Industrial Demonstrators (Demo)

- Programme head: O-G. Lademo

Background
The research areas defined in the Centre address the fundamental and generic aspects of the behaviour and modelling of an impact loaded structure, i.e. material models and response characteristics of generic components and joints, with emphasis on numerical solution techniques. In real structures a wide range of loading modes, materials and types of connectors has to be considered. Furthermore, each component might have been subjected to a thermomechanical process in the form of shaping and ageing, the effect of which must be captured in the numerical model. The applicability and feasibility of the various models can only be assessed when tested on full-scale industrial systems, here denoted demonstrators. The main objectives of this research area are: 1) to establish a link between the basic research and real structures for validation and possible refinements of the developed technology and 2) to facilitate industrial implementation of the developed modelling concepts.

SIMLab Material Model library and MatPrePost
Some principal outcomes of the research in the Centre are numerical implementations of constitutive models and failure criteria. The annual report for 2009 presents the developed model library, consisting of models for metals, polymers, solid geomaterials, crushable foams and castings along with tailored development on the associated solution techniques. Interest is expressed by several of the consortium partners to make use of the material models in various industrial contexts and in various FE programmes. A project was initiated in 2010 to maximize the industrial benefit of the researchbased models, for minimized efforts for both SIMLab personnel and the industrial partners. In other words, the overall aim of the activity is to bring the SIMLab models to a higher (or the highest) ‘technology readiness' level.

In 2010 focus has been paid to the various, and so far co-existing, metal plasticity models within the SIMLab model library (WTM/STM-2D, GSTM, MJC, HAZ-2D). It was decided to restructure and combine these models into a single, highly versatile model thus being applicable to a broad range of mechanical problems. In order to do this a rewrite of current source codes was called for. In parallel, a re-structuring of parameter identification and visualization procedures was initiated.

Solution strategy and product interdependencies:
A value chain for non-linear numerical analyses was chosen, as illustrated in the figure below. Four principal steps are distinguished, as seen along the timeline axis (abscissa). In short these are: 1) Experimental testing, 2) Parameter identification, 3) Numerical analysis and 4) Post processing. The desired outcome of the chain is processed results that allow for proper engineering decisions. It is important to optimize the individual elements of the chain, e.g. appropriate experimental tests must be defined, proper identification procedures must be available, accurate model representation of physical response and supportive postprocessing must be facilitated. The optimization of the chain, as a whole, must also be ensured, in the sense that the ‘pre-fabric results' are fed efficiently along the chain.

As illustrated in the figure, three interdependent products will be developed:

ResOrg:
‘Result Organizer' that supports experimental planning, execution and processing. This tool is, in principal, built since the initiation of SIMLab in 2007.

MatPrePost:
Tool for parameter identification and tailored pre- and post-processing. MatPrePost is developed based on different in-house programs and spreadsheets (as indicated by the box ‘Untill 2010' in the figure), and tailored for efficient industrial application. The outcome of the pre-processing utility are for instance visualizations of the model concept, predicted Forming Limit Diagrams (FLDs) and fracture locus plots, and formatted input for the user-defined material model. This tool will at the end support output to several FE codes.

UMAT(s):
User-defined material model able to represent the physical phenomena of the engineering material in question. As stated earlier, we choose to define customized material models for the individual materials classes. The focus in 2010 has been on the metal plasticity model, which should be valid for rolled and extruded metals and alloys. The FE-code LS-DYNA is used as development platform, but the UMAT (as well as MatPrePost) will be adapted also for use with other FE-codes.

Status:
A ‘beta-version' of ‘MatPrePost' is established and released for internal use and evaluation. The preliminary appearance of the tool is illustrated in the next figure by some selected interface windows. A modular source code for the metal model has been written which so far includes a limited number of options. This work will be continued in 2011.