Real-time Equipment Health Monitoring

Intelligent Automation, Inc.

Rockville, MD

Tools and techniques to allow real-time prognostics. The approach is based on principle component analysis to reduce the dimensionality of the raw data, and a simple type of neural net to perform diagnosis in the resulting reduced dimensionality space. Our tools and techniques are relatively easy to apply, and are able to adapt to widely varying operational modes.

• Experiments have been done using actual data from helicopter gearboxes, and from rocket engines. For helicopter gearboxes, we achieved 98% correct detection and diagnosis of incipient failures significantly before they would have caused a failure on an actual helicopter. For rocket engines, we achieved 97% correct detection and diagnosis of incipient failures before they would have caused an actual failure. Other experiments using simulation also showed exceptional ability to detect and diagnose latent failures in complex systems.

• IAI was also awarded a small contract from Ford Motor Company for a pilot installation of our techniques for health monitoring of critical equipment at the Visteon Rawsonville, MI plant. We were also awarded a larger contract from Motorola to apply the same tools and techniques to real-time health monitoring of the Call Processing Components of the Iridium Satellite-based Worldwide Telephone System. The resulting software was very successful, being used operationally almost immediately after completion of the prototype.

• Based largely on our previous success, IAI won an SBIR award from NASA KSC to apply these previous results to other NASA systems, and to integrate our prognostics technology with logistics systems. In that work we demonstrated the application of our techniques to real-time monitoring of actual space shuttle components including the Auxiliary Power Unit (APU) and space shuttle solenoids. For the APU, we detected and correctly diagnosed 97% of the incipient failures, and for the solenoids, we achieved 100% correct results for the data sets available. IAI and Boeing have teamed in the development of a proposal to KSC for the extension of the original work to integrate it into the Checkout and Launch Control System (CLCS) used for Shuttle and other KSC launches.

Return to NASA SBIR Success Listings