Abstract.A novel and reliable theoretical model based on the Birnbaum-Saunders (BISA) distribution is presented from which the fatigue life can be determined. Experimental verification of the model is in progress and will be published in due course.
IntroductionAccording to a poll conducted in 1985 by the American Society for Quality Control, reliability was the second most important attribute among the ten most important products attributes [1]. This result is expected as the impact of product failure can range from minor injuries and/or loss to sever injuries/death and/orloss. Therefore, the precise prediction of failures can save lives and money. Usually, the reliability of the products is determined through one of several types of life testing. The primary objective of these tests is to quantify the reliability of the product, which can be used to determine whether a set of goals for the product are met or not. Typically the result of the reliability test is a set of failure times that is analyzed statistically to predict the reliability distribution through curve fitting. One of the disadvantages of this method is the high cost and time in some circumstances especially when regular life tests are used. Another disadvantage is the low accuracy of the test when accelerated life tests are used due to extrapolation. Moreover, this approach pays no attention to the strong relationship between the mechanical properties and the reliability distribution of the product as it uses parametric or nonparametric statistical methods to predict the reliability from the failure data and not from the mechanical properties of the product.Little work has been done to predict the life of products through methods other than life testing. Good references for fatigue life prediction methods can be found in [2]. Ref [3] predicted the fatigue life of the tibial component of a polymeric PMMA space rutilizingmultiaxial fatigue coupled with computational simulations and material properties. Ref [4] used the S-N curve and the crack growth rate curvet predict the fatigue life. Ref [5] proposed a method to predict the fatigue-life based on linear elastic analysis and used it to predict the fatigue for different materials subjected to constant amplitude multiaxial proportional loading. Ref [6] Used ΔK-N curve utilizing finite element analysis to predict the fatigue life of spot welds starting from coarse finite element meshes and ending at one unique ΔK-N curve.Among the little available non life-testing prediction methods, prediction through fitting the mechanical properties to the life distribution parameters' was absent. The primary objective of this paper is to propose a method for fatigue failure mode life prediction of carbon steel shafts through their mechanical properties. The primary advantage of this method is in saving time and money. Once the relationship between the mechanical properties and the reliability distribution parameters for the