Mining Hidden Assets:Making the Most of Big Data for MRO

by Charlotte Adams

BigData251Airplanes generate large amounts of data as they fly, but only a small part of that data is used in the maintenance process. And the historical maintenance information for legacy aircraft, some of it decades old, is also voluminous, varied and difficult to preserve, organize and search. This huge and ever-growing mass of information, aka “big data,” challenges the wherewithal of traditional computer systems. Yet these vast collections of information hold out tantalizing possibilities for those who can exploit them. Aviation, along with other industries, is trying to mine these lodes and extract more value from them, particularly in the area of predictive analytics.

Big data is data that is high in volume, velocity and variety, explains Svetlana Sicular, research director with Gartner Group, the company which reportedly invented the phrase. Opinions differ as to whether it includes structured data that fits into the rows and columns of a relational database, as well as unstructured data like engineering drawings, simulation files and engine telemetry. And experts differ as to the level of maturity of current tools and whether they are essentially prognostic or diagnostic in nature.

GE Aviation, which processes data for commercial and business aviation customers, further describes big data as “extremely large data sets that can be analyzed computationally to reveal patterns, trends and associations, especially relating to machine behavior and interactions.” Data analytics, by the same token, is “the process of examining large data sets containing a variety of data types, including big data, to uncover hidden patterns, unknown correlations, machine behavior and other useful business information,” according to the GE unit.

Among other characteristics frequently associated with big data and attempts to harness it are the use of the “industrial Internet” for the automated transfer of data from machines to computers using algorithms that preserve, sift, organize, search, analyze and visualize the data to provide insights and suggest solutions.

The challenges of big data range from dealing with floods of information in real time to making sense of streams of disparate data coming in at variable speeds. Service providers include many startups, such as Splunk, as well as giants such as General Electric (GE) and IBM, which have invested heavily in the field. Estimates of the market size for this growing area of computation vary widely, as it is difficult to measure. But IBM cites a forecast of $187 billion by 2015 across industries worldwide.

Although it’s become a popular phrase, big data has been around a long time, Sicular says. Financial institutions, for example, have been trying to sniff out fraud as long as they have existed. Big data is also relative, she points out. What’s valuable for one company might be worthless for another.

In the aviation industry engine manufacturers are known for their data intensity, but new technologies promise to help them exploit larger portions of these assets. The trend toward engine leasing also drives efforts to anticipate and control maintenance costs. Pratt & Whitney (P&W) has allied with IBM, which has invested some $24 billion in big data and analytics. GE also leverages both externally and internally developed technologies. Among the latter is the Predix platform for embedded analytics via the industrial Internet.

Peaxy, one of the many startups in the field of big data analytics, has gained the attention of the Gartner Group as a service provider with deep knowledge of industries rather than a purely entrepreneurial organization built around a new technology. Peaxy’s launch customer is GE oil and gas, which is a large a step—horizontally or vertically—towards aviation.

For Peaxy big data is essentially unstructured data. Although some components of big data may be relative, there are “crown jewel” data sets that are common across manufacturers, says the company’s CEO, Manuel Terranova. These data sets include geometry (such as engineering drawings), simulations (such as structural analyses), and telemetry, whether off the test bench or in the field, he says. The company’s Hyperfiler product helps turn big data into a “readily findable asset,” the company says.

[s2If is_user_logged_in()]
[s2If is_user_not_logged_in()]
[message_box title=”To READ the full article in our digital edition you need to be a registered subscriber to the magazine – If you are NOT then goto the SUBSCRIBE section of the site first before you can login” color=”green”]



Leave a Reply