Statistics and Probabilities – The heart of RAM analysis
Contribution from my colleague Varun Mittal:
There are two major applications of probability theories and statistics in everyday life, one being risk assessment and other in-trade on commodity markets. Risk assessment in our context refers to RAM analysis – the ability to manage and mitigate risk from plant downtime in order to maximize production and revenue.
What do we mean by statistics and probabilities?
Statistics: A statistic is a measurement from a pool of data or sample, which we use to estimate a chosen variable.
Probabilities: We like to think of probabilities as a chance or likelihood of getting a particular outcome.
The notion that the future is unpredictable is undercut every hour by the ease with which the past is explained. We make educated decisions everyday based on the statistics from previous incidents. RAM analysis exploits the very same idea – using historical data to predict future/life time events, as well incorporating of Monte Carlo Simulation where a large amount of random simulations are used to find an unknown probabilistic entity to which analytical solutions are not available. (A fantastic short video explaining the theory http://www.youtube.com/watch?v=Xaymy3Blnq4)
Another significant application of probability theory in everyday life is reliability. Most of the plant operators, whether upstream or downstream in the production line, need to utilize reliability theory in the design of their plant to reduce the probability of failure. Assigning probabilities to an event is a very risky task. We often get these wrong and are prone to blame decision makers for good decisions that worked out badly and give them little credit for successful moves that appear obvious only after the fact.
In our RAM tools, however, we can create catalogues reliability data from various sources: AEA, EIReDA, RAC, Cox, Lees as well as generic and in-house reliability data. The use of such records provides the much needed support and comfort in our decisions to predict the likelihood of equipment downtime and its effect on other dependent equipment downstream, so that the operators can maintain plant efficiency and suitable reliability levels. Maros and Taros also offer criticality assessment, so that the user can identify critical items in the system and take corrective measures to ensure adequate availability of equipment to reduce production loss.