Issue 41

D. Nowell et alii, Frattura ed Integrità Strutturale, 41 (2017) 197-202; DOI: 10.3221/IGF-ESIS.41.27 200 Our own work has so far focused on simpler one or two parameter models, and our view is that the introduction of different forces with different notations and signs in Fig. 3., whilst intended to link the model to physical phenomena, is not particularly helpful. What is helpful, however, is the concept of a plastic wake and its effect on the surrounding material. Hence, there is no need to introduce different y-direction forces at minimum load, but one can simply stick with the system of forces shown in Fig. 3a and drop the double subscripts. So, F y causes crack opening and a stress intensity, F s , exerts shear on the areas above and below the crack, and this is reacted by F x , which may be thought responsible for the bounded stress component in the x-direction (T-stress). These three forces neatly map onto the four terms tabulated on the previous page: F S Is responsible for K S . Physically, its link to crack propagation rates is difficult to see, so that one might consider it, at most a secondary effect. F x Is responsible for the T-stress. Again, a secondary effect, though perhaps more important than K S . F y Must clearly be responsible for K F and K R , and which may be separated using the CJP approach, or if preferred may remain as a single K I term. E XPERIMENTAL RESULTS AND DISCUSSION e have essentially already extracted the dominant K I term in the above simplified version of the CJP model, and an example result has already been presented in Fig. 2. To compare this with the results of the full model, we will use some results produced by Vasco Olmo [8]. He used a full-field DIC technique to extract K F and K R for the CJP model, using essentially the same material (Al4%Cu) and specimen geometry (Compact Tension) as in our own work. Figure 4 gives his results for a test conducted a load ratio, R = 0. It can be seen that the measured K F value is very close to the nominal elastic K, calculated using the usual standard solution. The K R value starts close to zero, but then becomes negative, and increases in magnitude until the peak load, decreasing again during unloading. Figure 4 : Results obtained by Vasco Olmo [8], showing the variation of K F and K R through a load/unload cycle for an Al4%Cu CT specimen. Results are normalized with respect to the nominal elastic K. If we now choose to plot a single crack driving force, K F + K R , the results are transformed to those shown in Fig. 5. Finally, we note that in our own work, the datum for displacements was the unloaded specimen with the crack present , so that we are unable to detect any pre-existing residual K. Hence the appropriate parameter to plot is  ( K F + K R ), and this is given in Fig. 6, along with our own experimental data (Fig. 2), re-plotted on the same axes. The comparison between the two sets of data, obtained independently on two different specimens in two different laboratories is striking, particularly when one considers that different DIC algorithms are used, and different post-processing routes were adopted. Comparison of these two sets of data, suggest that the four parameters of the mode I CJP model may usefully be reduced to three if one combines the K F and K R parameters, and that the three remaining terms each map clearly onto the effects of a force transmitted across the interface between the plastic zone and resulting wake, and the surrounding elastic hinterland. When viewed from this perspective, the combined K F + K R parameter appears to be very similar to the measured delta K in our own experiments carried out under similar conditions. Both our own laboratory and that of Vasco Olmo and Diaz Garrido in Jaén have a wealth of similar data, carried out for different loading conditions and specimen geometries, and more time is needed to make further comparisons of a similar nature. However, before we do so, it would be useful if the community could take a view on whether the full complexity of the W