Energy Balances for Kahn and M(T) Specimen The total external work is either stored as elastic energy or dissipated as work of plastic deformation or work of separation, W= Wel +Wpl +Wsep, (5) with . (6) Wel = σij Ý ε ij el dt 0 t ∫ V ∫ dV, Wpl = σe Ý ε e pl dt 0 t ∫ V ∫ dV, Wsep = B Γ0 ∆a The special advantage of the CZM is that the latter can be separated from the overall work of plastic deformation [12]. An analysis of these contributrions now shows, that plastic work constitutes more than 50 % and up to 70 % of the total mechanical work in the Kahn specimen whereas it is less than 20 % in the M(T) specimen, see Figure 4. Elastic energy amounts to 90 % in the M(T) specimen, but plays a minor role, 20 ÷ 40 %, in the Kahn specimen. Work of separation contributes only 4 % in the Kahn specimen and 0.6 % in the M(T) specimen at maximum crack growth. Thus, what is measured primarily in a JR-curve is plastic work in the Kahn and elastic energy in the M(T) specimen. No wonder that any attempt of transferring these data is senseless. And the really interesting material property of separation energy is not captured at all by the respective tests. The high amout of elastic energy stored in the M(T) specimen will result in a high influence of Young's modulus on the numerical result. As described above, the panels are not homogeneous but coated with a corrosion protection which comes up to 9 % of the total thickness. A parameter study showed that a reduction of Young's modulus by 9 % reduces maximum load by approximately the same factor. Any influence of the cladding on plastic deformation and dammage has not yet been studied. 0 10203040506 0 20 40 60 80 100 energy portion (%) ∆a (mm) M(T) specimen 0 1 2 3 4 5 6 0 20 40 60 80 100 ∆a (mm) Kahn specimen elastic plastic separation 0 Figure 4: Recoverable and dissipated portions of total external work for Kahn and M(T) specimens MODELING OF SLANT FRACTURE As mentioned above and shown in Figure 1, the actual crack plane shifted from a normal to a 45° inclined orientation, which was not accounted for in the two simulations. Damage models as well as cohesive zone models are capable of simulating slant fracture under appropriate modeling conditions, namely • a full 3D analysis is necessary and no symmetry conditions must be imposed in the ligament, • the number of elements over the thickness has to be large enough for the models of continuum damage, • a mode III component has to be added to the separation law and cohesive surface elements have to be placed along the faces of tetrahedral 3D elements to allow for 45° crack paths.
RkJQdWJsaXNoZXIy MjM0NDE=