|
What - Fault Tree models take low level
failures and propagate them out towards system-level and observable behavior.
They can include complex events as well as actual physical failures, and often
use both AND and OR logic symbols. Each node in a fault tree can be assigned a
failure probability so that the probability of the higest level effect can be
computed.
Why - Fault Trees are commonly used for
safety analysis in large systems such as nuclear reactors, space vehicles,
processing facilities, etc. They are useful because they can be easily created
in the conceptual phase in order to build in some safety considerations from
the beginning.
When - When performed in the concept phase,
Fault Trees are largely based on similarity analysis, or as a mockup. Often,
the failures are flipped into losses of functions for lack of failure mode data
at the early phases. When performed later, they tend to be analyses based on
actual designs and implementation plans, as a proof of the existance of safety
factors and remediation capabilities
When - Fault Trees first began appearing in
the 1960's, coincidentally with classic Dependency Modeling. Over the years,
they have grown in their applications to include complex logic, hierarchy and
sophisticated probability calculations.
Example:
Related Links -
Reliability Block Diagrams
|