Figure 10.6: Illustration of the results of the risk analysis of the power supply system.
The Bayesian probabilistic nets have several advantages, of which one of them is the ability to
provide diagnostics support. Providing evidence to the Bayesian net that the power supply has
failed may attain this support. By conditioning the state of power supply on Fail back
propagation may be made – explaining away – this information in the net and assess the most
probable cause of failure. The result is illustrated in Figure 10.7.
Figure 10.7: Result of diagnostics assessment given that the power supply has failed.
From Figure 10.7 it is seen that the most probable cause of power supply failure is failure of
the power distribution cables. The fuel support system is redundant.
Finally also the possibility of taking into account common cause failure is illustrated by
introducing a new variable into the network, namely the event of an earthquake. The
earthquake has the influences on the Bayesian network as illustrated in Figure 10.8.