• -55%
  • New
GMW8758
search
  • GMW8758

GMW8758

$43.00
$19.35 Save 55%

GMW GMW8758 [ Revised ] Calibrated Accelerated Life Testing (CALT)

standard by General Motors Worldwide , 10/01/2004

Quantity

Full Description

ENGLISH HISTORICAL VERSION

The CALT process is applicable to a wide array of failure mechanismand applies to a wide range ofproducts. Failure mechanisms such as fatigue, wear, thermaldegradation, corrosion, color-change,voltage effects, and other cumulative damage processes are goodcandidates for CALT. Generally thefailure mechanisms should be addressed "one at a time", however, theprocess can model the effectof two failure mechanisms occurring simultaneously. Two stress CALTprocesses increase the cost andtime demands of the CALT process. The following discussion willdescribe a single stress analysisonly, and can be performed without specialized software tools.Specialized software tools areavailable (see reference section) to increase the accuracy and speedof this process.

Note: Nothing in the specification supersedes applicable laws andregulations unless specificexemption has been obtained.

Note: In the event of conflict between the English and domesticlanguage, the English languageshall take precedence.

Purpose. The Calibrated Accelerated Life Test methodology describesthe design and analysis of atest process that allows time compression in life testing whileretaining correlation in resultinglife predictions. This standard is intended for use as a guideline inplanning, executing, andanalyzing Calibrated Accelerated Life Tests (CALT) where a significantreduction in test time isdesirable as well as quantification of reliability. The CALTmethodology proposed by Larry Edson inthe following text leverages the previous work of Wayne Nelson,William Meeker, and PantelisVassiliou. The CALT method focuses on situations needing extensivetime compression, thus requiringsignificant extrapolation. The desire to combine "fast learningcycles" in testing with"calibration to reality" is the fundamental underlying objective. Thismethodology is especiallywell suited for situations where a very high level of quantifiedreliability is required, thusbeing ideal for Reliability Design for Six Sigma.

Special situations involving rubber isolators and mixed failure modesare addressed and are theresult of product testing experience.

Definitions.

Acceleration Factor - A unitless value that relates life at normalstress to life under acceleratedstress:

Arrhenius Model - A math model developed for chemical reaction ratesby the Swedish physicalchemist Svandte Arrhenius in 1887. Generally used where temperature isthe accelerating stress.

Highest Stress Level - Highest level of stress used in testing that isless than a predeterminedfoolish limit.

Inverse Power Model - An empirically derived math model that evolvedfrom relating fatigue life tolevel of stress being applied. The Coffin-Manson relationship is aderivative of this basic model.

Lowest Stress Level - Lowest level of test stress used and the closestto the normal stress level.

Middle Stress Level - Second highest level of test stress used.

Miners Rule - This concept is based on the idea that a productcontains a certain amount of lifefor any given stress type and the product will fail when all of thatlife is used up. The totalquantity of life can be used up as fractional amounts at differentstress levels. When the sum ofthe fractional amounts equals "1" the product fails. Figure 6graphically explains this concept.

Normal Stress Level - The level of stress that the product is expectedto experience as defined bythe specification. This value should not be an average user, butshould represent an extreme leveluser.

Stress-Life Model - A mathematical model depicting the relationshipbetween the life of the productand the level of stress being applied.

S-N Slope - The slope of the stress-life line when plotted onlog-linear or log-log paper. Thestress-life line shows the relationship between Stress level and theNumber of cycles of life atthat stress. Example: The slope of the line for the IPL model iscalculated as shown:

Weibull Analysis - Fitting data to a distribution that has parametersfor location, variability,and shape. Waloddi Weibull, a Swedish engineer/scientist, introducedthis distribution to the worldin 1939 and gained popularity in 1951. Saab employed Mr. Weibull earlyin his career.

R00002609
chat Comments (0)
No customer reviews for the moment.