Prognostic Software Agents for Machinery Health Monitoring

1 IEEEAC paper #1311, Updated September 19, 2002 Abstract— Increasing levels of machinery automation for systems health monitoring are providing operators with larger amounts of raw data. However, transforming massive amounts of data into information useful for effective condition-based maintenance (CBM) remains an arduous task. New technology is needed to continually monitor machinery, to identify impending failures, and to accurately predict its remaining useful life. Prognostic software agents can satisfy this growing need as higher levels of machinery automation raise the cost requirements of continuous monitoring beyond the levels of human and company feasibility. Software agent technologies that can automatically perform useful work as human assistants and can readily be integrated into existing automation system environments, represent viable tools to improve machinery reliability and reduce maintenance costs.