CRE (Hong Kong Civil Service Exams) flashcards

32 cards. Read the term and the definition here. The deck at the top of this page is optional.

Reliability definition

The probability that an item performs a required function without failure for a specified time interval under stated operating and environmental conditions. It is a dimensionless probability between zero and one.

Exponential reliability function

R(t) equals e raised to the negative product of failure rate and time. It applies when the hazard rate is constant, giving the memoryless property and MTBF equal to one divided by failure rate.

Mean time between failures

Total operating time divided by the number of failures observed, or one divided by the constant failure rate. It is the reciprocal of failure rate and is not the same as the time by which half the units fail.

Weibull shape parameter

Beta below one indicates decreasing hazard (infant mortality), beta equal to one gives a constant hazard matching the exponential, and beta above one indicates increasing hazard from wear-out. Beta describes failure mode, not scale.

Weibull characteristic life

The scale parameter eta, the time at which reliability equals approximately 0.368 regardless of shape parameter. It is often called the characteristic life and differs from mean time to failure unless beta equals one.

Bathtub curve regions

Three sequential hazard regimes: infant mortality with decreasing hazard, useful life with roughly constant hazard and random failures, and wear-out with increasing hazard. Screening and burn-in address the first region; replacement policy addresses the third.

Series system reliability

System reliability equals the product of component reliabilities, and system failure rate equals the sum of component failure rates. Any single component failure causes system failure, so reliability is always lower than the weakest component.

Parallel redundancy reliability

For n identical active units, system reliability equals one minus the quantity one minus unit reliability raised to n. The system fails only when all units fail, so redundancy raises reliability at added cost and weight.

k-out-of-n system

The system succeeds when at least k of n identical units succeed. Reliability sums binomial terms from k to n; k equal to one is full parallel redundancy and k equal to n is a series system.

Hazard rate function

The instantaneous failure rate at time t, equal to the probability density divided by reliability. It has units of failures per unit time and describes conditional failure risk for units that have survived to time t.

Mean time to repair

The average corrective maintenance duration, equal to one divided by repair rate for exponential repair times. It excludes logistics delay and administrative downtime unless explicitly included in the definition.

Operational availability formula

Availability equals mean time between failures divided by the sum of mean time between failures and mean time to repair. Equivalently it equals repair rate divided by the sum of failure rate and repair rate.

Maintainability function

The probability that a failed item is restored to service within a stated time, given specified repair resources and procedures. For exponential repair times it equals one minus e raised to negative repair rate times time.

Reliability block diagram

A logic model showing how component success or failure combines to determine system success. Series blocks require all success; parallel blocks require only one success. It is a success-logic model, not a physical layout drawing.

Redundancy types

Active redundancy keeps all units operating simultaneously; standby redundancy activates a spare on failure. Standby reliability is higher only if switching is perfectly reliable; imperfect switching adds a series failure mode.

Failure modes and effects analysis

A bottom-up qualitative method listing component failure modes, their local and system effects, and severity, occurrence, and detection ratings. The risk priority number is the product of those three ratings and prioritizes corrective action.

Fault tree analysis

A top-down deductive method starting from an undesired top event and resolving causes through AND and OR gates. AND gates multiply probabilities for independent events; OR gates combine as one minus the product of complements.

Safety integrity level

A discrete target level for safety function performance, defined by average probability of failure on demand or by hourly failure rate. Higher SIL demands lower failure probability and greater architectural and diagnostic rigor.

ALARP principle

Risk must be reduced as low as reasonably practicable, balancing further risk reduction against the sacrifice in cost, time, and difficulty. Grossly disproportionate cost can justify accepting a residual risk.

Risk definition

The combination of the likelihood of harm and the severity of that harm. Risk matrices plot likelihood against consequence; risk is reduced by lowering probability, severity, or exposure, in that order of preference.

Common cause failure

Failure of multiple redundant units from a single shared cause such as common power, shared software, or identical environmental stress. It defeats redundancy, so diversity and physical separation are used as defenses.

Derating

Operating components below rated stress levels, such as voltage, temperature, or load, to reduce failure rate. Derating is a design reliability technique applied before hardware is built and validated by testing.

Burn-in screening

Operating or stressing units for a defined period before delivery to precipitate latent infant-mortality defects. It removes early failures from the population but does not change the inherent wear-out distribution of sound units.

Reliability growth testing

A test-analyze-fix process in which failures found during testing are corrected and the demonstrated mean time between failures increases. Growth models such as Duane and AMSAA track cumulative failure rate against cumulative test time.

Confidence interval for MTBF

A range that contains the true mean time between failures with a stated probability. For time-terminated testing with exponential failures, chi-square values with twice the failure count degrees of freedom set the limits.

Demonstrated versus predicted reliability

Predicted reliability comes from part stress and parts-count models before test; demonstrated reliability is statistically estimated from test or field failure data. Demonstrated values carry sampling uncertainty and require confidence statements.

Preventive maintenance interval

The scheduled replacement or servicing time chosen to reduce wear-out failures. Optimal age replacement balances the cost of planned replacement against the expected cost of unplanned failure before the interval.

Failure reporting analysis corrective action

A closed-loop system recording failures, analyzing root cause, implementing corrective action, and verifying effectiveness. It supplies the field data used to update failure rates and reliability predictions.

Sneak circuit analysis

A technique that detects unintended current paths or logic conditions causing functions to occur without command or to be inhibited. It targets design errors rather than component failures and is common in aerospace and safety systems.

Human error probability

The likelihood a required human action fails given the task, environment, and available time. It enters fault trees as basic events and is reduced by procedures, feedback, and human factors design rather than exhortation.

Redundancy versus diversity

Redundancy duplicates identical elements; diversity uses different technologies, vendors, or algorithms. Diversity is the stronger defense against common cause failure because identical units share identical latent design defects.

Product limit reliability estimate

The Kaplan-Meier nonparametric estimator of reliability from censored life data. It multiplies survival factors at each failure time, correctly handling units removed from test before failing.