NASA 1996 SBIR Phase I

PROJECT TITLE : Real-Time Propulsion Diagnostic System Designer

TECHNICAL ABSTRACT (LIMIT 200 WORDS)

The Real-Time Propulsion Diagnostic System Designer is a powerful operations support tool designed to reduce operations costs, expedite test and launch processing, and increase the safety, reliability and availability of NASA ground test and space transportation systems. Over the last ten years, NASA and its contractors have developed advanced algorithms and methods targeted for improved launch vehicle propulsion system health management. Lacking in this research has been the substantial development of an integrated top-level framework for automated diagnostic assessment of features and events observed in the data or detected by these health management algorithms and tools. We solve several significant research problems enabling real-time capable automated diagnostic assessment of NASA flight systems through the development of an innovative set of software tools which: 1) enable rapid, intuitive computer-aided modeling of an expert's diagnostic knowledge using Bayesian belief networks; 2) automate methods to evaluate and optimize the user's diagnostic belief network; 3) automatically compile the user's belief network into fast, compact code for real-time embedded decision processing. The technical approach builds on our family of established system health management tools which automate crucial steps in the diagnostic process, including real-time sensor data validation and automated real-time event detection.

The Diagnostic Designer enables automated real-time diagnostic support for many high-value and safety critical process control applications. The U.S. Bureau of the Census reports an annual domestic market for 'special purpose industrial controls' of $2.1B. Lacking in the market is an integrated set of tools enabling operations personnel to develop and maintain automated, event driven machinery and process diagnostic systems. Every process control system is vulnerable to the deleterious effect of sensor failures and other process threatening events. For all but the most safety critical or highest value systems, automated diagnostic monitoring has been previously unaffordable. Our product lowers the affordability point to open an untapped niche market within the process industry. Our method enables plant operations engineers to design, develop, deploy and maintain sophisticated diagnostic monitoring systems without the requirement for a large staff of software development professionals.

Randall L. Bickford
Expert Microsystems, Inc.
7932 Country Trail Drive, Suite 1
Orangevale, CA 95662-2120

Expert Microsystems, Inc.
7932 Country Trail Drive, Suite 1
Orangevale, CA 95662-2120