The analysis of a large quantity of experimental data from a thorough test of a complex system can be time-consuming and expensive. This paper describes how expert system technology can be applied to automate the detection of anomalies in a standard test of a torpedo. The purpose of the FLEET/DRAS (Data Reduction Analysis System) project is to develop an expert system to aid a Navy technician in the quick-look analysis of data from torpedo sea trials. Each sea trial is expected to generate about 50 megabytes of data. Many sea trials are expected. Anomaly detection, which has been performed by an expert engineering analyst, is to be automated. A prototype expert system for the automatic detection of anomalies in the control system has been developed by Honeywell. It follows the rules of the expert analyst. It detected all of the anomalies that it was designed to discover. It also discovered two subtle anomalies that had previously gone undetected. The feasibility of the system has been demonstrated.
An expert system for data reduction
Ein Expertensystem fuer Datenreduktion
1985
5 Seiten, 5 Bilder, 2 Tabellen, 5 Quellen
Conference paper
English
Expert system technology as a data processing tool
NTRS | 1988
|Tema Archive | 1987
|British Library Conference Proceedings | 1999
|Expert system data bases in flight test engineering
AIAA | 1995
|Locomotive data expert diagnosis analysis method, device and system
European Patent Office | 2021
|