OptimalAvailability Studio™ Wiki / Reliability engineering
The frame
Reliability engineering
Reliability engineering asks four questions of an asset: what must it do, how can it fail to do it, how much does each failure matter, and how do we design and maintain so it keeps doing it. RCM, FMECA, RBD, Weibull and RAM are the methods that answer them.
The disciplines
| RCM | Reliability-centred maintenance. Decides the right maintenance task for each failure mode from its consequences, to SAE JA1011. |
|---|---|
| FMEA and FMECA | Failure mode and effects analysis, with criticality. Finds and ranks the ways an item fails and what each failure causes. |
| RBD | Reliability block diagram. Models how component reliabilities combine into system reliability through series, parallel and k-out-of-n logic. |
| RAM | Reliability, availability and maintainability modelling, usually by Monte Carlo simulation, to predict system and production availability. |
| Weibull and life-data | Reads the failure pattern and characteristic life out of failure history, which tells you whether a wear-out age even exists. |
| Maintainability | How quickly and surely an asset can be restored: the MTTR and repair-time distribution that sits behind availability. |
Where it sits: the design and proactive domains
On the DIPF curve, resistance to failure is set in the design domain, where around 80% of lifecycle cost is fixed, and preserved in the proactive domain by precision installation and defect elimination. Condition monitoring, the predictive domain, comes later and only catches what was left in. OptimalAvailability Studio™ works the first two domains: it designs reliability in and decides how to preserve it, then hands the on-condition tasks to RCM and to OptimalTREND™. See the modelling page for the maths and the DIPF curve in the OptimalTREND™ wiki for the full picture.
Inherent versus operational reliability
An asset has an inherent reliability fixed by its design: the best it can do if built, installed and operated perfectly. What it actually delivers, its operational reliability, is that inherent figure minus everything installation, operation and maintenance take away. Reliability engineering raises the inherent number where it can and, more often, closes the gap to it. You cannot maintain your way past the inherent reliability of a poor design, which is why the design domain has the most leverage.
The measures
| Reliability R(t) | The probability an item performs its function to time t without failure. |
|---|---|
| Availability A | The share of time an item is in a state to perform its function. Inherent, achieved and operational forms differ by what downtime they count. |
| Maintainability M(t) | The probability a failed item is restored within a given time. Summarised by MTTR. |
| MTBF and MTTF | Mean time between failures for repairable items, mean time to failure for non-repairable. The reliability yardstick. |
The formulae and the three availability definitions are on the modelling page.
Where OptimalAvailability Studio™ fits
OptimalAvailability Studio™ is the workbench for all of the above: one connected model in which functions, failure modes, tasks, block diagrams and life data live together, so a change in one propagates to the maintenance strategy and the predicted availability rather than sitting in a disconnected spreadsheet.