Fault Tree Analysis
Fault Tree Analysis
FTA is a top-down risk analysis that is done by the aid of a logical diagram. A fault
tree diagram shows the potential failure of a system to identify the root causes of
that failure. The failure of the specified system is the undesirable event which can
also be called a Top Event while the causes of the top event can be named as Basic
Events. Since each system includes some components and sub-components, these
constitute the FTA basic events (Wang & Roush, 2000). As one can understand, the
relation between failure of a system, components, and the barriers are shown by
fault tree diagram. It should be noted that the basic event is not necessarily a
systems component failure; it could be a human error or another failure that has
occurred.
A fault tree diagram consists of graphical symbols that are representing the top and
basic events of the system in addition to their relations. These graphical symbols
are called logical gates, and although they are varying in different standards,
however, the most common, and significant ones are shown in Figure 3.3 along with
their interpretation.
Figure 3.3: Fault Tree Analysis Symbols (Aven, 2008)
To construct a fault tree diagram, the start point is the top event and then the
possible failures that can directly cause the top event shall be determined. These
failures must be linked to the top event via logical gates. Further, the work will be
continued respectively to reach the basic events. The ongoing process will be
stopped once one reaches the appropriate stage of details. Meaning that depending
on the severity of the analysis and organization policy, the basic events of fault tree
analysis for a nominated system might be varying. From one hand, the most
important point in constructing the fault tree analysis is that the FTA is a deductive
analysis. One shall repeatedly ask questions like What are the direct causes of this
failure? or How this event can happen?
On the other hand, the most common mistake in FTA happens when one tends to do
analysis and attain basic events too fast. It means that the development of fault
tree branches must be done consistently by considering all the sub-events of each
systems components. The fault tree analysis can be done in both qualitative and
quantitative ways as described in the following, which is based on (Rausand &
Hyland, 2004): Qualitative FTA For the qualitative fault tree analysis, there is a
need firstly to identify the cut sets and secondly the minimal cut sets. In fault tree, a
collection of basic events that their occurrence leads to the top event is called a cut
set. If the cut set cannot be decreased and still the top event will happen, this is the
minimal cut set. Once the minimal cut sets of the system are identified, one can
determine which combination of basic events may lead to system failure for doing
further risk treatment. The criticality of minimal cut sets is specified based on the
number of the events that they hold. This number of events is named the order of
the cut set. In terms of ranking, a minimal cut set with one event is more serious
than a cut set with two events since the system fails only by the occurrence of one
single failure. The shorter minimal cut sets are riskier and more undesirable ones.
However, one should not just rely on the number of events in the cut set since the
probability of the events occurrence is not considered. For instance, the shorter
minimal cut set might have a lower probability of happening, so there is a lower risk
of system failure. Thus, a pure qualitative FTA can be misleading, and a quantitative
FTA is needed to have a proper and reliable risk analysis. Quantitative FTA For a
quantitative fault tree analysis, the probability of basic events occurrence shall be
determined, which provides a basis for calculating: 1. The criticality of basic events
in the tree; 2. The top event probability of occurrence. For calculating the top
events probability of occurrence, there is a formula that will give a good
approximation. After determining each basic events probability of occurrence, the
probability of each cut set shall be calculated by summing these probabilities and
the top event occurrence probability will be gained.
Engine Start