A Structural Algorithm for Finding Testable Sub-models and Multiple Fault Isolability Analysis
Structural methods have previously been used to perform isolability
analysis and finding testable sub-models, so called Minimal
Structurally Overdetermined (MSO) sets, Analytical Redundancy
Relations (ARR), or Possible Conflicts (PC). The number of MSO sets
grows exponentially in the degree of redundancy making the task of
computing MSO sets intractable for systems with high degree of
redundancy. This paper describes an efficient graph-theoretical
algorithm for computing a similar, but smaller, set of testable
submodels called Test Equation Supports (TES). A key difference,
compared to an MSO based approach, is that the influence of faults is
taken into account and the resulting number of testable models as well
as the computational complexity of finding them can be reduced
significantly without reducing the possible diagnosis performance. It
is shown that the TESs in a direct way characterize the complete
multiple fault isolability property of a model and thus extends
previous structural approaches from the single-fault case.
Mattias Krysander, Jan Åslund and Erik Frisk
21st International Workshop on Principles of Diagnosis (DX-10),
2010
Informationsansvarig: webmaster
Senast uppdaterad: 2021-11-10
LiU Homepage
