Fault tree analysis (FTA)

Fault tree analysis (FTA) is a method by which scientists can plot and examine the likely causes of a problem. This analysis produces a chart showing an undesired event, such as a product failure, which is then connected to one or more likely immediate and basic causes. Fault tree analysis was developed in the early 1960s and quickly became a staple of industrial science. Many organizations use fault tree analysis to evaluate their products and procedures and increase safety by attempting to eliminate problems before they occur. This analysis can be used to create safer and more efficient products, structures, vehicles, and processes of all types.

Brief History

The fault tree analysis system began in the early 1960s during a US Air Force study into the Minuteman launch control system, a new missile technology. Due to the complexity of the launch system and the great power of the technology, scientists searched for new ways to find possible flaws in the system and avoid malfunctions or other potentially disastrous failures. H.A. Watson of Bell Telephone Laboratories developed the basic form of fault tree analysis for this project, and his plan was approved by the Air Force. Soon, other scientists, including Dave Haasl of the Boeing Company, adopted fault tree analysis procedures for other complex and risky tasks such as aircraft design.

By 1965, the idea of fault tree analysis had been adopted by major organizations and its popularity was spreading in science and industry. Traditionally, experts created new technology and planned complicated procedures from start to finish, hoping that sound plans would make failure virtually impossible. Now, facing science more advanced—and potentially more deadly—than ever, many experts favored the new approach of identifying possible failures beforehand and tracing their possible roots to help eliminate the possibility of problems developing. This was an early step in a new increased focus on safety procedures in science and industry.

Scientists expanded the basic fault tree analysis system created by Watson to increase its accuracy and improve its applicability to many fields. The advent of modern computing technology helped scientists create, alter, study, and compare fault tree charts with greater ease, as well as to calculate new equations to assess the likelihood of various failures. By the 1970s, the nuclear power industry adopted fault tree analysis programs to help avoid nuclear catastrophes. Automobile, railroad, and chemical companies quickly followed, making fault tree analysis an indispensable part of industrial science around the world.

In the twenty-first century, fault tree analysis remains a common feature in many sciences and industries. Computers, robots, aircraft, trains, automobiles, power facilities, factories, weapons, and much more are designed with the aid of fault tree analytical procedures.

Overview

Fault tree analysis is a relatively simple, inexpensive, and effective way to identify problems before they occur. (It may also be used to determine the cause of problems that have already occurred, as in accident investigation.) This form of analysis is crucial for making products and procedures in many fields safer. Fault trees can show where products or procedures are weak, where they may be improved, how human error may influence an event, and how failures may be avoided.

Traditionally, scientists mostly employed inductive modeling, which involves the forward motion of an event to its consequences. Fault tree analysis does the opposite. It uses deductive modeling, which examines the backward motion from an event back to its causes. Here, a scientist identifies an event and then attempts to determine what might have caused that event. This process is aided by several mathematical concepts, including probability theory (the study of randomly distributed quantities) and Boolean algebra (the study of true or false variables).

The events examined in fault tree analysis are failures and other undesired events. These events may be as small as everyday inconveniences (such as a faulty lightbulb) or as large as deadly catastrophes (such as explosions at nuclear power plants). In either case, scientists examine these undesired events and connect them with their most likely immediate causes. From there, scientists may expand their study further back to determine even more basic causes.

For example, a scientist may identify a potential failure as "the rocket will run out of fuel before reaching its destination." One immediate cause of that failure may be that not enough fuel has been stored on the rocket. A more basic cause of that may be that the rocket was not designed to be powerful enough to carry the necessary amount of fuel. There may be many potential causes, both immediate and basic, for a failure, all of which may be explored via the fault tree analysis procedure.

All the information gathered about events and causes is plotted on a diagram that shows the relationships between these various factors. Although the topics being investigated may be complex, the diagram was intended to display this information in a relatively simple structure. The fault tree chart resembles a flow chart or organizational chart, with a variety of shapes (representing events) connected by lines (showing causality and other relationships).

The top shape is a large rectangle in which the undesired event being studied is written. Beneath that is a shape called a logic gate. Logic gates help to organize the information by showing how the various factors are related. For example, some logic gates show that factors may work alone, while others show that factors must work together to produce the given undesired event. Below the logic gates are rectangles showing immediate causes of the undesired event. These may be connected to other logic gates that are in turn connected to circles that identify the basic causes of the undesired event. Together, this information effectively provides a visual description of the potential problem and its foreseeable causes.

Bibliography

Ericson, Clifton A. "Fault Tree Analysis—A History," Proceedings of the 17th International System Safety Conference, 1999, web.archive.org/web/20110723124816/http://www.fault-tree.net/papers/ericson-fta-history.pdf. Accessed 3 Jan. 2017.

"Fault Tree," Smartdraw, www.smartdraw.com/fault-tree/. Accessed 3 Jan. 2017.

"Fault Tree Analysis." Weibull.com, www.weibull.com/basics/fault-tree/. Accessed 3 Jan. 2017.

Limnios, Nikolaos. Fault Trees. Wiley-ISTE, 2013.

Marshall, Jane. "An Introduction to Fault Tree Analysis (FTA)." University of Warwick, www2.warwick.ac.uk/fac/sci/wmg/ftmsc/modules/modulelist/peuss/slides/section‗11b‗fta‗lecture‗slides‗compatibility‗mode.pdf. Accessed 3 Jan. 2017.

Pilot, Simha. "What Is a Fault Tree Analysis?" Quality Progress, Mar. 2002, asq.org/quality-progress/2002/03/problem-solving/what-is-a-fault-tree-analysis.html. Accessed 3 Jan. 2017.

Stamatelatos, Michael, et al. "Fault Tree Handbook with Aerospace Applications." NASA Office of Safety and Mission Assurance, Aug. 2002, www.hq.nasa.gov/office/codeq/doctree/fthb.pdf. Accessed 3 Jan. 2017.

Vesely, Bill. "Fault Tree Analysis (FTA): Concepts and Applications," NASA, www.hq.nasa.gov/office/codeq/risk/docs/ftacourse.pdf. Accessed 3 Jan. 2017.