ABSTRACT
Fault tree analysis (FTA) begins with a known event (referred to as the top event) and describes possible combinations of events and conditions that can lead to this event. The top event in the fault tree can be the loss event under investigation or a specific event that is involved in the incident. The fault tree looks backward in time to describe the potential causes of the top event. AND and OR logic is used to graphically show potential combinations of events and conditions leading to the top event. It is commonly used proactively during risk assessments to identify dominant potential contributors. For incident investigation applications, however, the smallest possible tree is developed. As soon as a branch is shown not to be credible, the development of that branch is stopped. Most reactive and proactive analysis techniques only identify single-event failures. Examples of fault trees are provided below, as well as an explanation about the building blocks of such trees and a procedure for constructing a tree.
