PSI - Issue 5

Jan Raška et al. / Procedia Structural Integrity 5 (2017) 225–232 Jan Raška / Structural Integrity Procedia 00 (2017) 000 – 000

228

4

On the contrary, the application of the special CRAC2D NASTRAN element [6], developed for the stress intensity factor – K I and K II – computing, is very fastidious. The CRAC2D element replace 4 elements, neighboring the crack tip (yellow ones on the figure 2). Therefore, at the first step, this 4 elements are removed from de FE model . At the second step, the CRAC2D is created, as demonstrated on the figure 3. However, this operation can not be made by the help of current commercial FEM pre- postprocessors. In the practice, the user have to add the several lines directly in the input NASTRAN text file to define the CRAC2D element (example on the picture 4). This last operation is very fastidious and above all, generate many potential errors (typically, node numbering errors, sequence node errors, etc.).

Fig. 4. CRAC2D definition in a text file

This procedure leads to only a single value of the stress intensity factor – K I and K II – for defined crack path. In practice, the stress intensity factor computing must be repeated manifold (often hundredfold and more – see fig. 1). The above-mentioned manual procedure is very time-consuming and error susceptible. For this reason, the automation of this procedure was realized and the appropriate software was developed. 3. Software principle and use The fundamental data input of the software is the virgin FE model – it means the FE model without modeled crack. However, this FE model must be adapted so, that the crack path pass on the element borders. Moreover, in the neighborhood of the crack path, only the 4-node, linear shell elements (QUAD4 in NASTRAN) must be applied (see figure 5). The second basic data input is the crack path definition. For this goal, the user has to numerate appropriately the nodes at the crack path. Logically, the numbering must be consecutive from the crack beginning (initiation) to the maximum crack opening. In the example on the figure 5, the crack is initiated on the node 1000 and is opened step by step up to the node 1005. This figure 5 demonstrate the practical manner (but not the necessary condition) to do it: The node numbers on the crack path are much greater, then the rest of the node numbers.

Fig. 5. Input FE data – crack path definition [3]