What RAM Analysis Actually Measures

RAM stands for reliability, availability, and maintainability. Reliability is the probability that equipment keeps running without failure over a given time. Availability is the fraction of time it is actually able to produce, accounting for both failure and repair. Maintainability is how quickly and predictably it can be brought back after a failure. On their own, these three numbers describe a single piece of equipment. The step most plants skip is connecting them to the process flow, so the analysis says something about what the facility as a whole can deliver, not just how often one pump breaks.

That connection matters because facilities rarely fail outright, they degrade. A three-train compressor loses one train and keeps running at two-thirds rate. Production availability captures this by modeling the plant as a flow network, spare capacity and export logistics included, and reporting actual production divided by potential production over the asset's life. Equipment-level RAM, one failure rate per component with no model of how that failure moves through the process, cannot show this at all.

How the Same Failure Can Mean Two Different Things

A simple example makes the point concrete. Two export pumps, each rated at half the platform's 100 bbl/day design capacity, run redundant against each other. Early in field life, one pump failing cuts export to 50 bbl/day, a real production loss. Later, once natural decline has already brought demand down to 50 bbl/day, the same failure costs nothing, because the remaining pump covers it. Identical equipment, identical failure rate, opposite production impact, depending only on where the field sits on its decline curve. A maintenance report that rolls up MTBF by equipment class has no way to tell these two failures apart, even though one of them matters and the other does not.

Working Through the Redundancy Math

Take a unit with 95 percent availability and compare three setups built to meet full demand.

• 1 x 100 percent, no spare. System availability equals the unit's own: 0.95.
• 2 x 50 percent, both required for full output. Both up: 0.95 x 0.95 = 0.9025, worse than a single unit. At least one up, covering half output: 1 - (0.05 x 0.05) = 0.9975.
• 1 x 100 percent plus a full-rated standby (N+1). At least one up for full output: 0.9975, same math as above, now applied to the whole duty.

Splitting a unit into 2 x 50 percent does not raise the odds of hitting full capacity. It behaves like two units in series. What it buys is a much better chance of partial output, useful once partial output is enough. Real availability gain for full capacity only comes from a genuine spare, not from splitting the rating. Common cause failures (shared power, shared cooling), standby units that fail to start on demand, and platform weight limits all sit outside this calculation and still decide which configuration gets built.

Why International Reference Data Matters

Indonesian operators rarely build this analysis from local failure data alone, there usually is not enough of it. The international reference most RAM studies lean on is OREDA, the Offshore Reliability Data project, running since the early 1980s and now in its sixth published handbook edition, covering topside and subsea equipment reliability across a pool of international operators. ISO 20815, Petroleum, Petrochemical and Natural Gas Industries, Production Assurance and Reliability Management, sets the lifecycle framework these studies are expected to follow, from early design through operations. A system-level RAM model built on OREDA failure rates and run against ISO 20815's lifecycle structure holds up to scrutiny in a way a one-off equipment vendor number does not.

Starting Point for Teams New to This

None of this requires new instrumentation. It requires taking the failure data a plant already collects and modeling it against the facility's flow network and current demand, instead of reporting it component by component. That reframing is usually the hardest part for teams used to equipment-level KPIs, more a change in how the analysis is set up than in what data goes into it.

If your team is building this capability, or wants a second opinion on a RAM or production availability model before it goes into a project sanction package, Cliste Rekayasa Indonesia works on this with operators beyond just running the numbers, including how to structure the study, read the results, and explain them to the people approving the budget. Reach out for a consultation or training session.

Contact us now!

Author: Jen Megah Bremanda S. (Reliability Engineer)

References

• SKK Migas Targets Eight Upstream Oil and Gas Projects to Go Onstream in 2026, Indonesia Business Post, 2026
• Production Availability and Capacities, DNV, 2024
• Standards and Publications, OREDA, 2015

Clarification Note: This article was developed with AI assistance for drafting, research support, and language refinement. The topic framework, technical direction, and key points presented were determined by the author.