Location via proxy:   [ UP ]  
[Report a bug]   [Manage cookies]                
Scribd
Upload
42 views

Component System Reliability

This document discusses component reliability analysis from both frequentist and Bayesian perspectives. It defines component reliability as an item's ability to perform its intended function during a mission. Methods to improve reliability include strengthening components, reducing harmful environments, minimizing loads, and preventive maintenance. The document also covers reliability functions, failure rates, common reliability distributions like exponential, Weibull, gamma, and normal, and how to select reliability models based on field and test data.

Uploaded by

ImamHambaliii
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
42 views

Component System Reliability

This document discusses component reliability analysis from both frequentist and Bayesian perspectives. It defines component reliability as an item's ability to perform its intended function during a mission. Methods to improve reliability include strengthening components, reducing harmful environments, minimizing loads, and preventive maintenance. The document also covers reliability functions, failure rates, common reliability distributions like exponential, Weibull, gamma, and normal, and how to select reliability models based on field and test data.

Uploaded by

ImamHambaliii
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 34

Component System Reliability

Pramudita Satria Palar, Ph.D.


Kelompok Keahlian Fisika Terbang
Fakultas Teknik Mesin dan Dirgantara
Institut Teknologi Bandung
Component reliability analysis

FREQUENTIST BAYESIAN

The definition of component encompasses a broad group


of physical items (i.e., components, subsystems, systems,
etc)
Component reliability analysis

• Reliability: an item’s ability to successfully perform an


intended function during a mission.

• How to improve reliability? Stronger and more durable


elements, reducing harmful environmental conditions,
minimizing loads and stress applied, and providing a
preventive maintenance program.
Component reliability analysis

Reliability Designated
conditions
Designed period of time/
number of cycles for the
item application
Reliability function: PoF of the item as a function
of time

Unreliability

• F(t) = Probability that the item will fail sometime up to time t.

Reliability function / survivorship function


P-level quantile
• P-level quantile is defined as the value tp such that

• The median is the quantile of the level p = 0.5

• 100p th percentile of time to failure is the point at which the probability of an item
failure is equal to p.

• B10 life = time at which 10% of the components are expected to fail.
• Other popular percentiles are the 1st , 5th, 10th and 50th percentiles.
Mean time to failure
• The basic characteristics of time-to-failure distribution and basic
reliability measures can be expressed in terms of pdf, f(t), cdf, F(t), or
reliability function R(t)

Expected value of the


density function

MTTF refers to the length of time a system can be expected to


run without failure.
Conditional reliability function
• Conditional reliability function is the probability that the item will survive
for time tau , given that it has survived for time t.

Conditional reliability
function

Conditional probability of
failure
Example: exponential Let’s say lambda = 1/10, then

distribution
PDF

CDF
Reliability
function
MTTF
Failure rate
• Failure rate is the conditional probability that a component fails in a small
time interval, given that is has survived from time zero until the beginning
of the time interval.

Failure PDF
rate Reliability function

- Failure rate is the (instantaneous) rate of failure for the survivors to time t during
the next instant of time
- The hazard rate refers to the rate of death for an item of a given age (x)
Average failure rate

Average
failure rate
Bathtub curve
IFR (increasing
DFR failure rate)
(decreasing
failure rate)

- Burn-in is characterized by early failures attributable to defects in design,


manufacturing, or construction.
- Chance-failure region exhibits a reasonably constant failure rate due to random
failures.
- Wear out is characterized mainly by complex again phenomena, e.g., deterioration
due to accumulated fatigue
Effect of stress in a typical bathtub
curve

- As the stress level increases, the chance-failure region decreases and premature
wear out occurs.
- It is important to minimize stress factors, such as a harsh operating environment,
to maximize reliability.
Component System Reliability

Common distributions
Exponential distribution
• Gives a constant hazard rate model that is
often realistic.
• Can simply represent the chance-failure
region (after burn-in)

PDF

CDF
Reliability
function
MTTF
Weibull distribution
• Covers a variety of shapes, flexibility of describing hazard rates for all
three regions of the bathtub curve.

Scale

Shape

• Applications include: Corrosion resistance studies, time to failure of many hardwares


(electronics and mechanical), time to failure of basic elements of a system
Weibull distribution
Gamma distribution
• A generalization of the exponential distribution

• Applications include: The distribution of times between recalibration of an


instrument, time between maintenance of items that require maintenance after k
uses.
• Alpha shows the number of shocks required before a failure occurs, and beta
represents the mean time to occurrence of a shock
Gamma distribution Check the applet
https://homepage.divms.uiowa.edu/~mbognar/apple
ts/gamma.html
Normal and lognormal distribution
• Normal distribution is mostly used in stress-strength analysis (time-independent)

• Normal distribution is mostly used in stress-strength analysis (time-independent)

• The lognormal distribution represents the distribution of an r.v. whose logarithm follows the
normal distribution.
• Suitable for failure processes that are the result of many small multiplicative errors (e.g.
fatigue cracks)
Component System Reliability

Reliability data and model


selection
Reliability data and model selection
• A probability model describes how an r.v. is spread over its range in
terms of probabilities.
• Let us discuss how field and test data can support the selection of a
probability model for reliability analysis.

Parametric • Based on a selected distribution function

Nonparametric • Do not need a particular


distribution function
Nonparametric graphical procedure
• Attempts to directly estimate the reliability characteristics (e.g., the
PDF, reliability, and hazard rates) from a sample.
• Useful for exploratory data analysis. Small sample case
Nonparametric graphical procedure
Nonparametric graphical procedure: large samples

Number of surviving components in the interval starting from ti

Number of observed failures in the interval ti Δ + t


Probability plotting
• A simple graphical method of displaying and analyzing observed data.
• Each distribution has its own probability paaper.
• From the plotted line, one can roughly estimate the parameter of the
hypothesized distribution.
Probability plotting: exponential distribution
Probability plotting: Weibull

You might also like