands, the system can send a message and we can have the part waiting and replaced, thereby minimizing the turnaround time. On our customer side, we can use Big Data to help them optimize maintenance programs, pilot and crew scheduling to improve their overall operational efficiency. This all ties back into design, as we can use the information of the existing fleet to help design the next generation of airplanes more efficiently and be more cost-effective as well.
Q. What is one of the biggest challenges of Big Data?
Bicos: It’s the sheer amount of data and how do you get information out of it in time to make meaningful decisions, so that we take the right actions. Whether in production systems or flights, majority of value comes from minority of data. As the size of the data increases exponentially, the amount of data that’s not valuable also becomes significant. For example, 787s generate terabytes of date per trip, and that’s a lot of data. There is also a challenge in handling sparse data because inferring the right decisions from regions of data that have few observations is an issue. Both extremes are challenging.
Q. Does the aerospace industry have the right talent capable of interpreting the data to find meaningful business insights? How can engineering students prepare for data-related careers?
Bicos: We do have the right folks to move us in the right direction. We also need to rely on talent in the other industries and universities to keep us moving forward and accelerate us as the amount of data grows. As we go through applying these technologies across more and more production, operations, and maintenance systems, we do foresee a shortage of skilled data scientists. Some universities have started programs to tackle that. Whether they want to become design, manufacturing, or maintenance engineers, it’s important for all engineering students to really become comfortable with dealing with the sheer amounts of data and being able to understand patterns and trends of using the systems so that they can make better decisions.
Q. How will Big Data drive the aerospace industry going forward?
Bicos: Big Data is going to help enable improvements across all stages of the product lifecycle, from very early conceptualization, to design, production, after-market support as well as in-flight operations. As aerospace becomes more automated, we need to harness Big Data and do it more efficiently and more of it in real time. As our products and production systems become more complex, we need to become more efficient and cost-effective in dealing with all of this data. Future will be dominated with modeling and simulation and how we apply the sheer volumes of the real system data to validate the simulations will become important. What we would like to get to is to be able to build a product inside the computer and validate it using Big Data before we build actual hardware. I think in short, Big Data is going to a better enabler to design of our products, more efficient manufacturing and operations, and better business decisions and internal operations. It will be about getting the right information to the right decision maker at the right time.
Additive manufacturing is transforming every aspect of product development from product design and manufacturing to quality control. How can you prepare for this technology to impact your business? Learn more at ASME’s AM3D Conference & Expo
Big Data is going to help enable improvements across all stages of the product lifecycle, from very early conceptualization, to design, production, after-market support as well as in-flight operations.Dr. Andrew S. Bicos, Boeing Research & Technology