TABLE OF CONTENTS We have already encountered the problem of uncertainty several times during the course of the present text. In Chapter 3 we discussed and modeled the scatter in S-N data, in Chapter 4 we treated the fatigue loading as a stochastic process. In Chapter 5 we discussed the uncertainty in the Miner s summation rule and the problem of accuracy when determining the geometrical stresses by finite element analysis (FEA). The highly variable and irregular local weld toe geometry was also discussed. In Chapter 6 we presented the probability distributions for the growth parameter C in the Paris law, as well as for the initial crack depth. Finally, we briefly touched the probability of detection (POD) concept for characterizing the performance of a given inspection technique. All these issues underline the need for stochastic modeling when describing the fatigue behavior of welded joints. Stochastic modeling means taking account the various sources of uncertainty in a logical, rational, and consistent manner. Our tools are statistical methods and probability calculations. In this chapter we shall go into more details about the methods of stochastic modeling. If the analysis is concerned with the probability of survival (or failure) only, the calculation scheme is called reliability analysis. If the probability of failure is weighed against the consequences of a fracture it is called risk analysis. The latter concept is now extensively used for risk-based inspection planning.
The objective of the present chapter is to...
