AS IEC 61078:2017 pdf free.Reliability block diagrams.
Warning: the calculations given above are valid only with asymptotic block availabilities. They are not valid with ordinary average availabilities.
Then, when the RBD does not reach a steady state, the average availability has to be calculated by using the general Formula (32).
A special case occurs when the RBD is used through recurring phases such as:
succession of seasons: winter, spring, summer and autumn;
test intervals for periodically tested items.
Figure 26 ilustrates a system with three recurring phases such that the same pattern of three phases is repeated with a time interval equal to ρ= T1+T2+Tg. The availability of such a system has no asymptotical value but the average availability Aavg[np,(n+ 1)p] generally reaches a limit value when n is large enough: As Aavg[np,(n+ 1)p] decreases when n increases, AIM(p) provides a good conservative approximation of the average availability over a time interval [0, 7] encompassing several sets of recurring phases.
Therefore, the techniques described in this standard can be used to calculate the average unavailability of safety systems and this makes the link with the functional safety standards (IEC 61508 or IEC 61511) which require such calculations for safety instrumented systems and where the average unavailability is called PFD  avg (average of the probability of failure on demand). This is described in Clauses C.4 and C.5.
10.3.3 Reliability calculations
When repaired blocks are considered, the calculation of the reliability Rs() implies that the repairs of the blocks in the system (RBD) need to be considered only as long as the system remains in the up state (see in 10.1).
This can be ilustrated by the simple redundant system modelled by the RBD on the left hand side of Figure 27. With regards to the calculation of Rg(t)， when the block B fails, it can be repaired only if S in the up state (i.e. if A is in the up state). In the same way when the block A fails, it can be repaired only if S in the up state (i.e. if B is in the up state). Therefore when one block fails, its repair depends on the state of the system S which, in turn, depends on the states of all the blocks. This systemic dependency between the blocks is modelled in the Markov graph presented on the right hand side of Figure 27. It is equivalent to the RBD presented on the left hand side.AS IEC 61078 pdf download.