Home > Research > Publications & Outputs > d-matrix based fault modeling for cryogenic loa...

Links

View graph of relations

d-matrix based fault modeling for cryogenic loading systems

Research output: Contribution to conference - Without ISBN/ISSN Conference paper

Published
  • Anu Kodali
  • Peter Robinson
  • Anna Patterson-Hine
  • Dmitry Luchinsky
  • Michael Khasin
  • Ekaterina Ponizhovskaya
  • Jose Perotti
  • Barbara Brown
  • Anupa Bajwa
Close
Publication date2015
Number of pages10
<mark>Original language</mark>English

Abstract

The study is motivated by NASA plans to develop technology for an autonomous cryogenic loading operation including online fault diagnostics as a part of Integrated Health Management system. For years, the diagnostic modeling effort is performed in many paradigms. None of these paradigms independently can provide a complete set of efficiency metrics: better diagnostics, lower run-time, etc. D-matrix, a causal 0-1 relationship between faults and tests, is proposed as a single representation between different model-based diagnostic methods for comparison and communication. This framework is suitable to create a common platform for communication via D-matrix for systems engineering process. The knowledge transfer between modeling techniques is done via D-matrix. In addition, D-matrix provides a common paradigm to compare the embedded knowledge and performance of heterogeneous diagnostic systems. D-matrix is generated from physics models to be used with faster run-time performance D-matrix based diagnostic algorithms. Additionally, we will also investigate if the derived D-matrix and thereby the physics model is sufficient and accurate for efficient diagnostics via iDME tool.