Issue 41

J.A.O. González et alii, Frattura ed Integrità Strutturale, 41 (2017) 227-235; DOI: 10.3221/IGF-ESIS.41.31 228 that Elber’s plasticity-induced crack closure (PICC) is the sole or at least the dominant cause for all load order effects on FCG, whereas many others deny that PICC may be even relevant in such problems, or in FCG for that matter. Since this work does not aim to review or analyze such conflicting claims, see for instance [3-5] for a small sample of the arguments that support them. The objective here is to propose and analyze a simple and easily reproducible test that can clearly identify the actual role of crack closure in FCG, at least under constant amplitude loads (CAL). Kemp states that tests to support (or to deny) that the effective SIF range  K eff is (or is not) the FCG driving force should always include proper closure measurements [6]. He evaluates mechanical (compliance), optical, ultrasonic, electrical (potential drop), and metallographic techniques used to measure opening loads P op , and lists many closure mechanisms and models used to quantify it. He says that compliance techniques are the most reliable to measure P op , but claims that such tests must be carefully made and analyzed. He also says that, besides on P op , crack closure depends on many factors, such as the cracked component thickness, alloy strength, grain structure, crack morphology, load conditions, and/or the environment. Therefore, if FCG rates are really controlled by  K eff , crack closure dependence on too many factors could be a major issue for structural design and analyses, since FCG rate predictions cannot assume that similar {  K , K max } loading conditions induce identical opening loads (and thus  K eff ) in all cracked components. Moreover, if PICC is indeed the FCG driving force, due to variable plasticity-induced transversal contraction restrictions, K op should be larger and FCG rates should be lower in thin pieces, i.e. under predominantly plane stress ( pl-  ) conditions, than in thick ones, where most of the crack front grows under plane strain ( pl-  ). Likewise, OLs should affect much more the cracked piece surfaces, because their plastic zones pz OL are larger in pl-  than in pl-   regions along the crack tip. Some authors even attribute all or most OL-induced delay effects in thick pieces to their surface behavior [7-8]. However, not all load order phenomena can be well explained by surface PICC arguments. An important detail can illustrate this claim: how the delayed crack fronts can remain almost parallel after OLs, as they usually do. Indeed, if the crack is driven by  K eff , and if  K eff varies along the crack front due to higher OL-induced closure effects in pl-  regions near the cracked component faces, why then do the central parts of the crack fronts that grow under pl-  at a higher  K eff not propagate faster and gradually increase their curvature? In some cases they do, see Figs. 1-2 [9]. Figure 1 : Pre-cracked SE(T) specimen loaded in pure bending to partially close its crack. Figure 2 : Successive crack fronts propagated in transversal bending from an initially straight shape. Fig. 1 schematizes edge cracks initially grown in SE(T) specimens under pure tension loads at R  0.05 , generating approximately straight fronts. Then these pre-cracked specimens were repositioned and reloaded under 4-point bending (maintaining the same R ), to work as a flat beam with a side crack. Fig. 2 shows the resulting crack fronts.