Accelerated Testing: Statistical Models, Test Plans and Data Analysis [Kietas viršelis]

Preface		xi
Introduction and Background		1	(50)
Survey of Methodology and Applications		3	(9)
Types of Data		12	(3)
Types of Acceleration and Stress Loading		15	(7)
Engineering Considerations		22	(15)
Common Accelerated Tests		37	(6)
Statistical Considerations		43	(8)
Problems		49	(2)
Models for Life Tests with Constant Stress		51	(62)
Introduction		51	(2)
Basic Concepts and the Exponential Distribution		53	(5)
Normal Distribution		58	(2)
Lognormal Distribution		60	(3)
Weibull Distribution		63	(2)
Extreme Value Distribution		65	(3)
Other Distributions		68	(3)
Life-Stress Relationships		71	(4)
Arrhenius Life-Temperature Relationship		75	(10)
Inverse Power Relationship		85	(7)
Endurance (Fatigue) Limit Relationships and Distributions		92	(3)
Other Single Stress Relationships		95	(3)
Multivariable Relationships		98	(7)
Spread in Log Life Depends on Stress		105	(8)
Problems		107	(6)
Graphical Data Analysis		113	(54)
Introduction		113	(1)
Complete Data and Arrhenius-Lognormal Model		114	(14)
Complete Data and Power-Weibull Model		128	(6)
Singly Censored Data		134	(5)
Multiply Censored Data		139	(6)
Interval (Read-Out) Data		145	(22)
Problems		154	(13)
Complete Data and Least Squares Analyses		167	(66)
Introduction		167	(3)
Least-Squares Methods for Lognormal Life		170	(12)
Checks on the Linear-Lognormal Model and Data		182	(7)
Least-Squares Methods for Weibull and Exponential Life		189	(14)
Checks on the Linear-Weibull Model and Data		203	(7)
Multivariable Relationships		210	(23)
Problems		229	(4)
Censored Data and Maximum Likelihood Methods		233	(84)
Introduction to Maximum Likelihood		234	(8)
Fit the Simple Model to Right Censored Data		242	(13)
Assess the Simple Model and Right Censored Data		255	(10)
Other Models and Types of Data		265	(19)
Maximum Likelihood Calculations		284	(33)
Problems		302	(15)
Test Plans		317	(60)
Plans for the Simple Model and Complete Data		317	(11)
Plans for the Simple Model and Singly Censored Data		328	(21)
Evaluation of a Test Plan by Simulation		349	(12)
Survey of Test Plans		361	(3)
ML Theory for Test Plans		364	(13)
Problems		371	(6)
Competing Failure Modes and Size Effect		377	(48)
Series-System Model		378	(5)
Series Systems of Identical Parts		383	(2)
Size Effect		385	(2)
Nonuniform Stress		387	(5)
Graphical Analysis		392	(15)
ML Analysis for Competing Failure Modes		407	(6)
ML Theory for Competing Modes		413	(12)
Problems		417	(8)
Least-Squares Comparisons for Complete Data		425	(26)
Hypothesis Tests and Confidence Intervals		426	(3)
Graphical Comparisons		429	(5)
Compare Log Standard Deviations		434	(3)
Compare (Log) Means		437	(4)
Compare Simple Relationships		441	(4)
Compare Multivariable Relationships		445	(6)
Problems		448	(3)
Maximum Likelihood Comparisons for Censored and Other Data		451	(42)
Introduction		451	(1)
One-Sample Comparisons		452	(6)
Two-Sample Comparisons		458	(7)
K-Sample Comparisons		465	(5)
Theory for LR and Related Tests		470	(23)
Problems		488	(5)
Models and Data Analyses for Step and Varying Stress		493	(1)
Survey of Theory for Tests with Varying Stress		494	(1)
Step-Stress Model and Data Analyses		495	(11)
Varying-Stress Model and Data Analyses		506	(15)
Problems		513	(8)
Accelerated Degradation		521	(28)
Survey of Applications		521	(2)
Degradation Models		523	(11)
Arrhenius Analysis		534	(15)
Problems		544	(5)
Appendix A. Statistical Tables		549	(12)
A1. Standard Normal Cumulative Distribution Function Φ(u)		550	(2)
A2. Standard Normal Percentiles zp		552	(1)
A3. Standard Normal Two-Sided Factors Kp		552	(1)
A4. t-Distribution Percentiles t(P;v)		553	(1)
A5. Chi-Square Percentiles χ2(P;v)		554	(2)
A6a. F-Distribution 95% Points F(0.95;v1,v2)		556	(2)
A6b. F-Distribution 99% Points F(0.99;v1,v2)		558	(2)
A7. Probability Plotting Positions Fi = 100(i-0.5)/n		560	(1)
References		561	(18)
Index		579

Wayne B. Nelson, PHD, is a leading expert on analysis of reliability and accelerated test data. Formerly with General Electric Research & Development for twenty-three years, he now privately consults on and teaches engineering applications of statistics for many companies, professional societies, and universities. For his outstanding contributions to reliability data analysis and accelerated testing, he was elected a fellow of the Institute of Electrical and Electronics Engineers, the American Society for Quality, and the American Statistical Association. DR. WAYNE NELSON IS AWARDED THE SHEWHART MEDAL American Society for Quality awarded Dr. Wayne Nelson of Schenectady, New York the 2003 Shewhart Medal. The Medal honors his outstanding technical leadership, particularly for innovative developments and applications of theory and methods for analyzing quality, reliability, and accelerated test data, and for widely disseminating such developments through his books and many publications, talks, and courses. The Shewhart Medal for outstanding technical leadership is named after Dr. Walter A. Shewhart, who pioneered statistical methods for controlling and improving the quality of manufactured products. These methods contributed significantly to the United States' war effort in World War II. Subsequently taken to Japan by Dr. W. Edwards Deming, these methods revolutionized Japan's industries. Today these methods are part of widely used Six Sigma training on how to improve the quality of products and services. The American Society for Quality is the world's largest professional society dedicated to the improved quality of products and services. It serves its members and the public through a variety of educational activities, including conferences, training courses, journals, and books. Dr. Nelson is a graduate of the California Institute of Technology (Caltech) and the University of Illinois. Formerly with GE Research & Development, he now privately consults and gives courses for companies, professional societies, and universities. For his technical contributions, he was elected a Fellow of the American Society for Quality, the American Statistical Association, and the Institute of Electrical and Electronic Engineers. He recently spent four months in Argentina on a Fulbright Award, lecturing on analysis of product reliability data.