600 Pages 310 B/W Illustrations
    by Chapman & Hall

    Since the publication of the second edition of Applied Reliability in 1995, the ready availability of inexpensive, powerful statistical software has changed the way statisticians and engineers look at and analyze all kinds of data. Problems in reliability that were once difficult and time consuming even for experts can now be solved with a few well-chosen clicks of a mouse. However, software documentation has had difficulty keeping up with the enhanced functionality added to new releases, especially in specialized areas such as reliability analysis.

    Using analysis capabilities in spreadsheet software and two well-maintained, supported, and frequently updated, popular software packages—Minitab and SAS JMP—the third edition of Applied Reliability is an easy-to-use guide to basic descriptive statistics, reliability concepts, and the properties of lifetime distributions such as the exponential, Weibull, and lognormal. The material covers reliability data plotting, acceleration models, life test data analysis, systems models, and much more. The third edition includes a new chapter on Bayesian reliability analysis and expanded, updated coverage of repairable system modeling.

    Taking a practical and example-oriented approach to reliability analysis, this book provides detailed illustrations of software implementation throughout and more than 150 worked-out examples done with JMP, Minitab, and several spreadsheet programs. In addition, there are nearly 300 figures, hundreds of exercises, and additional problems at the end of each chapter, and new material throughout.

    Software and other files are available for download online

    Basic Descriptive Statistics
    Populations and Samples
    Histograms and Frequency Functions
    Cumulative Frequency Function
    The Cumulative Distribution Function and the Probability Density Function
    Probability Concepts
    Random Variables
    Sample Estimates of Population Parameters
    How to Use Descriptive Statistics
    Data Simulation

    Reliability Concepts
    Reliability Function
    Some Important Probabilities
    Hazard Function or Failure Rate
    Cumulative Hazard Function
    Average Failure Rate
    Units
    Bathtub Curve for Failure Rates
    Recurrence and Renewal Rates
    Mean Time to Failure and Residual Lifetime
    Types of Data
    Failure Mode Separation

    Exponential Distribution
    Exponential Distribution Basics
    The Mean Time to Fail for the Exponential
    The Exponential Lack of Memory Property
    Areas of Application for the Exponential
    Exponential Models with Duty Cycles and Failure on Demand
    Estimation of the Exponential Failure Rate λ Exponential Distribution Closure Property
    Testing Goodness of Fit—the Chi-Square Test
    Testing Goodness of Fit—Empirical Distribution Function Tests
    Confidence Bounds for λ and the MTTF
    The Case of Zero Failures
    Planning Experiments Using the Exponential Distribution
    Simulating Exponential Random Variables
    The Two-Parameter Exponential Distribution
    Test Planning Via Spreadsheet Functions
    Determining the Sample Size
    EDF Goodness-of-Fits Tests Using Spreadsheets
    KS Test

    Weibull Distribution

    Empirical Derivation of the Weibull Distribution
    Properties of the Weibull Distribution
    Extreme Value Distribution Relationship.
    Areas of Application
    Weibull Parameter Estimation: Maximum Likelihood Estimation Method
    Weibull Parameter Estimation: Linear Rectification
    Simulating Weibull Random Variables
    The Three-Parameter Weibull Distribution
    Goodness of Fit for the Weibull
    Using a Spreadsheet to Obtain Weibull MLES
    Using a Spreadsheet to Obtain Weibull MLES for Truncated Data
    Spreadsheet Likelihood Profile Confidence Intervals for
    Weibull Parameters

    The Normal and Lognormal Distributions

    Normal Distribution Basics
    Applications of the Normal Distribution
    The Central Limit Theorem
    Normal Distribution Parameter Estimation
    Simulating Normal Random Variables
    The Lognormal Life Distribution
    Properties of the Lognormal Distribution
    Lognormal Distribution Areas of Application
    Lognormal Parameter Estimation
    Some Useful Lognormal Equations
    Simulating Lognormal Random Variables
    Using a Spreadsheet to Obtain Lognormal MLEs
    Using a Spreadsheet to Obtain Lognormal MLEs for Interval Data

    Reliability Data Plotting

    Properties of Straight Lines
    Least Squares Fit (Regression Analysis)
    Rectification
    Probability Plotting for the Exponential Distribution
    Probability Plotting for the Weibull Distribution
    Probability Plotting for the Normal and Lognormal Distributions
    Simultaneous Confidence Bands
    Order Statistics and Median Ranks

    Analysis of Multicensored Data

    Multicensored Data
    Analysis of Interval (Readout) Data
    Life Table Data
    Left-Truncated and Right-Censored Data
    Left-Censored Data
    Other Sampling Schemes (Arbitrary Censoring: Double and Overlapping Interval Censoring)—Peto–Turnbull Estimator
    Simultaneous Confidence Bands for the Failure
    Distribution (or Survival) Function
    Cumulative Hazard Estimation for Exact Failure Times
    Johnson Estimator
    Obtaining Bootstrap Confidence Bands Using a Spreadsheet

    Physical Acceleration Models

    Accelerated Testing Theory
    Exponential Distribution Acceleration
    Acceleration Factors for the Weibull Distribution
    Likelihood Ratio Tests of Models
    Confidence Intervals Using the LR Method
    Lognormal Distribution Acceleration
    Acceleration Models
    The Arrhenius Model
    Estimating ΔH with More Than Two Temperatures
    Eyring Model
    Other Acceleration Models
    Acceleration and Burn-In
    Life Test Experimental Design
    An Alternative JMP Input for Weibull Analysis of High-Stress Failure Data
    Using a Spreadsheet for Weibull Analysis of High-Stress Failure Data
    Using A Spreadsheet for MLE Confidence Bounds for Weibull Shape Parameter
    Using a Spreadsheet for Lognormal Analysis of the High-Stress Failure Data Shown in Table 8.5
    Using a Spreadsheet for MLE Confidence Bounds for the Lognormal Shape Parameter
    Using a Spreadsheet for Arrhenius–Weibull Model
    Using a Spreadsheet for MLEs for Arrhenius–Power Relationship Lognormal Model
    Spreadsheet Templates for Weibull or Lognormal MLE Analysis

    Alternative Reliability Models

    Step Stress Experiments
    Degradation Models
    Lifetime Regression Models
    The Proportional Hazards Model
    Defect Subpopulation Models
    Summary
    JMP Solution for Step Stress Data in Example 9.1
    Lifetime Regression Solution Using Excel
    JMP Likelihood Formula for the Defect Model
    JMP Likelihood Formulas for Multistress Defect Model Example

    System Failure Modeling: Bottom-Up Approach

    Series System Models
    The Competing Risk Model (Independent Case)
    Parallel or Redundant System Models
    Standby Models and the Gamma Distribution
    Complex Systems
    System Modeling: Minimal Paths and Minimal Cuts
    General Reliability Algorithms
    Burn-In Models
    The "Black Box" Approach—An Alternative to Bottom-Up Methods

    Quality Control in Reliability: Applications of Discrete Distributions
    Sampling Plan Distributions
    Nonparametric Estimates Used with the Binomial Distribution
    Confidence Limits for the Binomial Distribution
    Normal Approximation for Binomial Distribution
    Confidence Intervals Based on Binomial Hypothesis Tests
    Simulating Binomial Random Variables
    Geometric Distribution
    Negative Binomial Distribution
    Hypergeometric Distribution and Fisher’s Exact Test
    Poisson Distribution
    Types of Sampling
    Generating a Sampling Plan
    Minimum Sample Size Plans
    Nearly Minimum Sampling Plans
    Relating an OC Curve to Lot Failure Rates
    Statistical Process Control Charting for Reliability

    Repairable Systems Part I: Nonparametric Analysis and Renewal Processes

    Repairable versus Nonrepairable Systems
    Graphical Analysis of a Renewal Process
    Analysis of a Sample of Repairable Systems
    Confidence Limits for the Mean Cumulative Function (Exact Age Data)
    Nonparametric Comparison of Two MCF Curves
    Renewal Processes.
    Homogeneous Poisson Process
    MTBF and MTTF for a Renewal Process
    MTTF and MTBF Two-Sample Comparisons
    Availability
    Renewal Rates
    Simulation of Renewal Processes
    Superposition of Renewal Processes
    CDF Estimation from Renewal Data (Unidentified Replacement)
    True Confidence Limits for the MCF
    Cox F-Test for Comparing Two Exponential Means
    Alternative Approach for Estimating CDF Using the
    Fundamental Renewal Equation

    Repairable Systems Part II: Nonrenewal Processes

    Graphical Analysis of Nonrenewal Processes
    Two Models for a Nonrenewal Process
    Testing for Trends and Randomness
    Laplace Test for Trend
    Reverse Arrangement Test
    Combining Data from Several Tests
    Nonhomogeneous Poisson Processes
    Models for the Intensity Function of an NHPP
    Rate of Occurrence of Failures
    Reliability Growth Models
    Simulation of Stochastic Processes

    Bayesian Reliability Evaluation

    Classical versus Bayesian Analysis
    Classical versus Bayes System Reliability
    Bayesian System MTBF Evaluations
    Bayesian Estimation of the Binomial p
    The Normal/Normal Conjugate Prior
    Informative and Noninformative Priors
    A Survey of More Advanced Bayesian Methods
    Gamma and Chi-Square Distribution Relationships
    Problems
    Answers to Selected Exercises
    References
    Index

    Biography

    Dr. David C. Trindade is the chief officer of best practices and fellow at Bloom Energy. He was previously a distinguished principal engineer at Sun Microsystems, senior director of software quality at Phoenix Technologies, senior fellow and director of reliability and applied statistics at Advanced Micro Devices, worldwide director of quality and reliability at General Instruments, and advisory engineer at IBM. He has also been an adjunct lecturer at the University of Vermont and Santa Clara University, teaching courses in statistical analysis, reliability, probability, and applied statistics. In 2008, he was the recipient of the IEEE Reliability Society’s Lifetime Achievement Award.

    "I have used the second edition of this book for an Introduction to Reliability course for over 15 years. … The third edition … retains the features I liked about the second edition. In addition, it includes improved graphics … [and] examples of popular software used in industry … There is a new chapter on Bayesian reliability and additional material on reliability data plotting, and repairable systems’ analysis. … The book does a good and comprehensive job of explaining the basic reliability concepts; types of data encountered in practice and how to treat this data; reliability data plotting; and the common probability distributions used in reliability work, including motivations for their use and practical areas of application. … an excellent choice for a first course in reliability. The book also does a good job of explaining more advanced topics … the book will continue to be a popular desk reference in industry and a textbook for advanced undergraduate or first-year graduate students."
    Journal of the American Statistical Association, June 2014