Engineering StatisticsEngineering Statistics

Where the spread in engineering data comes from and how to describe it: measurement error and gauge studies, honest summaries of process data and probability plots that check the normal model before you trust it.

1. 1Measurement Error: Accuracy and Precision
2. 2Summarizing Process Data
3. 3Checking Normality with Probability Plots

The workhorse models for counting defects and timing failures: binomial and Poisson counts with yield, the exponential lifetime and its Poisson process and the Weibull distribution for parts that die young or wear out.

1. 4Counting Defects: Binomial and Poisson
2. 5Time to Failure: The Exponential Model
3. 6The Weibull Distribution

How part-to-part variation adds up in assemblies and calculations: means and variances of sums, worst-case and root-sum-square tolerance stacks, propagation of error through formulas and Monte Carlo simulation when the formulas run out.

1. 7Sums of Random Variables and Tolerance Stack-Up
2. 8Propagation of Error
3. 9Monte Carlo Tolerance Analysis

How long things last and how to keep them running: reliability and hazard functions, the bathtub curve and MTTF, series, parallel and k-out-of-n systems, and repair, availability, spares and preventive replacement.

1. 10Reliability and Hazard Functions
2. 11Series, Parallel and k-out-of-n Systems
3. 12MTBF, Repair and Availability

Confidence intervals for a mean and for σ, prediction intervals for the next part, tolerance intervals for nearly every part, and choosing the sample size and power before the data come in.

1. 13Intervals for a Mean and for Variability
2. 14Prediction Intervals
3. 15Tolerance Intervals
4. 16Planning Sample Sizes and Power

Hypothesis tests for engineering decisions: a process mean against a spec, two processes or a before-and-after change, and variability with χ² and F tests.

1. 17Testing a Process Against a Spec
2. 18Comparing Two Processes
3. 19Testing Variability: χ² and F Tests

Least squares on engineering data: calibrating sensors and running the curve backward, reading residuals and spotting influential points, fitting several inputs at once, and straightening power laws and Arrhenius models with logs.

1. 20Calibration Curves
2. 21Residuals and Diagnostics
3. 22Multiple Regression
4. 23Linearizing Engineering Models

Planning experiments that answer cause-and-effect questions: randomization, replication and blocking, two-level factorial designs with their main effects and interactions, and fractional factorials for screening many factors cheaply.

1. 24Randomization, Replication and Blocking
2. 252ᵏ Factorial Designs
3. 26Fractional Factorials and Screening

Watching a process with X̄, R, p and c control charts, judging whether it can meet its spec with Cp, Cpk and ppm, and deciding whether to accept incoming lots with sampling plans and OC curves.

1. 27X̄ and R Control Charts
2. 28p and c Charts for Defects
3. 29Process Capability: Cp and Cpk
4. 30Acceptance Sampling and OC Curves