System reliability tells how likely a system is to work given the reliability of its parts and the way they are arranged. The system reliability calculator takes the reliability of three components and of a k-out-of-n group, and gives the reliability of a series system, a parallel system, a series system with a duplicated component, and a k-out-of-n system, plus the failures per 1,000 missions.
How the calculation works
- Each component has a reliability: the probability that it works for the mission. Enter it as a percentage.
- In a series system all components must work, so the reliability is the product: R = R₁ × R₂ × R₃. It is never higher than the weakest component.
- In a parallel system the system works if at least one component works: R = 1 − (1 − R₁) × (1 − R₂) × (1 − R₃). Duplicating a component in parallel with itself gives 1 − (1 − R₁)².
- In a k-out-of-n system of identical components, at least k of n must work: R = P(X ≥ k) from the binomial distribution with n trials and probability R of each.
Example
Three components in series have reliabilities of 95 %, 90 % and 98 %, so the system works 83.79 % of the time, or about 162 failures per 1,000 missions. In parallel they would reach 99.99 %. Duplicating component 1 in parallel lifts the series system to 87.98 %. A system of 5 identical components of 90 % reliability that needs 3 to work reaches 99.14 %.
Why it matters
Designers need to know how much each choice of arrangement buys, for example whether a backup pump is worth its cost. The calculator makes the effect of series, parallel and voting arrangements concrete before the system is built.
Good to know
The calculation assumes that components fail independently and that the reliability is the same for the whole mission. Common-cause failures, such as a shared power supply, a fire or a software bug, reduce the real reliability of redundant systems below the figure given. Repairable systems are better described by availability.
Related tools
The MTBF and availability calculator links failure rates and repair times, the Weibull distribution calculator models how reliability changes with time and the binomial calculator gives the probabilities behind the k-out-of-n result.