Page 264 - T. Anderson-Fracture Mechanics - Fundamentals and Applns.-CRC (2005)
P. 264

1656_C005.fm  Page 244  Monday, May 23, 2005  5:47 PM





                       244                                 Fracture Mechanics: Fundamentals and Applications


                          There are two major problems with the weakest link model that leads to Equation (5.27a) and
                       Equation (5.27b). First, these equations predict zero as the minimum toughness in the distribution.
                       Intuition suggests that such a prediction is incorrect, and more formal arguments can be made for
                       a nonzero threshold toughness. A crack cannot propagate in a material unless there is sufficient
                       energy available to break bonds and perform plastic work. If the material is a polycrystal, additional
                       work must be performed when the crack crosses randomly oriented grains. Thus, one can make an
                       estimate of threshold toughness in terms of energy release rate:

                                                         G    ≈ 2γφ                              (5.28)
                                                          c(min)  p

                       where φ is a grain misorientation factor. If the global driving force is less than G c(min) , the crack
                       cannot propagate. The threshold toughness can also be viewed as a crack arrest value: a crack
                       cannot propagate if K  < K .
                                            IA
                                        I
                          A second problem with Equation (5.27a) and Equation (5.27b) is that they tend to overpredict
                       the experimental scatter. That is, scatter in experimental cleavage fracture toughness data is usually
                       less severe than predicted by the weakest link model.
                          According to the weakest link model, failure is controlled by the initiation of cleavage in the
                       ferrite as the result of the cracking of a critical particle, i.e., a particle that satisfies Equation (5.17)
                       or Equation (5.18). While weakest link initiation is necessary, it is apparently not sufficient for
                       total failure. A cleavage crack, once initiated, must have a sufficient driving force to propagate.
                       Recall Figure 5.22, which gives examples of unsuccessful cleavage events.
                          Both problems, threshold toughness and scatter, can be addressed by incorporating a conditional
                       probability of propagation into the statistical model [42, 43]. Figure 5.26 is a probability tree for
                       cleavage initiation and propagation. When a flawed structure is subject to an applied K, a microcrack
                       may or may not initiate, depending on the temperature as well as the location of the eligible cleavage
                       triggers. The initiation of cleavage cracks should be governed by a weakest link mechanism, because the
                       process involves searching for a large enough trigger to propagate a microcrack into the first ferrite grain.

































                       FIGURE 5.26  Probability tree for cleavage initiation and propagation.
   259   260   261   262   263   264   265   266   267   268   269