The most notable of these is the patented Watchdog Agent Toolbox. Dr. Mohamed AbuAli of the Center explains, "We wanted to use a more non-intrusive approach and find a way to do this without having to put sensors on the machines. Take a power-based system. We take power readings to determine when the machine begins to degrade, based on power usage. It's starting to use more power. Why? We're using power usage over time to infer when the machine will begin to degrade and may need maintenance."
The toolbox contains 20 algorithms in four categories: signal processing and feature extraction; performance prediction; health assessment; and health diagnosis.
Although most of the work so far has been done with rotary equipment, AbuAli says that the algorithms would be effective in the process industries as well.
The Center has partnered with National Instruments to commercialize the toolkit. The algorithms will be linked to LabView. More information on the toolkit and a downloadable evaluation version for LabView is available at http://sine.ni.com/nips/cds/view/p/lang/en/nid/210191.
