Topology Optimisation Limitations and Future Expectations

Due to the free and organic forms that can be created by TO (Figure 1), some are often difficult to manufacture through traditional subtractive or formative manufacturing techniques. Thus an integration with manufacturing constraints must be considered for the feasibility of the geometry production. As an alternative to overcome this limitation, additive manufacturing[1] is a promising alternative.

According to Zhao, Li & Liu (2017):

…On the other hand, rapidly developing additive manufacturing technologies have the promise to overcome the barrier between the potentiality that the topology optimization approaches can provide and the limitations that traditional manufacturing technologies can fabricate. In reality, additive manufacturing is a natural counterpart to topology optimization in that they have very versatile capability to quickly generate and realize new components not existing before… (p.1) However, is important to consider that the additive manufacturing has its limitations, such as the necessity of supporting material or anisotropic mechanical behaviour, depending on the technology used. Therefore, new constraints on algorithm must be applied to additive manufacturing seeing the good practices of DfAM[2].

  

               

Figure 1 - Example of a complex structure to be manufactured, created by Topology Optimisation. FONT: ARUP

          Other limitations of the most commercial software are about the conversion of the final optimised results into useful continuo\us CAD geometries. When designers must create and smooth the resulting structure, it demands time (and cost), and features from optimised results may get lost, extra weight added, or even lose performance. When the software has smoothing algorithms, such as Laplace iterative smoothing, marching cube algorithm and the smoothing algorithm proposed by Wang and Wu, that it is a field still in improvements, some information from the optimised geometry could also get lost and mechanical behaviour influenced (Fiebig et al., 2015).

            The benefits of the adoption the TO in design process must be analysed counterpoint to how much it is possible to save with weight/mass reduction or gain in performance and how much is necessary to spend/invest to make feasible the solution found through CAD post-treatment and manufacturing process.  But, if it would be necessary to compare all current limitations, make the geometry viable to manufacture is one of the most apropos.

As defended in recent years by Gu (2013):

The manufacturing challenge prompts further thoughts on the use of Additive Manufacturing (AM) in order to retain much, if not all, of the original organic yet complex optimal topology. This emerging area of combining TO and AM is very promising for designing better… (p.2).

…the integration of TO into the structural design process will be an area for continuous developments in the future (p.4).

…TO needs further integration with the manufacturing process in order to be able to fabricate the highest performing topology at reasonable cost (p.5).

Recent years have witnessed an increased variety of applicable conventional manufacturing constraints implemented in the commercial software, such as Member Size Control, Draw Directions, Extrusion Constraints, Pattern Repetition, Pattern Grouping implemented in OPTISTRUCT. However, these manufacturing constraints are still limited. They are also constrained by the simplified TO model that usually does not have finer meshes to count on small manufacturing features. To address this challenge, the commercial TO tools will keep improving and adding its manufacturing constraint capability (p.5).

… design and manufacturing, using modern information technology, would be integrated into one step, consistent with the modern trend of employing unitized structures (p.6).

Despite the evolution of the mathematical formulation of TO over the last decade, there are still limitations on achieving practical solutions for most of the real-life engineering problems. Most of the commercials software do not have complete solutions to perform optimisation of nonlinearities, dynamics, crash, sheet structures, multi-physics interaction, fatigue, multi-scale and hybrid parts or material (Fiebig et al., 2015; Gu, 2013; Larsson, 2016). In these cases, it can be useful to combine the TO with other tools like Shape Optimization, FEM, to get closer to the optimum solution for the problem. Moreover, with several mathematical approaches and methods in development, the unify and standardisation of methods and approach is a necessity. As defended by Sigmund & Maute (2013): “… the topology optimization community should reunite, get together in joint ventures in the search for the “optimal optimization approach” and in this process use standard benchmarks as well as insightful and expert-based comparisons between methods” (p.1050).

           

Since 2013, Rozvany has defended that, for industrial applications, the dominant preferences for the TO tools are: low CPU time, the generality of applicability, the simplicity of implementation, reliability and the simplicity of obtained topologies. These recommendations became the preference for methods and software development and state of the art for the main commercial tools (Gu, 2013).

            For industry purposes, convergence to an integration of TO in CAD and CAE environments is a reality. The reasons for this are due to some benefits: all the functionalities and features within proved and reliable CAE solvers should also be accessible for the TO; the optimisation can utilise existing investments in hardware and software; make use of knowledge of the staff responsible for developing CAD and CAE. With these, the TO becomes an add-on module of commercial CAD and CAE solvers, as predicted in 2006 (Bendsøe, Olhof, & Sigmund, 2006).

            Finally, in the product design engineering environment, to find the optimum balance between the form provided by TO and the form requirements from design, most of them intangibles and qualitative are still a challenge for designers and engineers and no software are provided with these constraints at all.  

Bibliography:

Bendsøe, M. P., Olhof, N., & Sigmund, O. (2006). IUTAM Symposium on Topological Design Optimization of Structures, Machines and Materials. Dordrecht: Springer.

Fiebig, S., Sellschopp, J., Manz, H., Vietor, T., Axmann, J. K., Schumacher, A., & Ag, V. (2015). Future challenges for topology optimization for the usage in automotive lightweight design technologies. 11th World Congress on Structural and Multidisciplinary Optimization, (June), 1–8.

Gu, W. (2013). On Challenges and Solutions of Topology Optimization for Aerospace Structural Design. 10th World Congress on Structural and Multidisciplinary Optimization, 1–7.

Larsson, R. (2016). Methodology for Topology and Shape Optimization : Application to a Rear Lower Control Arm.

Sigmund, O., & Maute, K. (2013). Topology optimization approaches: A comparative review. Structural and Multidisciplinary Optimization, 48(6), 1031–1055. https://doi.org/10.1007/s00158-013-0978-6

Zhao, D., Li, M., & Liu, Y. (2017). Self-supporting Topology Optimization for Additive Manufacturing. Retrieved from http://arxiv.org/abs/1708.07364

[1] The additive manufacturing is a process in which the material is deposited following some coordinate system, usually controlled by computer and typically layer by layer. It is also popularly called of 3d printing.

[2] DfAM means Design for Additive Manufacture. It is a conjunct of good practices to consider during designing some object with aim to achieve the best performance possible to additive manufacturing. DfAM is part of the DfX (Design for Excellence) methodologies.