This year, the contest is looking into the visualization-supported discovery of novel recycle-based metallic materials: aluminium alloys for combustion engine pistons are high-performance materials with complex multi-element chemical compositions. Due to the current shift towards e-mobility, a large amount of pistons made of these alloys will become available as combustion engines reach their end of life. As a result, in the upcoming years these alloys, with their broad range of compositions and a myriad of valuable chemical elements, will represent a large and diverse materials source that can be exploited for developing new recycle-based alloys. The hypothesis in terms of the material discovery challenge is that a mixture of several scrap materials (3-6 alloys) can yield a sustainable path for novel alloy designs with superior properties across different application fields.
Challenge 1
Material Scientists are increasingly employing simulations to predict the properties of novel alloy candidates using high-throughput combinatoric approaches. These simulations cover a large search space of chemical compositions and materials properties and, therefore, researchers have to deal with multidimensional data to find optimum and robust solutions. Currently, mainly parallel coordinate plots as well as scatter plots are used to get an impression of the composition–microstructure–properties relationship of the alloys designed virtually, i.e., only selected input and output parameters are visualized to find a correlation.
Task 1
For a proper global determination of interactions between all considered parameters, more sophisticated multi-dimensional visualization techniques are required. Contestants are expected to develop multi-dimensional data visualization methods to generate overview on all simulated candidate materials. Research directions could, e.g., follow the paths of dimensionality reduction and visualization of respective embedding spaces or provide suitable means to visualize correlations in in- and outputs as well as the sensitivity of specific inputs on respective outputs.
Task 2
In order to help reveal promising material design candidates explorative visualization techniques need to be provided which facilitate the exploration and comparison of candidates with respect to composition – microstructure – properties of the alloy compositions (i.e. between input and output data) generated from scrap mixtures.
Challenge 2
Due to the complexity of the simulation and respective computational efforts, only subsets of combinations can be sampled. The complexity of respective simulations is even further elevated when mixing more scrap materials. Subtle changes in the material composition may yield strong impacts in the simulated parameters. With the target of identifying chemical compositions that possess ideal target parameters for a specific application, optimization algorithms are used. In this challenge (intermediate) results from the multi-dimensional visualization and analysis should therefore be used to communicate with respective optimization algorithms.
Task 1
For the exploration of alloy combination methodologies of visual steering should be explored. Interactive tools should be developed and applied to help understand how to determine the direction in which the optimization should go. Related research targets for this purpose could, e.g., be the exploration of the sensitivity of inputs versus the generated output or the stability / volatility of output regions with respect to their inputs.
Besides the overall first, second, and third place winners, this year’s jury will additionally award prizes to the:
- best report cover visualization
- most innovative approach