Statistics pattern-based big data analytics framework for IoT-enabled cybermanufacturing
Jin Wang1, Q. Peter He2, Anthony Skjellum3, Devarshi Shah1, Carlos Lemus3
1Department of Chemical Engineering, Auburn University, AL 36849
2Department of Chemical Engineering, Tuskegee University, AL 36088
3Department of Computer Science and Software Engineering, Auburn University, AL 36849
With the emergence of the Industrial Internet of Things (IoT) and ever advancing computing power and expansion of wireless networking technologies, a new generation of networked, information-based technologies, data analytics, and predictive modeling are providing unprecedented embedded computing capabilities as well as access to previously unimagined potential uses of data and information. These capabilities provide possibilities for new, radically better ways of doing manufacturing. Although there are different names used to describe next generation manufacturing systems, such as cybermanufacturing and smart/advanced manufacturing, the essence of these is the application of increasingly powerful and low-cost computation and networked information-based technologies in manufacturing enterprises. There is a general consensus that factories and plants connected to the Internet are more efficient, productive and smarter than their non-connected counterparts.
Manufacturing process operation databases are massive because of the use of process operation and control computers and information systems. With ever-accelerating advancement of IoT devices and other communication and sensing devices and technologies, it is expected that the data generated from cybermanufacturing systems will grow exponentially. 4 V's are often used to characterize the essence of big data: Volume (from terabytes (~1012) to zettabytes (~1021)), Variety (from structured to unstructured), Velocity (from batch to online streaming), and Veracity (from well calibrated and cleansed data to less trustworthy and uncleansed data).
Big Data is arguably a major focus for the next round of the transformation of advanced manufacturing. According to research by McKinsey Global Institute and McKinsey's Business Technology Office, the analysis of large data sets will become a key basis of competitiveness, productivity growth, and innovation.
In a newly funded NSF EAGER project, we propose a novel, statistics pattern-based process data analytics framework with the aim to provide smart diagnostics and prognostics for cybermanufacturing. As part of this effort, we also propose to establish an IoT-enabled manufacturing technology testbed (MTT) to explore and establish a proof-of-concept for the proposed framework. The theoretical background of the framework, the setup and test of the MTT, as well as the preliminary results obtained will be presented and discussed.