Systems Integration

GE at Leading Edge of Big Data Challenge

Lessons Learned as End User and Equipment Supplier Translate into Effective Solutions for Industry.

By Keith Larson

The era of the Industrial Internet and "big data" isn't coming. It's here. GE, for one, already leverages remote connectivity and analytics to support more than 200,000 connected devices around the world -- from gas and steam turbines and aircraft engines to medical equipment and locomotives. "We have to solve the big data problem every day," noted Jim Walsh, general manager of software and services for GE Intelligent Platforms at the company's recent Connected World North America event in Chicago.

As a manufacturing and infrastructure company that "also happens to do software," Walsh noted that GE is uniquely positioned to help industrial equipment suppliers and end users manage the challenge and opportunity presented by pervasive connectivity and big data. "GE has been dealing with tremendous amounts of data for a long time," Walsh said. "We generate data; we store data; we analyze data. We're well down the road to enabling a Connected World."

Also Read: An Internet for Industry

Walsh offered as evidence a mining company customer that used data-fed analytics to boost throughput of a milling circuit by 5%. Stabilizing unit feed rates also resulted in better grind quality, higher energy efficiency and improved asset life. Meanwhile, mine truck monitoring systems and predictive analytics at another customer are heading off engine, transmission and bearing issues before they lead to equipment failure. "They can finally begin to get at no unplanned downtime," Walsh said.

As a leading adopter of what it calls the Industrial Internet, GE can offer guidance to its customers in their journey, Walsh said. The most common problem is that customers don't make enough time up front for strategic prioritization, he said. They want to jump right to the analytics that will drive huge amounts of value -- but they have to get the foundation right first, Walsh said.

This foundation consists of first identifying what data is critical to the process. "You'll find some unexpected correlations," Walsh counseled, "but you have to use your domain expertise to decide what's likely to be important." Next, develop a strategy for gathering, efficiently storing and accessing that critical data. Desired outcomes such as throughput, efficiency and quality as well as asset life and equipment availability must be prioritized as well. "Only then can you begin to develop the visualization strategy to get the right information into the right hands at the right time," Walsh said.