| name | rcm-reliability |
| description | Reliability-Centered Maintenance (RCM) — failure mode and effects analysis (FMEA), Weibull analysis (β/η parameters, B10 life), failure consequence categorization (safety/operational/economic), maintenance task selection (condition monitoring, scheduled restoration, run-to-failure), P-F interval, MSG-3 aviation RCM, IEC 60300-3-11 RCM standard, FMECA severity/occurrence/detection ranking, preventive maintenance optimization, and SAE JA1011/1012 RCM criteria. |
| metadata | {"priority":7,"promptSignals":{"phrases":["reliability centered maintenance","RCM","FMEA reliability","Weibull analysis","maintenance optimization","failure mode effects"],"minScore":3}} |
Reliability-Centered Maintenance (RCM) — Complete Skill
RCM Fundamentals
Definition and Objectives
RCM (SAE JA1011): structured process for determining what maintenance must be done to ensure physical assets continue to fulfill their intended functions in their present operating context
Seven RCM questions (Nowlan-Heap):
- What are the functions and associated performance standards? (primary + secondary)
- In what ways can it fail to fulfill its functions? (functional failures)
- What causes each functional failure? (failure modes)
- What happens when each failure occurs? (failure effects)
- In what way does each failure matter? (failure consequences)
- What can be done to predict or prevent each failure? (proactive tasks)
- What should be done if a suitable proactive task cannot be found? (default actions)
RCM output: maintenance task schedule + intervals; or documentation of run-to-failure decision
Failure Mode Analysis
FMEA (Failure Mode and Effects Analysis):
For each component: list all failure modes; for each mode: effect on system, consequences
Functional failure: inability to fulfill a required function to a minimum performance standard
Failure mode: event causing functional failure (e.g., bearing seizure, seal leak, fatigue crack)
FMECA (FMEA + Criticality Analysis):
RPN (Risk Priority Number) = Severity × Occurrence × Detection [scale 1–10 each]
RPN > 100: high priority maintenance or redesign
Severity 10 = catastrophic (safety, regulatory); Occurrence 10 = > 1 per week; Detection 10 = undetectable
Weibull Analysis
Weibull Distribution
Two-parameter Weibull CDF:
F(t) = 1 - exp(-(t/η)^β) [cumulative failure probability; t = time to failure]
f(t) = (β/η) × (t/η)^(β-1) × exp(-(t/η)^β) [PDF]
Parameters:
β (shape parameter): β < 1 = infant mortality (decreasing failure rate); β = 1 = random/exponential (constant failure rate, no wear); β > 1 = wear-out (increasing failure rate)
β = 1.0: exponential (electronic failure); β = 1.5–2: early wear-out; β = 3–4: well-designed mechanical (normal-like)
η (characteristic life, 63.2 percentile): time at which F(η) = 63.2%; η ≈ MTBF when β = 1
Mean time to failure (MTTF):
MTTF = η × Γ(1 + 1/β) [Γ = Gamma function; Γ(1.5) = 0.886; Γ(2) = 1.0; Γ(3) = 2.0]
For β = 3: MTTF = 0.893 × η; for β = 1: MTTF = η (= MTBF)
B-life (percentile life):
B₁₀ = η × (ln(10/9))^(1/β) = η × (-ln(0.90))^(1/β) = η × (0.1054)^(1/β)
B₁₀: time by which 10% of population fails; widely used for bearings
B₅₀ = median life = η × (ln 2)^(1/β) = η × (0.693)^(1/β)
Example:
β = 2.5, η = 10,000 hr → B₁₀ = 10,000 × (0.1054)^(0.40) = 10,000 × 0.476 = 4,760 hr
Weibull Parameter Estimation
Median rank plotting:
Rank failures by time: t₁ < t₂ < ... < tₙ
Median rank ≈ (i - 0.3) / (n + 0.4) [Bernard approximation; i = rank; n = sample size]
Plot on Weibull paper: ln(ln(1/(1-F))) vs. ln(t) → straight line with slope β, intercept at t = η
Maximum Likelihood Estimation (MLE):
For complete data: Σ t_i^β × ln(t_i) / Σ t_i^β - 1/β - Σ ln(t_i)/n = 0 [solve numerically for β]
η = (Σ t_i^β / n)^(1/β)
Software: Weibull++ (Reliasoft), Python scipy.stats.weibull_min, R CRAN survival
Goodness of fit: Kolmogorov-Smirnov or Anderson-Darling test; p > 0.05 → Weibull adequate
Failure Consequence Categorization
RCM Decision Logic
Hidden vs. evident failures:
Hidden failure: not evident to operating crew during normal operation (e.g., standby system, backup)
Evident: immediately noticed by operator (pump stops; alarm activates)
Consequence categories:
A — Safety consequence (HSE): hidden + safety; evident safety + environment
B — Operational consequence: affects output, quality, cost; evident operational
C — Non-operational consequence: economic only; no production impact
D — Hidden failure, non-safety: reduces redundancy; find via failure finding tasks
Default actions by consequence:
Hidden safety (A): failure-finding task (periodic testing to reveal hidden failure) if possible; redesign if not
Evident safety: scheduled restoration or discard if technically feasible; redesign if not
Operational/non-operational: scheduled task only if cost-effective; otherwise run-to-failure
P-F Interval and Condition Monitoring
P-F Concept
P-F interval: time from detectable potential failure (P) to functional failure (F)
Maintenance task interval: T_task ≤ P-F interval / 2 [conservative; ensures detection before failure]
Condition monitoring technologies:
| Technology | P-F interval | Application |
|---|
| Vibration analysis | 1–6 months | Rotating machinery: bearings, imbalance, misalignment |
| Oil analysis (spectrometry) | 1–6 months | Engine wear metals; contamination |
| Infrared thermography | 1–12 months | Electrical joints, heat exchangers, refractory |
| Acoustic emission | Days–weeks | Crack initiation; bearing early failure; vessel leaks |
| Ultrasonic (UT thickness) | Months–years | Corrosion, erosion, wall thinning |
| Motor current signature | 1–6 months | Rotor bar cracking, bearing looseness |
| Process parameter trending | Hours–days | Pump efficiency drop; compressor surge approach |
Vibration alarm levels (ISO 10816):
| Machine class | Good (mm/s rms) | Acceptable | Alarm | Trip |
|---|
| Small machines (< 15 kW) | < 2.3 | 2.3–4.5 | 4.5–7.1 | > 7.1 |
| Medium (15–75 kW) | < 4.5 | 4.5–9.3 | 9.3–14.7 | > 14.7 |
| Large rigid (> 75 kW) | < 7.1 | 7.1–11.2 | 11.2–18 | > 18 |
Maintenance Task Selection
Task Types
Scheduled restoration: restore to new condition at fixed interval; effective if β > 1 (wear-out); useless if β ≤ 1
Scheduled discard (replacement): replace at fixed interval; applicable for items where cost of restoration > new
Condition-based maintenance (CBM/PdM): monitor parameter; act when condition threshold crossed; requires P-F > 2× monitoring interval
Failure finding: periodic test of hidden function; test interval T = target MTBetween hidden functional failures × average probability of finding failure at test
Run-to-failure: no proactive action; acceptable only for non-safety, non-operational consequence items
Optimum Maintenance Interval (Age-Based)
Optimal replacement age (for β > 1):
Minimize total cost rate C(T) = (C_pm + C_cm × F(T)) / (MTTF_truncated) [C_pm = PM cost; C_cm = CM cost]
Numerical optimization: dC/dT = 0 → optimal T* from:
C_cm × f(T*) × R(T*) + C_pm × F(T*) × f(T*) = 0 [implicit; solve numerically]
Rule of thumb: optimal T ≈ 0.5 × B₁₀ for typical rotating machinery with β = 2–3
MSG-3 (Aviation RCM)
MSG-3 (Airlines for America): structured process for commercial aviation maintenance program development
Applied during aircraft certification; defines minimum maintenance tasks for FAA/EASA approval
Consequence categories: Safety; Airworthiness; Economic; Operational
ATA iSpec 2200: maintenance task format standard for airlines
Maintenance Review Board (MRB): airline regulatory process; approves MSG-3 results as basis for maintenance program
Applies to: commercial transport aircraft, helicopter, engine, APU
Standards and References
| Standard | Scope |
|---|
| SAE JA1011 | Evaluation criteria for RCM processes |
| SAE JA1012 | A guide to the reliability-centered maintenance standard |
| IEC 60300-3-11 | Dependability management — RCM application guide |
| MSG-3 (A4A) | Operator/manufacturer scheduled maintenance development |
| MIL-STD-1629A | FMECA procedures for military hardware |
| IEC 60812 | FMEA technique |
| ISO 10816 | Vibration evaluation of machinery (in-situ measurements) |
Output
Provide: system/asset description (function; operating context; criticality), FMEA table excerpt (top 5–10 failure modes; functional failure; mode; effect; RPN), consequence categorization (hidden/evident; safety/operational/economic; consequence class A/B/C/D), Weibull parameters (β [shape]; η [hr characteristic life]; B₁₀ life [hr]; MTTF [hr]; data source: field data/Weibull++ estimate), P-F interval (technology; P-F [months]; recommended task interval [months]), task selection (type: CBM/scheduled restoration/discard/failure finding/run-to-failure; justification per decision logic), optimal maintenance interval T* [hr] (cost optimization: C_pm [$/task]; C_cm [$/failure]; β; t*), condition monitoring thresholds (vibration [mm/s rms] alarm/trip; or oil particle count; or IR temperature limit), and applicable standard (SAE JA1011/JA1012, IEC 60300-3-11, MIL-STD-1629A).
