What Is Fault Tree Analysis?

Fault tree analysis (FTA) is a deductive reliability and safety engineering method in which an engineer begins with an identified undesired system event and works backward through the causal chain to find the combinations of component failures and external events that could produce it. The result is a tree-shaped diagram, the fault tree, whose root is the top-level failure and whose branches are connected by Boolean logic gates that represent the conditions under which failures at lower levels propagate upward to the root. FTA is used during system design and in post-incident investigations to quantify risk, identify single points of failure, and prioritize where redundancy or defect control measures offer the greatest improvement to system reliability.

The method was originally developed in 1962 at Bell Laboratories to analyze the Minuteman I missile launch control system, and it subsequently spread into nuclear power, aerospace, chemical process, and pharmaceutical engineering. Today FTA is formalized in IEC 61025 and is a required analysis technique in many safety-standard frameworks, including those governing aviation software and nuclear instrumentation.

Logic Gate Structure

A fault tree uses two primary gate types. An OR gate indicates that the output event occurs if any one of its input events occurs; this models situations where multiple independent failure paths lead to the same outcome. An AND gate indicates that the output event occurs only when all of its input events occur simultaneously; this models redundant designs where all replicas must fail before the system fails. Basic events at the leaves of the tree represent individual component failures, with associated probability or failure-rate data. Undeveloped events are leaf nodes whose causes have not been analyzed further, either because they are outside the analysis scope or because data is unavailable.

Qualitative analysis of the tree identifies minimal cut sets: the smallest combinations of basic events whose simultaneous occurrence is sufficient to cause the top event. A minimal cut set of size one represents a single point of failure with no redundancy. Identifying these sets directly guides design improvements and is the most immediately actionable output of an FTA study. The structure of cut sets and their connection to Boolean function reduction is described in a guide to fault tree analysis published by the University of Nottingham Resilience Engineering group.

Quantitative Assessment

When component failure rates and probabilities are available, FTA yields a quantitative estimate of the top-event probability over a specified mission time. Using the minimal cut set probabilities and the inclusion-exclusion principle, an analyst computes the system failure probability accounting for the correlation introduced by shared basic events across multiple cut sets. This quantitative output feeds directly into safety cases, regulatory submissions, and reliability growth programs. Software tools automate the Boolean reduction and probability computation for trees that contain thousands of nodes, a scale at which manual analysis is impractical. Fault tolerant computing architectures use FTA to verify that hardware and software redundancy measures have reduced the top-event probability below required safety targets.

The ScienceDirect overview of fault tree analysis methods and the IEEE Xplore catalog of fault tolerance publications both document the progression from manual qualitative FTA to fully automated, probabilistic digital tools used in high-consequence engineering programs.

Applications

Fault tree analysis has applications in a wide range of fields, including:

• Nuclear power plant probabilistic risk assessment and licensing submissions
• Aerospace system safety analysis under MIL-STD-882 and ARP4761
• Chemical and petrochemical process hazard analysis
• Medical device reliability assessment under IEC 62304 and FDA guidance
• Fault tolerant computing system design, verifying redundancy against identified cut sets