Nowlan & Heap 1978: the report that killed "overhaul everything."

Why this report still matters

In the late 1960s, civil aviation had a cost problem. The widebodies were coming - far more systems, far more components - and under the reigning doctrine ("everything has a safe life; overhaul it before that life is up") their maintenance programs would have been ruinously expensive. So the airlines and the FAA formed Maintenance Steering Groups to build the new programs from decision logic instead of habit: MSG-1 produced the Boeing 747's program in 1968, and MSG-2 generalized the approach to other types in 1970.

What the steering groups found in their own overhaul records was the real story: for most items, fixed-interval overhaul had no measurable effect on safety or reliability. When the Department of Defense wanted the definitive account of this new discipline, it went to two United Airlines engineers who had lived it. F. Stanley Nowlan and Howard F. Heap delivered in 1978 - and gave the discipline its name: Reliability-Centered Maintenance.

The numbers were brutal for the overhaul-everything doctrine. Only about 11% of failure modes (patterns A to C) show a wear-out zone where a fixed age limit can help. Around 89% (patterns D to F) show no wear-out zone at all - there is no age at which replacement makes them more reliable. And the largest single group, 68%, follows pattern F: high infant mortality followed by a constant failure rate. We keep the six-patterns chart on one page only - see the figure in our RCM II summary, where Moubray popularized it for industry.

Sit with what pattern F means for a plant. Every "precautionary" teardown restarts the riskiest part of a component's life. The conveyor gearbox you rebuild every two years "to be safe" comes back with fresh seals that can leak, bearings that can be damaged on installation, and contamination that was not there before. The 1978 data said it plainly: intrusive overhauls can increase failures. If your PM program is a wall of time-based rebuilds, you may be paying money to make reliability worse.

• On-condition tasks. Inspect for evidence of a potential failure and act on condition, not on the calendar. This is the seed of everything we now call condition monitoring, and of the P-F interval thinking that schedules it (see our P-F & D-I-P-F summary). Vibration routes on your compressors and oil analysis on your gearboxes are 1978 logic wearing 2026 sensors.
• Hard-time tasks, only where earned. Fixed-interval restoration or discard is reserved for items with a demonstrated wear-out zone - patterns A to C, proven with data, not assumed.
• Failure-finding tasks. Hidden functions - the standby pump, the trip, the relief valve - fail silently and only reveal themselves when needed. The answer is periodic functional tests to find those failures before the demand does.
• Applicability and effectiveness logic. Every task must pass two gates: it must be technically capable of addressing the failure mode (applicable), and worth doing given the consequences (effective). The decision diagram that walks each failure mode through these gates is the machinery that RCM II later industrialized.

The safety logic is just as important: safety consequences are separated from economic ones. Economic failures get a cost test. Safety-critical failures do not - if no effective task exists for one, the answer is not a bigger PM but a redesign. That single rule, written for aircraft, transfers to your plant's protective systems without changing a word.

Two branches are worth naming. MSG-3 (1980, maintained today by Airlines for America) is how the scheduled maintenance program of every current airliner is still developed - when you board a plane, you are trusting a direct descendant of this report. On the industrial side, Moubray's RCM II carried the method into mining, process and manufacturing plants, SAE JA1011 fixed the criteria for what may legitimately be called RCM, and IEC 60300-3-11 turned it into an international application guide (see our IEC 60300-3-11 summary).