Issue 41

J.A.O. González et alii, Frattura ed Integrità Strutturale, 41 (2017) 227-235; DOI: 10.3221/IGF-ESIS.41.31 229 These figures clearly demonstrate that a partially closed fatigue crack can grow in the portion of its front that opens under tensile loads, while the portion that is under compression stands still. These bent cracks partially close their fronts under the compressive stresses induced by the bending loads, so they can verify how fatigue cracks grow under variable driving forces along their fronts. In fact, since the partial closure of the crack front is induced by compression caused by external bending loads, it occurs independently of Elber’s PICC or of any other type of internal closure mechanism. The cracks tested in such way severely distorted their initially (quasi) straight fronts while they grew after changing the load applied on the pre-cracked plates from pure tensile to pure transversal bending. Indeed, the successive crack fronts depicted in Fig. 2 clearly show that, although the pre-crack front was initially almost linear, it slowly assumed an increasingly elongated L- shape after propagating partially closed during its subsequent 2D FCG under pulsating bending loads. This is exactly the expected behavior of a crack front that gradually advances by fatigue influenced by the local value of the crack driving force, which clearly varies along it. Questions like that indicate that the hypothesis “  K eff controls FCG” cannot be accepted dogmatically, to say the least. Moreover, it is not evident that all closed fatigue crack tips always remain unloaded, let alone that they can only grow after fully opened. Albeit only careful K op measurements can identify if and how  K eff affects FCG, to prove PICC is their only cause it must be separated from other closure mechanisms whereas all other delay mechanisms must be ruled out as well. To question the actual role of  K eff in FCG certainly is a most important practical issue. Traditional da/dN procedures assume fatigue cracks loaded under fixed {  K , K max } grow under identical rates in any piece of a given material. The SIF- based similarity principle could not be used in FCG modeling if that was not so. However, identical loading conditions do not imply in identical crack opening loads, since P op may depend e.g. on the thickness and on the size of the residual ligament rl that forces the crack to close. So, whereas for a given geometry and crack size  K and K max can be calculated from  P and P max using catalog expressions,  K eff cannot. Hence, the use of  K eff to predict FCG can pose major problems in practical applications. In other words, albeit Elber’s PICC remains the most popular mechanism to explain load order effects on FCG under VAL, there are reasonable doubts whether it is the sole or even if it is the main one, which are explored in the following simply because they are too important to be ignored. Moreover, the main problem with this concept is how to use it in practice: there is no foolproof universal method yet to reliably calculate K op and  K eff in the complex structural components engineers must deal with. Since only simplified models are available to estimate K op values based on an idealized behavior of very simple geometries, this is indeed a major problem for  K eff -based FCG predictions. T ESTS TO VERIFY IF  K EFF IS THE ACTUAL FCG DRIVING FORCE lthough Elber’s PICC certainly is a plausible mechanism to explain many peculiarities of the FCG behavior, it has, at least, some limitations. Indeed, if  K eff controls FCG, and if closure is larger at the cracked piece surfaces, then the fatigue lives of thin pieces (with a large pz/t ratio) that work under pl-  FCG should be larger than the lives of similar but thicker pieces that work under equally fixed {  K , R } loading conditions in pl-  . Even under such simple conditions, if the crack starts to grow under pl-  , as it usually does, and gradually changes to a pl-  dominated stress state as its size increases, then FCG rates should vary in the same piece between these two limit cases. Hence, not even experimental da/dN   K eff data would provide enough information on the FCG behavior of structural materials. In fact, without the K -similarity it would be very difficult to reliably model FCG lives of most structural components even in very simple practical applications. Simple and easily reproducible FCG tests have been performed to verify the actual role of  K eff in FCG. Cracks were grown under quasi-constant {  K , R } in thin and thick DC(T) SAE 1020 steel specimens, carefully measuring FCG rates da/dN and crack opening loads P op along the crack path. The thickness t of the specimens was chosen to guarantee pl-  (making pz/t  1 ) in the thin and pl-  conditions in the thicker ones (making t > 2.5(K max /S Y ) 2 ), assuming this classic ASTM E399 requirement can be used in fatigue as well. All the DC(T) specimens were cut from an as-received 76mm wrought round bar with yield strength S Y  262MPa and ultimate tensile strength S U  457MPa .  K and R were maintained almost fixed adjusting the load at small crack increments, following standard ASTM E647 procedures. The crack length was measured by compliance and by optical methods, using a strain gage bonded on the back face of the specimens and a traveling microscope. In four tests, the crack opening load P op was redundantly measured by compliance techniques using linearity subtractor procedures [8, 10-11] on the signals of the back-face strain gage, and of a strip with several gages bonded ahead of the notch tip, see Fig. 3. In other four specimens, the opening loads were also simultaneously measured using digital image correlation (DIC) techniques, see Figs. 4 and 5. The P op values measured by A