Search Results

Real-time algorithms which provide the earliest possible indication of off-nominal Space Shuttle Main Engine (SSME) conditions could improve Shuttle safety and reliability by providing more time for corrective action. Multi-parameter fault detection techniques have been targeted because they do not rely on a single parameter for fault information and thereby improve confidence in the detection. Furthermore, no assumptions regarding failure modes are required, permitting the detection of previously unencountered or unanticipated failures. The Clustering Algorithm, a multi-parameter fault detection approach that was originally trained and validated on SSME ground test firing data, was slightly modified and applied to SSME historical flight data; the application is documented in this report. Preliminary studies were conducted to assess the impact of different engines, different missions and different thrust profiles on the performance of the Clustering Algorithm.

A health monitoring expert system software architecture has been developed to support condition-based health monitoring of rocket engines. Its first application is in the diagnosis decisions relating to the health of the high pressure oxidizer turbopump of the Space Shuttle Main Engine. The post test diagnostic system runs offline, using as input the data recorded from hundreds of sensors, each running typically at rates of 25, 50, or .1 Hz. The system is invoked after a test has been completed, and produces suggestions, analysis, and an organized graphical presentation of the data with important effects highlighted. The overall expert system architecture has been developed and documented so that expert modules analyzing other line replaceable units may easily be added. The architecture emphasizes modularity, reusability, and open system interfaces so that it may be used to the analyze other engines as well.