RAM RAM Data Analysis & Statistics 2 — Questions and Answers
Question 1: What is the Kaplan-Meier estimator used for in reliability analysis?
- Predicting future spare parts demand
- Non-parametrically estimating the survival function from life data including censored observations (Correct answer)
- Calculating the optimal preventive maintenance interval
- Fitting a Weibull distribution to accelerated test data
Correct answer: Non-parametrically estimating the survival function from life data including censored observations
The Kaplan-Meier method estimates the empirical survival function directly from field or test data—including censored units—without assuming any parametric distribution.
Question 2: What is the median rank method used for in reliability probability plotting?
- Estimating the average failure rate of a fleet
- Assigning cumulative failure probability estimates to ordered failure times for plotting on probability paper (Correct answer)
- Calculating the confidence interval around an MTBF estimate
- Determining the minimum sample size for a reliability demonstration test
Correct answer: Assigning cumulative failure probability estimates to ordered failure times for plotting on probability paper
Median ranks provide the best unbiased estimate of cumulative failure probability for each ranked failure time, enabling points to be plotted on Weibull or other probability paper.
Question 3: What does a Q-Q (quantile-quantile) plot assess in reliability data analysis?
- The cost distribution of spare parts over time
- How well the observed failure data fits a hypothesized theoretical distribution (Correct answer)
- The relationship between temperature and failure rate
- The number of censored observations in a dataset
Correct answer: How well the observed failure data fits a hypothesized theoretical distribution
A Q-Q plot compares quantiles of the observed data against quantiles of a reference distribution; points falling close to a straight line indicate a good distributional fit.
Question 4: What is the chi-square goodness-of-fit test used for in RAM statistical analysis?
- Estimating Weibull parameters from field data
- Testing whether observed failure frequency data is consistent with a hypothesized theoretical distribution (Correct answer)
- Calculating confidence intervals for availability estimates
- Optimizing spare parts reorder points
Correct answer: Testing whether observed failure frequency data is consistent with a hypothesized theoretical distribution
The chi-square test compares observed failure counts in intervals to expected counts under the hypothesized model, indicating whether the distribution assumption is statistically acceptable.
Question 5: What does a confidence interval around an MTBF estimate communicate to a reliability engineer?
- The exact true MTBF value of the population
- A range of values within which the true MTBF lies with a specified probability, reflecting uncertainty due to limited sample size (Correct answer)
- The minimum acceptable MTBF per the contract specification
- The range of repair times observed in the field
Correct answer: A range of values within which the true MTBF lies with a specified probability, reflecting uncertainty due to limited sample size
A confidence interval quantifies estimation uncertainty—a 90% CI, for example, means 90% of intervals constructed this way from repeated sampling would contain the true MTBF.
Question 6: What does assuming a constant failure rate (λ = constant) imply about the underlying probability distribution?
- The failure times follow a Weibull distribution with β = 2
- The failure times are exponentially distributed with no aging or infant mortality effects (Correct answer)
- The failure times follow a normal distribution
- The failure rate is increasing linearly with time
Correct answer: The failure times are exponentially distributed with no aging or infant mortality effects
A constant failure rate corresponds exactly to the exponential distribution, which is memoryless—past operating time provides no information about future failure probability.
What is the Kaplan-Meier estimator used for in reliability analysis?