曝光臺 注意防騙 網曝天貓店富美金盛家居專營店坑蒙拐騙欺詐消費者

analysis because the Weibull distribution is the
most commonly used distribution in life data
analysis. Other distributions available in Relex
Weibull include the exponential, Rayleigh,
normal, lognormal, Gumbel upper and lower,
gamma, logistic, and log-logistic distributions.
These distributions can be used to describe
the behavior of a wide range of products in
many industries, including electromechanical
parts, medical devices, electronics, materials
properties, chemical processing equipment,
products subject to maximum loads or
corrosion, parallel systems, series systems,
communications systems, and more.
What makes this kind of analysis so powerful
is its ability to make a number of predictions
and conclusions about a whole population
of products given observed data from a
small representative group of products. Relex
Weibull offers all of this functionality in an
intuitive, easy-to-use interface framework.
Relex Weibull 2009 provides functionality to view multiple plots side-by-side, or to view two data sets on a single plot,
for easy comparisons. Plots display data points, distribution curves, parameter values, confidence, and goodness of fit.
QuarterFlash Online: www.relex.com/news/qflash.asp www.relex.com 7
Quarter Flash
What Is a Distribution?
When the life data related to a product is
found to follow a known distribution, its
behavior is said to be well defined by the
mathematical functions of that distribution.
These functions provide both a graphical
representation of the product's behavior
throughout its lifetime and an equation that is
used to make predictions about the behavior
of members of the population from which the
sample data was taken.
From the functions related to the distributions,
an overall picture of the population's behavior
may be obtained, including when failures are
most likely to occur over time, the growth or
decline of the product's failure rate over the
course of its life, the product's probability
of failure during a certain interval of time,
and the relationship between time and the
probability of failure given the parameters of
the distribution.
Each distribution supported in Relex Weibull
has between one and three parameters, or
numerical values that define its characteristic
behavior. The most familiar parameters
belong to the normal distribution. These are
mu (the average or central tendency of the
data) and sigma (the spread or dispersion of
the data). Parameters define the shape of the
distribution, such that the normal distribution
assumes the familiar “bell curve" shape.
Likewise, other distributions assume their own
characteristic shapes based on the values of
their parameters.
Life Data Analysis in Relex
After entering life data (usually failure times
from observation or testing) from a sampling
of products, Relex Weibull automatically
uses statistical techniques to determine the
parameters of the distribution to which the
data belongs. Usually, the analyst knows
the expected distribution based on previous
testing, the past behavior of like products, or
other information. If the expected distribution
is not known, Relex Weibull provides a Best
Fit Distribution Analysis tool, which ranks all
of the distributions based on how well they fit
the data.
Using the calculated parameters, Relex
Weibull provides statistical information in
both graphical and numeric form. Graphs of
the data describe its characteristic qualities,
including distribution shape, cumulative
failure rate, and instantaneous failure rate. The
Summary Calculator provides key probability
metrics, including the probability of failure or
the reliability of the product at a specified time
in its life.
Other Features of Relex Weibull
Relex Weibull offers additional analytical
capabilities closely related to life data analysis,
including:
• Non-parametric life data analysis, which
follows a process similar to life data analysis
without assuming a distribution.
• Reliability growth analysis, which is
commonly employed during product design
and development to track the improvements
made to the reliability of a product as its
design is perfected.
• Degradation analysis, which is used to
estimate the time to failure of an item from
its rate of decline, without the need to test
the item all the way to failure.
• Warranty analysis, which performs future
returns forecasting and extrapolates life data
from data entered in the warranty format,
　
中國航空網 www.k6050.com
航空翻譯 www.aviation.cn
本文鏈接地址：航空資料37(39)