RAMS Core

RAMS Core glossary

Every symbol, abbreviation and term used across the five RAMS Core concept topics, defined in one place. Entries link to the topic that develops them, so this page works as a front door as well as a lookup.

The five properties

The RAMS quartet plus the enabling property, each defined as the standards define it.

Reliability: The ability of an item to perform as required, without failure, for a given time interval, under given conditions. Quantified, it is the probability of surviving a stated mission. Reliability
Availability: The ability to be in a state to perform as required, under given conditions. Quantified, it is the fraction of time the system is fit for use. Availability
Maintainability: The ability to be retained in, or restored to, a state to perform as required, under given conditions. Quantified, it is the probability that a failed item is restored within a stated time. Maintainability
Safety: Freedom from risk which is not tolerable. Not freedom from all risk: absolute safety does not exist, so safety is always defined against a tolerability judgment. Safety
Testability: The design characteristic that allows an item's status to be determined and its faults to be located in a timely and confident manner. Testability
Dependability: The umbrella term in the international vocabulary for the ability to perform as and when required, covering reliability, availability, maintainability and maintenance support together.

Reliability measures and symbols

The functions and point measures that describe time to failure.

Reliability functionR(t): The probability of surviving beyond time t. Starts at 1 and can only fall. Reliability
Cumulative distribution functionF(t): The fraction of the population failed by time t. F(t) = 1 − R(t). Reliability
Probability density functionf(t): The derivative of F(t): where on the time axis failures concentrate. Reliability
Hazard rateh(t): The instantaneous failure rate among survivors, h(t) = f(t)/R(t). The conditional view, and the one with physical meaning: falling, flat or rising tells you which failure regime you are in. Reliability
Cumulative hazardH(t): The integral of the hazard rate. Gives the central identity R(t) = e^(−H(t)), from which every named life distribution follows. Reliability
Failure rateλ: The hazard rate, usually assumed constant. Quoted in failures per 10⁶ hours or in FIT. Reliability
Mean time between failuresMTBF: The mean operating time between failures of a repairable item. Not a lifetime and not a guarantee: about 63% of an exponential population fails before it. Reliability
Mean time to failureMTTF: The mean time to failure of a non-repairable item, equal to the area under the survival curve. Using MTBF for a non-repairable item is a category error. Reliability
Mean time between maintenanceMTBM: The mean time between all maintenance actions, corrective and preventive, unlike MTBF which counts only failures. Availability
Failures in timeFIT: One failure per 10⁹ device-hours. A component at 100 FIT has λ = 10⁻⁷ per hour. Reliability
Annualised failure rateAFR: The expected fraction of a population failing in a year. Common in storage and consumer electronics. Reliability
B-lifeB10: The time by which a stated percentage of the population has failed. B10 is the 10% point; B1 and B5 likewise. Often a more honest requirement currency than a mean. Reliability
Mission reliability: The probability of completing a defined mission without a failure that aborts or degrades it. Redundancy improves it while making basic reliability slightly worse. Reliability
Basic reliability: A count of every failure, because every failure drives maintenance and logistics, whether or not it affects the mission. Reliability

Distributions and life models

The statistical models fitted to failure and repair data.

Exponential distribution: The constant-hazard model, R(t) = e^(−λt). The only memoryless life distribution, which is why it makes system arithmetic linear and why it must be earned rather than assumed. Reliability
Weibull distribution: The general-purpose life model, R(t) = e^(−(t/η)^β). Its shape parameter diagnoses the failure regime directly from data. Reliability
Shape parameterβ: The Weibull exponent. β < 1 means falling hazard (early failures), β = 1 constant (random), β > 1 rising (wear-out). Each regime needs a different countermeasure. Reliability
Characteristic lifeη: The Weibull scale parameter: the time by which 63.2% of the population has failed, whatever the shape. Reliability
Lognormal distribution: The classic model for corrective repair times and for multiplicative degradation processes. Right-skewed, so its mean always exceeds its median. Maintainability
Repair rateμ: The rate at which repairs complete among still-open repairs, the maintainability mirror of λ. Mean repair time is 1/μ. Maintainability
Bathtub curve: The composite population hazard: a falling early-failure region, a flat useful-life region and a rising wear-out region. A vocabulary for regimes, not a law of nature for a device. Reliability
Censoring: Data in which units have not failed by the end of observation (right-censored) or failed between inspections (interval-censored). Fitting to failures alone while discarding survivors is biased. Reliability
Load-strength interference: The model in which failure occurs where the load and strength distributions overlap. Shows why variance matters as much as mean, and why a safety factor blind to scatter is inadequate. Reliability
Safety margin: The separation between mean strength and mean load, measured in units of the combined scatter: SM = (μS − μL)/√(σS² + σL²). Reliability

Maintainability and support

Repair-time measures and the support system that surrounds them.

Maintainability functionM(t): The probability that a maintenance action is complete by elapsed time t. The mirror of R(t): starts at 0 and rises. Maintainability
Mean time to repairMTTR: Mean active corrective maintenance time at a stated level of repair. Active time only: it is not the downtime the operator experiences. Maintainability
Maximum corrective maintenance timeM_max: A percentile of the repair-time distribution, conventionally the 90th or 95th. The commitment that actually protects a maintenance window, which a mean cannot. Maintainability
Mean downtimeMDT: Active repair plus logistics delay plus administrative delay. The number availability actually feels. Availability
Logistics delay timeLDT, MLDT: Waiting for spares, people, transport or test equipment. A support-system property, not a design property, and often larger than the wrench time. Availability
Administrative delay timeADT: Work-order processing, approvals, permits and priority queues. The cheapest slice of downtime to fix and the most embarrassing to measure. Availability
Maintenance man-hours per operating hourMMH/OH: The labour ledger, distinct from elapsed time: three technicians for one hour is one hour of MTTR and three of MMH. Maintainability
Line-replaceable unitLRU: A unit designed to be removed and replaced at the operating location, normally at organizational level. Maintainability
Shop-replaceable unitSRU: A module or board exchanged inside an LRU at an intermediate or depot facility. Maintainability
Levels of maintenance: The echelon structure, classically organizational (at the system), intermediate (a nearby shop) and depot (factory-grade). Every repair-time number is conditional on the echelon it was measured at. Maintainability
Maintenance concept: The programme document stating who maintains the system, where, with what skills, tools and echelons. Every repair-time claim silently cites it. Maintainability
Level of repair analysisLORA: The economic and non-economic decision of where each item is repaired, or whether it is discarded instead. Maintainability
Maintenance task analysisMTA: Step-level analysis of a maintenance action: the sequence, elapsed time, people, skills, tools and spares it consumes. Maintainability
Reliability centred maintenanceRCM: Choosing the preventive task for each failure mode from its consequences, rather than applying blanket schedules. Its founding study showed that most complex-equipment modes gain nothing from scheduled overhaul. Maintainability

Availability measures

The three named availabilities and the arithmetic that connects them.

Inherent availabilityAi: MTBF/(MTBF + MTTR). The design's promise under ideal support: active corrective repair only, no delays, no preventive maintenance. Availability
Achieved availabilityAa: Adds preventive maintenance to the downtime, still assuming ideal support. Penalises designs that demand frequent or long scheduled attention. Availability
Operational availabilityAo: Uptime divided by total time, counting everything: repair, preventive work, logistics and administrative delay. The number the mission feels and the only one that notices whether the spare was on the shelf. Availability
Materiel availabilityAm: The fleet-wide form: the percentage of the total inventory operationally capable at a given time, including units in depot or reserve. Availability
Point availabilityA(t): The probability the item is up at instant t. For constant rates it settles exponentially to the steady-state balance μ/(λ+μ). Availability
UnavailabilityQ, q: 1 − A. The better working currency, because small unavailabilities add in series and multiply in parallel, which makes downtime budgets allocatable. Availability
Nines: The shorthand for availability classes. Each added nine divides the annual downtime budget by ten: 99% is 87.7 hours a year, 99.999% is 5.3 minutes. Availability
Performance-based logisticsPBL: Buying the outcome (availability, or mission-capable hours) rather than spares and repairs as transactions, so the logistics ledger sits with the party that can act on it. Availability

Safety and risk

The hazard vocabulary, the integrity ladders and the tolerability framework.

Harm: Injury or damage to the health of people, or damage to property or the environment. Safety is defined against consequences to people and assets, not against loss of function. Safety
Hazard: A potential source of harm. A state or condition, not an event and not a failure. Safety
Risk: The combination of the probability of occurrence of harm and the severity of that harm. Two dimensions, always. Safety
Mishap: In defence practice, an event or series of events resulting in unintentional death, injury, occupational illness, damage to or loss of equipment or property, or environmental damage. Safety
Tolerable risk: The level of risk accepted in a given context based on the current values of society. The international guide treats tolerable and acceptable risk as synonymous. Safety
Risk assessment codeRAC: The pairing of one severity category with one probability level, which a matrix maps onto a risk level and therefore onto who may accept it. Safety
Hazard tracking systemHTS: The closed-loop record of every hazard: causes, initial and target risk, mitigations, verification and formal acceptance. A live control instrument that outlives the programme. Safety
Safety case: A structured demonstration that a system is acceptably safe for a defined application in a defined environment, made explicit as a claim, an argument and the evidence beneath it. Safety
Development assurance levelDAL: The aviation integrity ladder, A (most severe) to E, assigned from the failure condition classification. It governs the rigour of development, not a probability. Safety
Function and item development assurance levelFDAL, IDAL: The distinction between the assurance level carried by a function and that carried by each item implementing it. Independent architecture members can carry lower item levels than the function. Safety
Safety integrity levelSIL: The functional safety ladder, SIL 1 to SIL 4, combining a target failure measure with systematic capability and architectural constraints. Safety
Automotive safety integrity levelASIL, QM: The automotive ladder, QM and ASIL A to D, determined by looking up severity, exposure and controllability in the standard's determination table. Never a product of the three. Safety
Tolerable hazard rateTHR: The railway expression of an integrity target: a permitted rate of occurrence per hour for an identified hazard, apportioned from a system-level safety target. Safety
Probability of failure on demandPFD: The average probability that a dormant protective function fails to act when demanded. For a proof-tested channel it approximates λT/2. Safety
Probability of dangerous failure per hourPFH: The high-demand or continuous-mode counterpart of PFD, expressed as an average frequency rather than a probability per demand. Safety
As low as reasonably practicableALARP, SFAIRP: The duty to reduce risk unless the sacrifice in money, time or trouble is grossly disproportionate to the benefit. An argument to be documented, not a calculation. Safety
Random and systematic failures: Random hardware failures arise from physical degradation and have a computable rate; systematic failures are designed or built in and do not. The split is why development assurance exists. Safety
Fail-safe: Failing to a state known to be safe: signal to red, valve closed, brakes applied. Available only where a genuinely safe state exists and reaching it is not itself dangerous. Safety
Fail-operational, fail-passive, fail-silent: The other failure directions: continuing to perform the function, stopping without disturbance, or failing without producing wrong output. Safety
Common cause failureCCF: A single mechanism defeating several channels at once. Floors redundancy's benefit: once a shared fraction exists, adding channels stops helping. Reliability
Common cause analysisCCA: The aviation triad that attacks independence claims: zonal safety analysis, particular risks analysis and common mode analysis. Safety
Zonal safety analysisZSA: Asks what shares physical space with what, and whether the installation as built preserves the separation the drawings assumed. Safety
Particular risks analysisPRA: Asks which single external threats cross zone boundaries by their nature: fire, rotor burst, lightning, flooding, bird strike. Safety
Common mode analysisCMA: Asks what shared non-physical thing touched both channels: the same part batch, software build, procedure, calibration or wrong assumption. Safety
Cut set: A set of basic events whose simultaneous occurrence causes the fault tree's top event. Minimal cut sets are the design's real vulnerability list. Safety
Cut set order: How many events a cut set contains. Order one is a single point of failure, which civil aviation certification forbids for a catastrophic failure condition whatever its probability. Safety
Proof test: The periodic exercise that reveals dormant failures in a protective function. Its interval is a designed safety parameter with the same standing as a failure rate. Availability
Defence in depth: Independent successive layers each capable of preventing or mitigating the hazardous outcome, so the argument does not rest on any one of them. Safety

Testability and diagnosis

Coverage measures, built-in test and the honesty measures.

Fault detection coverageFFD: The failure-rate-weighted fraction of failures detected by the stated test means. Weighted by λ, never counted by mode. Testability
Fault isolation coverageFFI: Of detected failures, the fraction isolated to an ambiguity group of at most n replaceable units. Isolation to one and to three are different worlds for the technician. Testability
Ambiguity group: The set of replaceable units a symptom cannot distinguish between. With equal priors, resolving a group of n takes (n+1)/2 swap-and-retest attempts on average. Testability
Built-in testBIT: Test capability designed into the item itself, rather than applied by external equipment. Testability
Continuous, power-up and initiated BITCBIT, PBIT, IBIT: The three in-service BIT layers: always running in the background, run at start-up as a readiness gate, and run on command with injected stimulus. Testability
False alarm: A failure indication when nothing is wrong. Spends the credibility budget: a crew that has been burned learns to ignore the annunciation. Testability
Cannot duplicateCND: A field indication that maintenance cannot reproduce. The false-alarm rate as the field actually experiences it. Testability
Retest OKRTOK: A removed unit that tests good at the next level of repair. The isolation shortfall as the supply chain experiences it. Testability
No fault foundNFF: The umbrella for units removed and found serviceable. A major maintenance cost driver and the clearest symptom of a testability shortfall. Testability
Dependency model: The matrix of which test responds to which failure mode. From it fall detection coverage, isolation signatures, ambiguity groups and the undetected-λ list, before hardware exists. Testability
Undetected fraction: 1 − FFD: the failures nothing announces. A designed number with its own budget, not a residue, because it feeds the latent-unavailability and hidden-danger ledgers. Testability
Boundary scanJTAG: On-chip test circuitry providing a standard access path for testing assembled boards, standardised as IEEE 1149.1. Test access designed in as infrastructure. Testability
Partial stroke testing: Exercising a shutdown valve through part of its travel between full proof tests, raising effective coverage without stopping the process. Safety

Structures and system modelling

How component behaviour composes into system behaviour.

Series structure: Every element must work. Reliabilities multiply and, for small values, unavailabilities add. The reason complex systems disappoint. Reliability
Active parallel redundancy: All channels energised, any one sufficient. Unavailabilities multiply, so two independent channels give the q² benefit, provided the independence claim holds. Reliability
k-out-of-n: The system works while at least k of n items work. The 2-out-of-3 vote is the workhorse, tolerating one failure while out-voting a channel that fails dirty. Reliability
Standby redundancy: Spares held in reserve and switched in on failure: cold (unpowered), warm (partially energised) or hot. The switch's own reliability multiplies the whole arrangement. Reliability
Reliability block diagramRBD: The success-path model of a system, from which series, parallel and k-out-of-n arithmetic produces system reliability and availability. Availability
Fault tree analysisFTA: The deductive model working downward from an undesired top event through the gates and combinations that produce it, yielding cut sets and a quantified probability. Safety
Failure modes and effects analysisFMEA, FMECA: The inductive model working forward from each failure mode to its effects, with criticality added in the FMECA form. The shared inventory the other analyses consume. Safety
Functional hazard analysisFHA: Enumerates each function's failure conditions, classifies them by severity and derives the safety targets and assurance levels everything downstream is measured against. Safety
Markov and state modelling: Solving for state probabilities when repair, detection, switching or testing couple the states, where the product rules stop applying. Availability
Importance measures: Rankings of which basic events and which cut sets actually drive the top-event probability, converting a model from a score into an improvement agenda. Safety
HAZOP: A guideword-driven study asking what deviations from design intent are possible at each point of a process. Safety
STAMP and STPA: The system-theoretic accident model and its hazard analysis technique, treating safety as a control problem: accidents arise from inadequately enforced constraints, not only from component failures. Safety

Programme and lifecycle

The activities and artefacts that manage these properties.

Allocation: Apportioning a system-level target into budgets each designer can own. Applies equally to failure rates, repair times, downtime minutes and coverage. Reliability
Prediction: Estimating a quantity from the design before hardware exists: failure rates from parts and stresses, repair times from access and diagnosis features, coverage from a dependency model. Reliability
Demonstration: Converting a prediction into evidence on real hardware against a statistical accept-or-reject criterion, whether of reliability, repair time or fault detection. Maintainability
Reliability growth: The improvement achieved by a test, analyse and fix programme, tracked against cumulative test time. Growth happens at the rate the programme funds root-cause analysis. Reliability
Environmental stress screeningESS: Stress applied to production units to precipitate latent manufacturing defects before delivery. Protects design reliability; demonstrates nothing about it. Reliability
Derating: Operating components below their rated stresses. Buys margin against the whole load distribution, and is usually the cheapest reliability available. Reliability
FRACAS: Failure reporting, analysis and corrective action: the closed loop that keeps field experience connected to the analyses, and the field-phase growth mechanism. Reliability
Order of precedence: The mitigation hierarchy in decreasing effectiveness: eliminate by design, reduce by design alteration, engineered features, warning devices, then procedures and protective equipment. Safety
Development assurance: Regulating the rigour of development in proportion to consequence, because the probability of a systematic error cannot be computed. Safety

114 entries across the five concept topics. Terms are defined as the public standards define them; where a standard and common usage disagree, the topic pages say so. Back to RAMS Core.