Inherent steady-state availability¶

The long-run operating fraction of a repairable item calculated as MTBF divided by MTBF plus mean corrective repair time, excluding planned maintenance and support delays.

Version: v1 · 2026-10-17 · History
Domain-specific #: 14196
Domain group: Applied Sciences & Engineering
Origin domain: Engineering & Design (beyond software)
Subdomain: Availability Engineering → Engineering & Design (beyond software)
Aliases: Steady-state inherent availability

Core Idea¶

Inherent steady-state availability is the long-run fraction of time that a repairable item is in its operating state when the counted downtime is corrective repair. For the stated stationary failure-and-repair model, its mean-time form is \(A_i=\mathrm{MTBF}/(\mathrm{MTBF}+\mathrm{MTTR})\): mean operating time between failures divided by that mean plus mean time spent correcting a failure. The numerator and repair term must refer to the same item, operating conditions and failure boundary.^[1]

The word inherent marks the accounting convention. Scheduled preventive work, waiting for a replacement part and waiting for maintenance resources are excluded. A different convention may give an achieved or operational availability, even for the same item.^[1] The word steady-state marks the limiting horizon, not a claim that any particular observed month had exactly this up fraction. The NASA guidance presents a constant failure-rate and repair-rate model for its displayed formula; those hazard assumptions should be reported as model assumptions, not silently attributed to every field case.^[1]

Structural Signature¶

Sig role-phrases:

Repairable item and states. Identify the receiver, pump or other item and what counts as operating, failed and undergoing corrective repair.^[1] A vague service label cannot determine its mean times.
Failure-side mean. MTBF is the mean operating time between the defined failures. Changing the failure boundary changes the quantity.^[1]
Corrective-repair mean. MTTR is mean time to perform corrective repair for those failures.^[1]
Downtime accounting boundary. Preventive work and support waiting are excluded from this inherent measure; adding them changes the availability type.^[1]
Long-run ratio. Combine the two means as \(\mathrm{MTBF}/(\mathrm{MTBF}+\mathrm{MTTR})\) for a stated stationary model. The output is a dimensionless operating fraction between zero and one when the means are nonnegative and the denominator is positive.^[1]

These five checks implement four functional roles: item/state, MTBF, corrective MTTR and long-run ratio. The separate downtime-accounting bullet specifies the corrective-only boundary of the MTTR role; it is not an independent fifth input. A reported percentage without its item, failure definition, downtime convention and horizon is not enough to classify the calculation as this metric.

What It Is Not¶

It is not every form of availability. Achieved availability counts scheduled maintenance as well as corrective maintenance; operational availability also reflects support delays in an actual operating environment. These are useful quantities with different inputs, not interchangeable estimates of the inherent quantity.^[1]

It is not a finite-period uptime observation or a point-in-time availability probability. A demonstrated fraction for one period can be compared with a model prediction, but it does not by itself establish the limiting fraction.^[1] IBM's broad high-availability server illustration gives the ratio formula and a finite-month example whose stated month and outage hours do not reconcile.^[2] That page does not establish the corrective-only limiting boundary, so the example does not qualify as a full case here.

It is not reliability alone. MTBF describes time between failures; the repair duration also affects inherent availability. It is not the general ratio Prime either: the reliability-engineering identity fixes the numerator, denominator addition, corrective-only downtime and repairable-item boundary.

Scope of Application¶

The metric applies literally to repairable equipment when an analyst can define operating and failed states, estimate MTBF and corrective MTTR for the same boundary, and justify a long-run model. NASA's radar-receiver design appendix uses it to compare candidate reliability and maintainability allocations.^[3] An EPA wastewater-treatment report applies it to sewage pumps and motors and explicitly excludes administrative delay.^[4]

It can be calculated as a design estimate before field operation, but that estimate should not be labeled observed performance. A service calculation that includes preventive work or supply and maintenance waiting uses a different downtime convention.^[1] Other differently scoped downtime measures need their own stated accounting before the inherent formula can be claimed.

Clarity¶

The adjective pair answers two separate questions: which downtime counts, and what time horizon is claimed? Inherent retains corrective repair and excludes other downtime. Steady-state seeks the limiting operating fraction rather than the status at a chosen instant.^[1] A bare “99% available” claim answers neither question. Naming both boundaries makes two apparently conflicting reports comparable or shows why they are measuring different things.

Manages Complexity¶

Repeated cycles of operation and corrective repair can be summarized by two mean durations and one ratio. That compression makes design alternatives comparable: a longer MTBF raises the fraction at fixed MTTR, while a shorter MTTR raises it at fixed MTBF. NASA specifies the failure/repair mean-time model and distinguishes its downtime types.^[1] The compact number must remain attached to that failure definition, repair-time scope and model horizon; otherwise the comparison hides its assumptions.

Abstract Reasoning¶

For a specified item and fixed MTTR, increasing MTBF increases \(A_i\); for fixed MTBF, increasing corrective MTTR decreases it. The same target can be reached with different mean-time pairs. The NASA receiver study gives three such design alternatives at about 0.990, then compares their estimated initial and sustaining costs.^[3] Thus the target fraction constrains a design choice but does not determine which qualifying design has the lowest estimated cost.

The inference stops at the chosen downtime convention. From a high inherent value alone, one cannot infer a high operational value when spare-part queues or maintenance waits are substantial.^[1] Nor does the ratio demonstrate that constant failure and repair hazards were measured in the receiver or pump examples; those cases present calculations, not empirical validation of a hazard model.

Knowledge Transfer¶

Within reliability engineering, the test transfers between unlike repairable items: define the operating and failed states, use the mean operating interval between those failures, use the matching corrective repair duration, and state the limiting model. The NASA receiver and EPA pump calculations retain those roles despite different equipment and design decisions.^[3]^[4]

Outside that setting, the mathematical idea of a long-run up fraction may be useful, but the specialist name is earned only when the corrective-only time accounting survives. A calculation with another numerator, time horizon or downtime convention should not be relabeled by analogy without matching the typed roles.

Examples¶

Canonical: radar-receiver design alternatives¶

NASA CR-61382 Appendix C sets a radar receiver's target inherent availability at 0.990. Its proposed configurations pair MTBF and MTTR as 200 h with 2.02 h, 300 h with 3.03 h, and 350 h with 3.54 h. Each pair gives approximately 0.990 in the mean-time ratio. The receiver is the repairable item, the first number is mean operation between failures, the second is mean corrective repair, and the target is a modeled long-run design property. The report presents these as alternative receiver designs.^[3] They are not three observed operating histories.

Mapped back: item/state → repairable receiver; MTBF → 200, 300 or 350 h; corrective MTTR → 2.02, 3.03 or 3.54 h; limiting ratio → approximately 0.990 under the design model.

Applied: wastewater sewage pumps and motors¶

The EPA report uses 8760 hours MTBF and 13 hours MTTR for sewage pumps and motors, calls the result inherent availability, and says that administrative repair delays such as waiting for a part are excluded.^[4] The inputs give \(8760/(8760+13)\approx0.998518\), or about 0.9985 to four decimal places. The report prints 0.9986.^[4] The decimal here is independently recalculated from its displayed inputs. Its repairable-item, failure-interval, corrective-repair and ratio roles match the radar receiver in a different engineering setting.

Mapped back: item/state → repairable pumps and motors; MTBF → 8760 h; corrective MTTR → 13 h with administrative delay excluded; limiting ratio → approximately 0.9985 from the stated inputs.

Structural Tensions¶

T1: Meeting the target versus choosing the economical route. The receiver alternatives show that roughly equal inherent availability can result from different reliability and maintainability designs. Nelson then compares estimated initial and sustaining costs and selects one configuration on that basis. Improving the mean time between failures and shortening mean repair time both help the ratio, but different ways of obtaining those changes have different costs. A target availability alone does not identify the least-cost design; the answer depends on the item's actual design and support assumptions. The report's cost figures are estimates for its receiver scenario.^[3] Diagnostic: For a specified 0.990 target and this receiver's cost assumptions, which MTBF/MTTR configuration has the lowest estimated total cost? The report's figures are not a transferable price curve.

Structural–Framed Character¶

This entry is structural within reliability engineering. Its recognition test is a repairable item, matching MTBF and corrective MTTR, and a justified long-run mean-time ratio. NASA's radar receiver and the EPA pumps instantiate that structure with different equipment.^[3]^[4]

Evaluative weight: the fraction describes predicted operating time; whether a given value is acceptable depends on an external requirement.
Human-practice dependence: engineers choose the failure boundary and repair-time accounting, but the ratio's stated calculation remains checkable once they are fixed.
Institutional origin: NASA and EPA document applications; neither organization is a necessary role of the metric.
Vocabulary travel: availability travels widely, while its inherent corrective-only meaning must be checked rather than assumed.
Import versus recognition: applying the exact ratio to a new repairable item recognizes the same reliability relation; using the word for a different service percentage imports a label without the roles.

Its character: a specialized reliability measure: ordered division travels widely, while corrective-only downtime, matched failures and a limiting repairable-item horizon retain the reliability-engineering boundary.

Structural Core vs. Domain Accent¶

The portable core is an ordered ratio: mean up time divided by mean up time plus mean down time. The domain accent fixes the down time as corrective repair, the up time as operation between defined failures, and the horizon as a limiting fraction for a repairable item.^[1] Those additions do real classificatory work. A generic ratio lacks them; an operational service percentage can share the same numeric value while counting different periods.

The two unlike cases establish literal reuse within reliability engineering. They do not establish that the full identity travels unchanged across fields with different notions of failure, maintenance and support. A future Prime proposal would need materially unlike fields that independently use the same corrective-only long-run repairable-item roles, rather than only the generic quotient. Does such evidence exist without translating away the failure and repair boundaries?

This entry is a kind of Ratio.

Ratio captures the ordered division: MTBF occupies the numerator and the positive sum MTBF+MTTR the denominator. Ratio alone says nothing about corrective repair or a limiting up fraction. Mean time between failures supplies one input; it is not the availability measure because it omits repair time. Performance Measure concerns a declared evaluative quantity, but need not use this corrective-only mean-time construction. These neighboring identities help locate the calculation without erasing its specialist boundary.

Relationships to Other Abstractions¶

Current abstraction Inherent steady-state availability Domain-specific

Parents (1) — more general patterns this builds on

Inherent steady-state availability is a kind of Ratio Prime

Inherent steady-state availability is the ordered MTBF-to-total-cycle-time ratio for a repairable item under corrective-only downtime accounting.

Hierarchy path (1) — routes to 1 parentless root

Inherent steady-state availability → Ratio → Comparison

Neighborhood in Abstraction Space¶

Inherent steady-state availability sits in a sparse region of the domain-specific corpus (85^th percentile for distinctiveness): few abstractions share its structure, so a faithful description tends to retrieve it precisely.

Family — Supply Chain & Inventory Management (29 abstractions)

Nearest neighbors

Computed from structural-signature embeddings · 2026-10-17

Not to Be Confused With¶

MTBF is only the numerator, not the complete availability. NASA distinguishes corrective MTTR from downtime that also counts preventive work and support delays, and it distinguishes observed uptime from a steady-state prediction.^[1] The EPA source prints 0.9986 for 8760 and 13 hour inputs; their quotient is about 0.9985.^[4]

References¶

[1] Tim Adams, Availability: What Is It?, NASA Kennedy Space Center (June 2024), pp. 1–3, especially the type comparison on p. 2 and steady-state formulas and assumptions on p. 3. https://extapps.ksc.nasa.gov/Reliability/Documents/Availability_What_is_it.pdf registry ↩a ↩b ↩c ↩d ↩e ↩f ↩g ↩h ↩i ↩j ↩k ↩l ↩m ↩n ↩o ↩p

[2] IBM Cloud, “Understanding high availability,” formula and 30-day server illustration. https://cloud.ibm.com/docs/resiliency?topic=resiliency-understanding-ha registry ↩

[3] James H. Nelson, Methodology for the Systems Engineering Process, Volume III: Operational Availability, NASA Contractor Report CR-61382 (March 1972), Appendix C, printed pp. 75–82, especially radar-receiver alternatives on p. 76 and cost comparison and decision on pp. 78–82. https://ntrs.nasa.gov/api/citations/19730003260/downloads/19730003260.pdf registry ↩a ↩b ↩c ↩d ↩e ↩f

[4] Charles W. Mallory and Robert Waller, Application of Selected Industrial Engineering Techniques to Wastewater Treatment Plants, EPA-R2-73-176 (February 1973), section IX, printed p. 101 (sewage pump and motor inherent-availability calculation and administrative-delay boundary). https://nepis.epa.gov/Exe/ZyPURL.cgi?Dockey=9100TFS3.TXT registry ↩a ↩b ↩c ↩d ↩e ↩f