Fault Tree Analysis (FTA)
Also known as: FTA, Fault Tree Method, Top-Down Reliability Analysis, Hata Ağacı Analizi
Fault Tree Analysis (FTA) is a top-down, deductive reliability method that begins with an undesired top-level failure event and systematically traces backward through chains of contributing causes using Boolean logic gates (AND, OR). First formalized by Watson at Bell Telephone Laboratories in 1961 and later standardized by Vesely, Goldberg, Roberts, and Haasl in the landmark 1981 NRC Fault Tree Handbook, FTA has become a cornerstone of quantitative risk assessment in nuclear, aerospace, and industrial safety engineering.
Read the full method
Sign in with a free account to read this section.
Method map
The neighbourhood of related methods — select a node to explore.
+33 more
When to use it
FTA is appropriate when the system is well-defined and component failure data are available or estimable. It suits safety-critical applications—nuclear plants, aircraft, chemical processes—where exhaustive failure pathway identification is mandatory. Assumptions include component independence (unless common-cause extensions are added), binary failure states, and static system logic. FTA is ill-suited for highly dynamic systems with time-dependent dependencies; in such cases, dynamic FTA, Markov models, or Petri nets are preferred alternatives.
Strengths & limitations
- Provides a graphical, auditable map of all logical failure pathways to the top event.
- Yields quantitative failure probabilities and importance rankings from basic event data.
- Systematic identification of minimal cut sets highlights single-point failures and weak spots.
- Standardized notation and well-established algorithms enable tool automation and peer review.
- Assumes binary (failed/working) component states; partial degradation requires extensions.
- Standard FTA is static and cannot model sequence-dependent or time-varying failures without modifications.
- Tree construction quality depends heavily on analyst expertise and completeness of failure mode identification.
- Combinatorial explosion of cut sets can make analysis computationally intensive for large, highly interconnected systems.
Frequently asked
What is the difference between a cut set and a minimal cut set?
A cut set is any collection of basic events that, if all occur simultaneously, causes the top event. A minimal cut set (MCS) is a cut set from which no basic event can be removed while still causing the top event. MCS are preferred in practice because they pinpoint the most critical and economical failure combinations, avoiding redundancy in the analysis.
Can FTA handle common-cause failures?
Standard FTA assumes component independence, which can underestimate risk when multiple components share a common failure cause (e.g., same power supply, flood exposure). Common-cause failures are incorporated through extensions such as the beta-factor model or explicit common-cause basic events added to the tree, allowing their contribution to overall system failure probability to be quantified.
How does FTA relate to Event Tree Analysis (ETA)?
FTA and ETA are complementary. FTA traces backward from a single top event to its causes (deductive), while ETA traces forward from an initiating event through system responses to final outcomes (inductive). In integrated probabilistic risk assessment, FTA results are used to supply failure probabilities for the branch probabilities within event trees, linking the two methods into a unified risk model.
Sources
- Vesely, W. E., Goldberg, F. F., Roberts, N. H., & Haasl, D. F. (1981). Fault Tree Handbook (NUREG-0492). U.S. Nuclear Regulatory Commission. link ↗
How to cite this page
ScholarGate. (2026, June 2). Fault Tree Analysis (FTA). ScholarGate. https://scholargate.app/en/reliability/fault-tree-analysis
Which method?
Set this method beside its closest kin and read them side by side — the library lays the books on the table; the choice is yours.
- Bayesian NetworkBayesian↔ compare
- Event Tree AnalysisReliability↔ compare
- Reliability AnalysisReliability↔ compare