How to conduct a 'Fault Tree Analysis'

Fault Tree Analysis is a graphical technique used to determine the various combinations of hardware (and software) failures and human errors, which can result in an undesirable outcome. The specified undesirable outcome is referred to as a ‘top event’, where the deductive analysis about the general conclusions and their causes is often described as a ‘top down’ approach.

A Fault Tree Analysis begins with a construction of a fault tree, relating the sequences of events leading to the top event. This may be illustrated by considering the probabilities of events and by constructing a tree with AND and OR logic gates. Basically, the steps involved in a fault tree analysis are:

  1. Define the Top Event
  2. Know the system
  3. Construct the tree
  4. Validate the tree
  5. Evaluate the tree
  6. Study tradeoffs
  7. Consider alternatives and recommend actions

A fault tree analysis can also include human error contribution to the overall system, if the probabilities for human error are described in the same term as component and hardware failures.

Thus the main purpose of fault tree analysis is to evaluate the probability of the top event using analytical and statistical methods. By providing useful information concerning the likelihood of a failure and its means, efforts can be made to improve system safety and reliability. It also evaluates the effectiveness and the need for redundancy.

Hence, the resulting benefits of fault tree analysis to project management are reduction of analysis time and precision in identifying and correcting deficiencies.

Other resources

Loading time: 30 (0) ms