Download Probability Applications in Mechanical Design by Franklin E. Fisher and Joy R. Fisher easily in PDF format for free.
The main part of the book is Chapter 2, which applies probability and computer analysis to fatigue, design, and variations of both. The essence of this chapter is the ideas presented in Metal Fatigue (1959) edited by George Sines and J. L. Waisman and considers the problem of having to deal with a limited amount of engineering data.
The discussions of fatigue by Robert C. Juvinall in Stress, Strain, Strength (1967) and by J. H. Faupel and F. Fisher in Engineering Design (1981), as well as the books by Edward Haugen (1968) on the variation of parameters in fatigue, are successfully combined into a single treatment of fatigue. This book is an extension of Haugen’s book Probabilistic Mechanical Design (1980) with applications.
The concepts of optimization are developed in Chapter 3. The technique of geometric programming is presented and solutions to sample problems are compared with computer-generated non-linear programming solutions. Reliability, the topic Chapter 4, is developed for mechanical systems and some failure rate data is presented as it can be hard to find.
The book is influenced by the consulting work I performed at Hughes Aircraft Co. from 1977 to 1993. Some of the examples are drawn from this effort. Joy Fisher, worked in computer programming in the 1980s and 1990s keeping track of the changing state of the art in computing and writing for sections in this book dealing with programming This book was roughed out on a sabbatical leave in 1994 from class notes and in a summer institute taught by Edward Haugen in the early 1970s.
Credit also goes to many students from industry who labored to understand and use the information. The editorial and secretarial assistance of Ms. Cathy Herrera is gratefully acknowledged. Data for load carrying material properties can be modelled using any probability distribution function. Statistical goodness-of-fit tests should be applied to determine if the data set could be randomly drawn from that distribution.