OOF Workshop summary, August 24 & 25, 2006 The OOF workshop held at NIST on August 24 and 25 was attended by 34 people from 5 US industries, 3 US government laboratories (including NIST), 13 US universities, and 5 foreign universities. PRESENTATIONS The workshop began with presentations from developers about OOF2: what it does, how it differs from OOF1, a worked problem, and and overview of how to install and extend it. Introductions -- Steve Langer, NIST OOF2 Overview -- Andrew Reid, NIST An OOF2 Example Problem -- Steve Langer Installing and Extending OOF2 -- Steve Langer The afternoon session consisted of eight contributed talks on applications of OOF1 and OOF2. The applications discussed included Li-ion batteries, light emitting semiconductor devices, partially molten rock, silicon carbide composites, thermoelectric materials, and functionally graded materials. Particularly notable is that the thermoelectric calculations were done with an OOF2 extension written at Purdue. Allowing this kind of easy extensibility has been one of the goals of the OOF2 project. OOF Extensions and Applications to Multifunctional Materials and Devices: An Overview, Edwin Garcia, Purdue University Combining front-tracking with finite elements: Elle and OOF2, Jens Becker, University of Tuebingen, Germany OOF and Beyond: 2D Plasticity and 3D Microstructure-based Modeling, Adarsh Ayyar, Arizona State University Teaching Computation and Modeling in Materials Science, Shahriar Anwar, Arizona State University Microstructure Effects of Thermoelectric Nanowire Composites, Mara Howell, Purdue University Effect of External Fields on Microstructures: A Modeling Perspective, Nitin Singh, University of Virginia Functionally Graded Materials, Antonella Sola, Universita degli Studi di Modena e Reggio Emilia Residual Stresses in Binary Eutectics, Michael Muscarella, Penn State University A presentation on "Advances in Dental Research with OOF" was cancelled for medical reasons. Thursday afternoon ended with a discussion of future development, giving workshop attendees a chance to provide feedback to the developers. On Friday (after a rough start due to network problems caused by a major power outage at NIST) participants used OOF2 in the CTCMS computer room, working through the OOF2 tutorials or meshing micrographs of their own. A major goal of the workshop was for the developers to get feedback from the users about what features are important to them. This will help to guide future development. Here is a list of issues raised, with comments about the status of each: Material Properties: Strain-dependent elastic modulus and temperature-dependent conductivity. These can easily be added as OOF2 extensions. Polymer systems -- strain rate, strain hardening, and large deformations. These require improvements in OOF2's handling of time-dependent and non-linear properties, but are planned for version 2.1. Finite-Element and other Numerical Issues: Three dimensional calculations. This has always been on our agenda and work has already begun. Non-zero values for out-of-plane fluxes (eg, stress and heat flux). This will be straightforward to implement. Dividing a problem into subproblems which can be solved separately. This ability is planned for version 2.1. Lagrange multiplier fields. This is planned for a future release, possibly 2.1. Lagrange multipliers will allow OOF to use periodic boundary conditions and other constraints. (There are other ways of implementing periodic boundary conditions, though.) Interfacial elements and contact elements. Interfacial elements will be included when we allow subproblems. This will allow surface tension to be specified as a material property. Axisymmetric elements. The general feeling of the OOF developers is that microstructures aren't often axisymmetric, so there's not much point in including machinery for dealing with them. If it's easy to do we might do it anyway. Remeshing for large deformations. This will be part of the non-linear, time-dependent improvements. Conditional stopping criteria and other means of controlling time-dependent simulations. This should be feasible. Interface Issues: Material editor to write extension modules, and an output editor to build new output quantities graphically. These are big projects that may or may not be feasible. We'll think about them. Automatic generation of movies from time-dependent simulations or incremental distortions. This should be possible. Configurable keyboard shortcuts. This may already exist in the gtk+ library that we're using. It would be easy to add in any case. Hide additional GUI complexity. As the program gets bigger, the interface gets messier. Users should be able to choose which parts of the interface they see. Direct input of OIM (orientation imaging microscopy) data. This should be possible. We should write a general API so that users can easily write input extensions for OOF2, so that any data format can be used. Draw vector fields with arrows in the graphics window. This should be fairly easy in the current framework. Computation of area changes. This is possible currently, but not in an easy way. We should fix that. Renaming material properties. This is also currently possible, but must be done in two steps. We may simplify it, at the expense of cluttering the user interface. Other Issues: Microsoft Windows. OOF2 doesn't yet run on Windows, but we're working on it. We are making no promises, though. Quick and easy ways of automating calculations. A one-button method for going from a microstructure to an effective modulus would be nice, but it's not clear if it's feasible in any kind of general setting. This is a big project, and something that should be investigated. Package managers. Installing OOF2 is moderately complicated because of the 3rd party libraries and tools that it requires. This could be made easier by using a package manager. Maintaining packages for all of the different managers would be a headache, though. It was also suggested that the OOF website could distribute the required libraries directly, instead of linking to external sites. An OOF users web site (wiki) for exchanging extensions, material definitions, and tips. It's difficult to set up a wiki at NIST for security reasons. Craig Carter has offered to host an OOF wiki at MIT. CONCLUSIONS The workshop introduced new users to OOF2. Some participants were new to the OOF project, but many were users of OOF1, interested in making the transition to OOF2. Participant feedback largely ratified the goals of the development team, and indicated an appreciation of the modular, open-source nature of the OOF software. Many participants indicated that they planned to introduce their colleagues to the OOF software on their return to their home institutions. During the Friday tutorial phase, participants were also able to make contacts among themselves. Along with renewed awareness of the OOF mailing list and planned construction (encouraged by users) of an OOF wiki, these steps should encourage the growth of a self-sustaining OOF user community. ACKNOWLEDGEMENTS The workshop would not have been possible without financial support from the NIST Mathematical and Computational Sciences Division in the Information Technology Laboratory, and the NIST Center for Theoretical and Computational Materials Science in the Materials Science and Engineering Laboratory. We especially thank Robin Bickel and Susan Roth for their extraordinary administrative support. Steve Langer Andrew Reid Rhonald Lua