Sometimes you need to get away from the everyday interactions in a project to see what it achieved. Preparing a presentation on what the group has done recently was a fantastic opportunity to review our recent history, and caused me to think about where all this was all heading.
Multi-scale analysis is a continuing trend in our industry (I’ve blogged about it before). It’s a very interesting trend, and very much, from my point of view, a work in progress. There are numerous approaches and a wide range of people and organisations ploughing different furrows. It’s a classic example of a new technology developing; I don’t even think we are at the VHS vs Betamax stage yet.
My views on topology optimisation have probably been aired here before. (For the record I’m not sure that it’s universally applicable, and it’s certainly not new.) But these views aside, thinking about the development of multi-scale modelling, topology optimisation, especially when we talk about lattices, and additive manufacturing reveals a coherent and fascinating direction of travel.
With multi-scale simulation methods, we have the ability to look into the material response of components at a level of detail previously impossible. And not only can we look at what happens to a single material – we can see how the interactions within complex mixtures influence overall structural response. It’s bi-directional in the most comprehensive simulations: we can see how the complex material influences structural response, and we can see how the load on the structure manifests itself in the material at what is called the meso level. FE2 is the buzzword. The project that sparked off this train of thought was undertaken by SSA/TECHNIA for a customer through our Dutch sister organisation Simuleon. In it we created an approach, and some software, that allowed users to specify how a road surface was made up from its usual component materials. The really clever bit was breaking out of the scanned data loop by creating idealised RVE’s (essentially finite element models of small material samples) controlled by parameters such as particle size, mix ratio and constituent properties. And these definitions could be used to assess how the overall material would respond to the load of a vehicle tyre. The application of these techniques to composites, both long and short fibre, are obvious.
Understanding the influence of small scale structures within what we usually just smear into a simple description is critical in the success of loadbearing additive manufactured parts. Little wonder then that a significant focus of multiscale modelling is simulating how the process of additive manufacture affects the crystal structure of materials, and whether there will be voids, internal crack and the like, and how these influence part serviceability.
All this raises the question “where does material start and structure and geometry end?” With the ability to create and optimise fine structures at a lattice level we can effectively define the local structure; really, we are engineering the material. 3D printing fibre reinforced structures has also become a reality, but meaningful simulation of these structures needs multiscale approaches, and the same can be said of short fibre reinforced injection moulded parts, which are beginning to find wide application in automotive applications.
It looks like the next art to become possible will be the linking of scales automatically. (And it’s possible that CFD is ahead of the game here, as different length scales are automatically catered for in many turbulence models.) I’m not entirely convinced that all structural analyses will be FE2 any time soon, in fact it’s likely that routine design will never require this level of introspection, but in demanding scenarios where ultimate performance is necessary engineering the response down to material level is likely to be an essential part of the design process.
Stuart Nixon from Dassault Systems will be presenting on the subject of multiscale analysis at the SIMULIA Symposium 2019 on Wednesday 19th June at the NCC, Solihul.