GMW GMW8758 PDF

The CALT process is applicable to a wide array of failure mechanism and applies to a wide range of products. Failure mechanisms such as fatigue, wear, thermal degradation, corrosion, color-change, voltage effects, and other cumulative damage processes are good candidates for CALT. Generally the failure mechanisms should be addressed "one at a time", however, the process can model the effect of two failure mechanisms occurring simultaneously. Two stress CALT processes increase the cost and time demands of the CALT process. The following discussion will describe a single stress analysis only, and can be performed without specialized software tools. Specialized software tools are available (see reference section) to increase the accuracy and speed of this process.

Note: Nothing in the specification supersedes applicable laws and regulations unless specific exemption has been obtained.

Note: In the event of conflict between the English and domestic language, the English language shall take precedence.

Purpose. The Calibrated Accelerated Life Test methodology describes the design and analysis of a test process that allows time compression in life testing while retaining correlation in resulting life predictions. This standard is intended for use as a guideline in planning, executing, and analyzing Calibrated Accelerated Life Tests (CALT) where a significant reduction in test time is desirable as well as quantification of reliability. The CALT methodology proposed by Larry Edson in the following text leverages the previous work of Wayne Nelson, William Meeker, and Pantelis Vassiliou. The CALT method focuses on situations needing extensive time compression, thus requiring significant extrapolation. The desire to combine "fast learning cycles" in testing with "calibration to reality" is the fundamental underlying objective. This methodology is especially well suited for situations where a very high level of quantified reliability is required, thus being ideal for Reliability Design for Six Sigma.

Special situations involving rubber isolators and mixed failure modes are addressed and are the result of product testing experience.

Definitions.

Acceleration Factor - A unitless value that relates life at normal stress to life under accelerated stress:

Arrhenius Model - A math model developed for chemical reaction rates by the Swedish physical chemist Svandte Arrhenius in 1887. Generally used where temperature is the accelerating stress.

Highest Stress Level - Highest level of stress used in testing that is less than a predetermined foolish limit.

Inverse Power Model - An empirically derived math model that evolved from relating fatigue life to level of stress being applied. The Coffin-Manson relationship is a derivative of this basic model.

Lowest Stress Level - Lowest level of test stress used and the closest to 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 product contains a certain amount of life for any given stress type and the product will fail when all of that life is used up. The total quantity of life can be used up as fractional amounts at different stress levels. When the sum of the fractional amounts equals "1" the product fails. Figure 6 graphically explains this concept.

Normal Stress Level - The level of stress that the product is expected to experience as defined by the specification. This value should not be an average user, but should represent an extreme level user.

Stress-Life Model - A mathematical model depicting the relationship between the life of the product and the level of stress being applied.

S-N Slope - The slope of the stress-life line when plotted on log-linear or log-log paper. The stress-life line shows the relationship between Stress level and the Number of cycles of life at that stress. Example: The slope of the line for the IPL model is calculated as shown:

Weibull Analysis - Fitting data to a distribution that has parameters for location, variability, and shape. Waloddi Weibull, a Swedish engineer/scientist, introduced this distribution to the world in 1939 and gained popularity in 1951. Saab employed Mr. Weibull early in his career.