Advanced Analysis of Sensory Data Yields Focused Fault Indications for Industrial Machinery, Electrical Power Systems and Physical Infrastructure - Reducing data streams to information to aid maintenance planning

摘要

Monitoring machines, electrical power systems, and infrastructures for reliability and failure indications is an increasingly common practice in today's industrial operations. However, monitoring sensory data without sorting, filtering, prioritizing, analyzing and correlating the data only produces more data than the human expert is able to utilize. To address this problem, embedded processing capabilities are now available for data analysis and reduction at the data collection site. This paper presents a methodology for sensory data selection, data sorting techniques to group similar data sets, and a review of metrics useful in determining operational health of machines and structures. The paper concludes with graphical communications for management with the intent to optimize maintenance activities.