Fault Tree Analysis

Description

A top-down Boolean logic diagram that decomposes an undesired outcome into combinations of basic events, revealing minimal cut sets and single points of failure.

Fault tree analysis starts from a precisely defined top event, such as 'user sends payment to wrong account', and asks what immediate events, combined through AND or OR gates, could cause it. Each branch is decomposed until reaching basic events: user errors, software faults, missing data, environmental conditions. Boolean reduction gives minimal cut sets, the smallest combinations of events sufficient to cause the top event; a cut set of size one is a single point of failure. With probabilities attached, the tree yields a quantitative top-event probability and importance measures. For UX it disciplines causal thinking: it distinguishes conditions that must co-occur from alternatives, and shows exactly where an interface safeguard breaks a chain.

Attributes

Loss type
Unmigrated
Effort
Days

Steps

  1. Define the top event precisely, including scope and system boundary.
  2. Identify immediate causes and connect them with AND or OR gates, distinguishing user actions, system states and external conditions.
  3. Decompose recursively until reaching basic events that can be observed or estimated.
  4. Compute minimal cut sets by Boolean reduction, by hand or with FTA software.
  5. Optionally assign probabilities from logs or HRA methods and compute top-event probability and Fussell-Vesely importance.
  6. Target design changes at single-event cut sets first and verify that each change adds an AND condition the user must satisfy.

Sources

  • Vesely et al., Fault Tree Handbook (NUREG-0492, 1981)
  • Ericson, Hazard Analysis Techniques for System Safety (2005)

Also known as

  • FTA
  • Fault tree

Categories

  • Accessibility
  • Human Factors & Error