James, > I am a graduate student who will be defending in the not-too-distant > future, and afterwards I want to remain in the research community even > though I will be leaving academia for the foreseeable future. My research > group has an established cohesive-zone model (CZM) code that is used in > Abaqus, and I have been using this for all of my graduate work. The main > drawback to not switching over to deal.ii years ago was that it lacks the > ability to model crack-propagation (which is the main focus of my > research), and the resistance to changing from a code the group knows > already works. > There are at least some papers that use deal.II for crack propagations, e.g. "A primal-dual active set method and predictor-corrector mesh adaptivity for computing fracture propagation using a phase-field approach <https://www.sciencedirect.com/science/article/pii/S0045782515001115>".
With this in mind, would a user/developer be willing to guide me along the > path of writing a new FiniteElement? I have already written initial > traction-separation laws (in C++) that could be used in a "volume" > cohesive-zone element. I have collected papers of others who have already > written cohesive-zone finite-elements for the commercial software package > Abaqus, and my intention would be to replicate this work (with perhaps some > small modifications/improvements) for deal.ii so that I can continue work > in fracture mechanics. > It is probably best if you describe what you need and how your new finite element would look like. In general, you need to define a suitable (polynomial) ansatz space and how the degrees of freedom are evaluated. In case, the element uses tensor-product polynomials there are alreay at lot of prerequsites available. You might one want to look into the implementation of FE_Q (https://github.com/dealii/dealii/blob/master/source/fe/fe_q.cc) > Additionally, does deal.ii have the capabilities of the so-called A-FEM > (augmented finite element method)? A-FEM allows one element to change an > element from one finite-element into another finite-element during the > calculation. An example from fracture mechanics: starting with an initially > isotropic elastic calculation of generic shape, when a certain stress > threshold is met by an element, the elastic element is replaced with a CZM > element to allow for arbitrary crack-initiation and propagation to occur. > Note: other FEM-based methods appear to use this same idea of replacing one > element type with another during a calculation, but I am do not have > extensive knowledge of the nuanced differences between them. > deal.II support hp-adaptivity that requires the possibility to use different finite elements across the discretization. Have a look at https://www.dealii.org/current/doxygen/deal.II/step_27.html. Best, Daniel -- The deal.II project is located at http://www.dealii.org/ For mailing list/forum options, see https://groups.google.com/d/forum/dealii?hl=en --- You received this message because you are subscribed to the Google Groups "deal.II User Group" group. To unsubscribe from this group and stop receiving emails from it, send an email to dealii+unsubscr...@googlegroups.com. For more options, visit https://groups.google.com/d/optout.
