Reliability Modelling: A Statistical Approach

Rating:
Summary: Excellent coverage, some flaws
Review: Based on coverage, I'd give this a 4-star rating--it treats all the basics: life distributions, hazard function, fitting to empirical data, load/strength modeling, Markov processes, repairable systems, system reliability (cut/tie sets, fault trees, etc.), maintainability, and so forth; as well as a nice selection of advanced topics. Almost all the technical material relevant to the ASQ Relibility Engineering Certification (and the UK and European equivalents) is there. At about 250 pages, it's quite concise.

There are some notable flaws: Enough misprints in equations to create occasional, but serious, problems for the reader; the presentation is uneven--some mathematical results are derived in fair detail, others are simply presented with no explanation. I was unsure, in many places, whether I was at fault for missing the derivation, or whether there was a misprint (which happened in a number of cases), or whether the author was citing a result without proof. There are many examples, in a few cases confusing because the figure accompanying the example is wrong. The book has no exercises, so an instructor using it as a textbook should expect to invest time in constructing them. There are no tables, so a handbook of statistical tables is a necessary corequisite. (There are a few extracts from tables in the text.)

The publisher's summary has some misleading statements: though MINITAB was apparently used in preparation of the book, there is no explanation or code given. This was not a problem for me, since I don't use MINITAB; but anyone expecting to find details on MINITAB usage for reliability analysis will be disappointed. The required mathematical background is understated--though probability and statistics concepts are reviewed, someone who has never taken a P&S course should not undertake this book. Certain topics, such as the use of Laplace transforms in solving differential equations, are well beyond the advertised prerequisite of "early undergraduate mathematics" (at least in the U.S.--perhaps UK undergraduates are better prepared).

I would hope that in a future edition, the author could rectify some of these flaws, in which case the book would deserve a higher rating. In particular, it would be easy to set some of the examples, and all the missing derivations, as end-of-chapter exercises. For now, the reader with enough patience can expect to learn a lot from this book, but suffer some frustration in the process.

Rating:
Summary: The most readable book on reliability modeling
Review: This highly-readable book should serve as an excellent introduction to reliability modeling for engineers and for students of statistics as well as for students intending to specialize in the field. The author gives a reasonable sample of the most important topics in a relatively short space. This includes some discussion of sometimes-difficult issues about choosing the "right" model. My only quarrel with the author is that she has not quite made the transition away from a reliance on published tables to a fully-computer-oriented approach. In the next edition she should present algorithms for solutions of important quantities such as confidence intervals. Nonetheless, I strongly recommend the book.