Life Data Analysis Flashcards

Which plotting paper linearizes the normal distribution for life data probability plots?

Answer: Normal probability axes with linear time scale

Normal probability paper uses a transformed Y-axis based on the inverse standard normal so that normally distributed data plots as a straight line.

In life data analysis, 'Type I censoring' means the test is terminated:

Answer: After a fixed time period regardless of failures

Type I (time-terminated) censoring ends the test at a predetermined time, so the number of failures observed is random.

The cumulative distribution function (CDF) of the Weibull distribution F(t) equals 0.632 when t equals:

Answer: The characteristic life η (eta)

At t = η (the characteristic life), F(η) = 1 − e^(−1) ≈ 0.632 regardless of the shape parameter β.

When multiple competing failure modes are present, removing one failure mode from analysis will:

Answer: Increase the estimated reliability of the product

Eliminating a failure mode treats those failures as suspensions, which reduces the effective failure rate and increases estimated reliability from remaining modes.

A Fisher matrix confidence bound in Weibull MLE analysis is valid primarily for:

Answer: Large sample sizes where asymptotic normality holds

Fisher matrix (likelihood-based) confidence bounds rely on asymptotic normality of MLE estimates, which requires sufficiently large sample sizes.

The mean life (MTTF) for a Weibull distribution with shape β and characteristic life η is given by:

Answer: η · Γ(1 + 1/β)

MTTF = η · Γ(1 + 1/β) where Γ is the gamma function, accounting for the shape of the distribution.

In life data analysis, 'sudden death testing' refers to a strategy where:

Answer: Groups of units are tested and each group test ends at the first failure in that group

Sudden death testing runs multiple groups in parallel, stopping each group at its first failure, which reduces test time while providing distributional information.