Reliability: Its Purpose, Home, And Activities

Tim C. Adams
NASA Kennedy Space Center
321-867-2267

June 17, 2004

 Two Schools of Thought

♦ Determinism:
− The doctrine that every event, act, and decision is the inevitable
consequence of antecedents (past events) that are independent
of the human will.

♦ Probabilism:
− The doctrine that probability is adequate basis for belief and
action, since certainty in knowledge cannot be attained.

♦ Many of the tools and techniques in used in Reliability

are probabilistic.

NASA KSC, Tim C. Adams

Page 2
 Comparing Concepts

X: Scenario Y: Likelihood Z: Consequence

How likely is this to If it does happen, what

What can go wrong?
Risk happen? are the consequences?

The element under

Uses historical failure Historical failure data is
study either does or
Reliability data mathematically modeled
does not meet the
to predict failures
failure definition

Likelihood is based on Evaluates hazardous states

Relies heavily on the
Safety identification of hazards
judgment and a that could occur from both
qualitative scale correct and incorrect
element behavior

NASA KSC, Tim C. Adams

Page 3
 What is Reliability?

♦ Reliability is defined as
− the probability an item will perform its intended function for a
given time period under a given set of operating conditions.

♦ Another name for reliability (R) is the “probability of

success” (ps).

♦ Many organizations speak in terms of the “probability

of failure” (pf) being unreliability (U).

♦ Fundamental relationship: ps + pf = 1 or R + U = 1 .
NASA KSC, Tim C. Adams
Page 4
 Other Measures Related to Reliability

♦Reliability deals with reducing the frequency of breakdowns.

♦Maintainability deals with reducing the duration of breakdowns or

downtime.

♦Availability deals with mission readiness which is a function of

Reliability and Maintainability, that is…

 A = uptime/(uptime + downtime) = MTBF/(MTBF + MTTR),

 Where MTBF is mean time between failure and MTTR is mean

time to repair.

NASA KSC, Tim C. Adams

Page 5
 Where does Reliability originate?

♦ It is design engineering where most of the true reliability work is done.

♦ True reliability is built into the design and is called inherent reliability.

♦ Reliability requirements are an integral part of engineering

specifications. As the formal definition implies, at least four items must
be contained in a reliability requirement or specification, namely:
− The intended function (mission) to be performed.
− The desired mission time.
− The operating environment.
− The probability of success (ps) that the product will perform its intended
function.

NASA KSC, Tim C. Adams

Page 6
 Process
Design Phase Reliability Tools & Techniques

♦ To specify reliability: Use all four parts of the reliability definition.

♦ To prevent failures: Use the design strategies in the following order.
− Improve the design to eliminate the failure mode.
− Design for fault tolerance (redundancy).
− Design to be fail-safe (i.e., failure affects function but no injury or additional damage
will occur).
− Provide early warnings of failure through fault diagnosis.
− Note: If these strategies are not viable, the designer may choose to issue special
maintenance instructions and/or use Reliability-centered Maintenance (RCM).
♦ To improve reliability (part 1): Use the applicable design strategies.
− Zero Failure Design: Critical failures are entirely eliminated by design.
− Fault Tolerance: Redundant elements are used to switch over to a backup or
alternative mode.
− Derating: A component is used much below its capability rating.
− Durability: A component is designed a have a longer “useful life” or is designed for
damage tolerance.
− Safety Margins: Design for all applicable worst-case stresses and environments.

NASA KSC, Tim C. Adams

Page 7
 Process
Design Phase Reliability Tools & Techniques (cont.)

♦ To verify reliability: Use the applicable analytical tools.

− Design Reviews: Challenges the design from different viewpoints and identifies and assesses risk
(technical, schedule, and cost).
− Reliability Allocation, Modeling, And Prediction: Provides a hierarchy of design requirements
along with the distributed reliability goal, a model for system configuration, and estimated
(predicated) reliability of the configuration.
− Design Failure Mode, Effects, And Criticality Analysis (Design FMECA or FMEA): Starts at the
component level. Asks what can go wrong and how does it affect the system. Is an inductive (bottom
up) and systematic method and is mostly qualitative.
− Fault Tree Analysis (FTA): Starts at the system major failure or undesired event and decomposes to
it contributing fault occurrences. Is a deductive (top down) and unstructured method; uses symbolic
logic and is always qualitative (e.g., identifies cut sets) with the option of being quantitative.
− Sneak Circuit Analysis: Identifies failures not caused by part failures but are caused by logic flaws.
− Worst-Case Analysis: Typically, used on circuits to evaluates performance when components are at
their high and low values.
− Statistical Analysis: Uses time-to-failure distributions, pass-fail distributions, and stress-strength
distributions to measure predicated or demonstrated reliability.
− Quality Function Deployment (QFD): A method for converting customer needs into engineering
requirements.
− Robust Design (Design of Experiments, DOE): Parameters and tolerance ranges are scientifically
established to optimize performance so that the item is robust in a variety of conditions.

NASA KSC, Tim C. Adams

Page 8
 Process
Design Phase Reliability Tools & Techniques (cont.)

♦ To improve reliability (part 2): Use the applicable engineering tests.

− Reliability Growth Tests: A test that identifies problems and solves them as the
design progresses. Thus, is essentially, a “test, analyze, and fix” method that is
used in a closed-loop corrective action manner.
− Durability Tests – Typically, Accelerated Tests that determine the failure rate
for the entire expected life. Duplicates field failures by providing a harsher but
representative environment. Performed instead of testing under normal
conditions in order to eliminate testing that would otherwise take months or
years.
− Qualification Tests – Consist of stressing the product for all expected failure
mechanisms. The test can be stopped if there are no failures during the expected
life—thus, are performed to measure the achievement of the reliability
requirement. Note: Demonstration Tests or Design Approval Tests are similar
and usually require stressing during only a portion of the useful life. See the tests
used in the manufacturing phase.

NASA KSC, Tim C. Adams

Page 9
 Process
Manufacturing Phase Reliability Tools & Techniques

♦ To prevent or reduce failures: Use the following analytical tools.

− Process Failure Mode, Effects, And Criticality Analysis: Used on the
manufacturing process before it is installed. Similar to Design FMECA.
− Statistical Process Control: Designed to ensure that the manufacturing process
continues to produce products with no more than expected variation in the critical
parameters. Often considered a test for determining the control of quality instead of
reliability.
♦ To prove reliability: Use the applicable accounting tests.
− Environmental Stress Screening Tests: Also, known as Burn-in and Screening
Tests. Tests to catch “infant mortality” failures. If the product is manufactured
properly, these tests are not required. Note: These tests are also performed in the
Design Phase such that early failures do not mask the true reliability. Unfortunately,
these tests are sometimes used as the “final word.” As a result, the screening may not
be long enough and weak products may be provided to the customer.
− Production Reliability Acceptance Tests: Also, known as Failure Rate (MTBF)
Tests. Used to detect any degradation in the inherent reliability of a product over the
course of production and to assure products being delivered meet the customer’s
reliability requirements and/or expectations (by testing a production lot and accepting
or not accepting based on a sampling plan). Also, used to qualify new products.

NASA KSC, Tim C. Adams

Page 10
 Process
User’s Phase Reliability Tools & Techniques

♦ Use the following strategies in the User’s Phase:

− Failure Reporting, Analysis, Corrective Action System: Provides the
data needed to identify deficiencies for correction to ensure that inherent
reliability is not degraded. This system is typically used to record data
for product failures that occurred during all phases of testing as well as
in the field. Also, the data from this system is typically used to detect
trends as early as possible and to respond accordingly in a timely and
preventive manner.
− Note: Weibull Analysis, a type of statistical analysis, is a good tool for
identifying trends in non-repairable systems. For repairable systems
that are not repaired “good-as-new,” start with the Laplace Test.
− Warranties: An attribute where reliability easily affects the
manufacturer’s current and future revenues. One of the biggest
challenges facing manufacturers is competition due to longer warranties.

NASA KSC, Tim C. Adams

Page 11