Issue 33

J.M. Ayllon et alii, Frattura ed Integrità Strutturale, 33 (2015) 415-426; DOI: 10.3221/IGF-ESIS.33.46 424 The SIF calculated is in mode I in the local reference system of the crack. Therefore, at the centre of the crack there really is a mixed mode growth of mode I and mode III because the crack is inclined an angle of 11.4º. Nevertheless, this angle is so small that life estimation calculations have been performed using only mode I because the contribution of mode III can been neglected, as can be observed in Fig. 12. L IFE ESTIMATION AND COMPARISON OF RESULTS inally, a comparison of the lives estimated by applying the TCD and VIL models, together with the results obtained experimentally is shown in Fig. 13. From a general point of view it can be said that both models have a reasonable behaviour. As can be seen in Fig. 13, the VIL model presents a better performance in estimating the fatigue life of the dental implant system, being able to correctly reproduce the slope of the fatigue curve and the presence of fatigue limit, although in this case the model seems to underestimate this limit. The latter is a conservative result, while it has to be experimentally confirmed that the actual fatigue limit is located where the only trial that has been able to withstand 5 million cycles suggests. It is important to highlight that the entire process of crack initiation and propagation takes place within about half a millimetre, and that the model is able to reflect this in an acceptable manner. On the other hand, the TCD model overestimates the life and not collects properly the slope of the fatigue curve. Fig. 14 shows a comparison between the initiation length obtained using the VIL model and the critical distance calculated through the TCD. Although the values are similar, the meaning of these parameters is very different. The application of both models to other situations should show if this similarity is a coincidence or if there is a relation between them. As a result of the VIL model application, it has also been observed that in the short/medium fatigue regime (lower than 10 5 cycles), the duration of the crack initiation phase is negligible compared to the total fatigue life. This is due to the severe notch existing at the thread root of the implant body, which causes the crack to initiate within the first loading cycles. Figure 13: Fatigue tests in implants and theoretical estimates. Figure 14 : Initiation length and critical distance vs. estimated life. C ONCLUSIONS he results of this study show that the VIL life prediction model, also used in various other situations, is very versatile and robust, adapting to different circumstances. It is important to emphasize that this model fits the life prediction of implants, elements which, given their small size, can present problems of scale. Once fatigue and fracture properties of the material are known, using this model allows estimating life without presupposing the relative importance of the stages of initiation and crack propagation. The applicability of each model can be compared. Regarding the material properties needed, the VIL model requires an elastic-plastic FE analysis to determine the stresses and the strains. In this FE model, for example, a kinematic hardening curve is needed to describe the plastic behaviour of the material. Also, the crack growth properties (Paris parameters) of F T