Discipline: Mathematics
A system may be represented by an interconnection of blocks. When a fault occurs, the block where the fault originates is the root cause and the other blocks that manifest failure because of the root cause are reactive causes or propagators of a failure event.
Fault tree diagnostics typically operate based on a failure model of cause and effect relationships. By supplementing it with a system success model, comprehensive premises are made available to infer the root cause distinctively from reactive causes. To validate the conclusion, necessary or essential symptoms must be reproduced for user validation and unnecessary symptoms must be discounted or filtered out.
To illustrate the fuzzy algorithm mechanism, a simple cascaded system is selected and a single fault occurrence is assumed. The relational matrix has been designed to accommodate information from both failure and success models. A single stage fuzzy composition was found to be good enough to infer the root cause in a simple cascaded system where a combination of root causes is absent. The inferred root cause is reused as a premise to infer all essential symptoms.
A new composition has been made that effectively infers essential symptoms. It is a modification of Alpha and Epsilon compositions and it is called Alpha1. It requires that its input fuzzy value must be in the interval [0,1].
ISSN 1656-1996 (Print)
Philippine E-Journals
