Topographic optimization for 3DPrinting

Despite being the most cited benefit of 3DPrinting, the implementation of complexity freedom is not an easy or straight-forward process. Results from two studies that form part of the PhD research project that gave birth to this blog cannot support the hypothesis that complexity freedom is influential in the conceptualization of products for 3DPrinting. The objective of the studies was to observe whether if there were observable effects of complexity freedom in the ideation processes of entrepreneurs that planned to use 3DPrinting for fabricating their ideas. Neither of both studies found that entrepreneurs address functional complexity in the development of the product architecture. Thus, new product development (NPD) seems to be split in two stages: one of conceptual development, and a second one that adjusts the product architecture based on the generated concepts. Contrary to our expectations, this separation is not bridged with the use of 3DPrinting despite the capabilities of this technology to design and manufacture concurrently.

These results agree with the claims that describe NPD with 3DPrinting as a substitution fallacy where NPD only substitutes other manufacturing processes with 3DPrinting and does not develop architectures that correspond to 3DPrinting capabilities (Stern, 2015). As we have mentioned in other blog posts, this is called a partial approach to Design For Additive Manufacturing (DFAM). Complexity freedom has the potential to turn this partial approach into a global approach where NPD designs “with” and not “for” the 3DPrinter. In a global approach to DFAM designers determine the functional surfaces, the structural loads that each of them is submitted to, and merges the functional volumes without complexity restrictions. Merging the functional volumes in a global approach to DFAM can be done manually. However, an manual development to a global approach to DFAM does not guarantee an optimal architecture that completely exploits complexity freedom.

An alternative that makes a global approach to DFAM viable is the implementation of structural optimization algorithms. Structural optimization is a functional approach to the creation of structures centred in the economy of the designed architecture. The optimization process focuses on optimizing a target, which usually is structural mass or number of elements, without disrupting the constraints imposed to the system by the loads that operate through the structure. While structural optimization is a method that can also be implemented manually, the implementation of optimization modules of CAD software expands the scope of analysis and facilitates the creation of geometries that exploit the capabilities of 3DPrinting. Structural optimization engines are based in finite element analysis (FEA), a mathematical method used to solve complex models that can represent physics or structural problems. FEA splits the system in smaller “finite elements” that can be analysed individually. Combined with CAD models, the mathematical method is able to evaluate the performance of a 3D mesh by making a structural analysis of each of the elements in the mesh. Thus, it can provide valuable information about its structural performance. If on top of the FEA the system has a module that redesigns the structure and resubmits the result of the analysis to the FEA engine, the system is able to optimize the geometry under the given restrictions.

Michael Stern (2015) classifies structural optimization in three main categories: size optimization, shape optimization, and topographic optimization. Size optimization focuses on the elimination of elements under a within a reduction threshold without changing the shape of the model. Shape optimization works all the way around without eliminating material but instead optimizing the contour of the shape for a better performance. Topographic optimization is the most flexible of them because it departs from the definition of a starting design space that after an initial evaluation is redesigned and resubmitted. Continuous iterations remove material and modify the shape of the model resulting in an optimized topography. Results from this kind of analysis can be used to redesign architectures that are produced through traditional manufacturing methods. However, complexity freedom lets the 3D printer use these results as final models for fabrication.

For topographic optimization to work, the design space of the model contains all the information for the solver to work with. Within a global approach to DFAM, the definition of the design space is achieved by delimiting the functional surfaces, the loads that work with them, and the printing volume available in the machinery. Therefore, a topographic optimization machine can provide a solution for the correct implementation of a global approach to DFAM. Fortunately for us, more optimization modules are available everyday through new versions of CAD software. The following example used the module provided by Autodesk Inventor.

Mallet finger

A mallet finger is a common finger injury in sports. It happens when the fingers are subject to extreme impacts or forces that snap the tendons that extend the finger and keep it in place. It is common to see these injuries in sports where athletes handle balls at high speeds such as baseball, basketball, or volleyball. During Easter 2017 I suffered a mallet finger injury while playing beach volleyball in Orewa Beach, New Zealand. Despite being not that painful, a mallet finger is delicate because the injured finger must be immobilized for the tendon to attach again to itself or the bone. If not immobilized on time, the injury is treated through a surgery that fixes the tendon back to the bone with the help of a screw. It has a recovery time of 6-8 weeks which makes everyday activities that use the hands, such as cooking, writing, or typing, uncomfortable to do. Yet, the most common solutions available are plastic one-size-fits-all cone shaped splints that must be fixed with tape.

This problem was addressed through the combination of a topographic optimization module and a Fused Deposition Modelling (FDM) 3DPrinter. Using a global design approach to DFAM the development of the splint started with the 3D model of definition of the functional surfaces that could support the finger in place. In the case of a mallet finger injury, the finger is fixed by pressing the fingertip upwards against the middle of the finger. Therefore, the finger was 3D modelled and both the fingertip and the top middle segment of the finger were isolated as contact surfaces. To constrain the system a force equivalent to the full force of a closing hand was placed against the fingertip surface while the middle top was defined as an anchor surface. The design volume was defined as a tube with the shape of the finger in the middle and a 5mm thickness.

The structural optimization solver gave a solution that removed material and optimized the shape of the splint. However, a global approach to DFAM must also consider the restrictions of the used material and 3DPrinting method. In this case the splint was fabricated using a PLA plastic in an FDM machine. This meant that the geometry had to be oriented to remove support material. The particular orientation that was more convenient to the fabrication process created the material layers in a direction that is aligned with the sheer stress of the structure. Therefore, a trade-off existed between the orientation (speed and material cost) of the solution and the structural performance of the model. To alleviate such contradiction the model was redesigned based on the structural optimization result but with greater thicknesses in the weakened parts. The redesign of the splint also allowed the incorporation of extra features such as edge rounds and holes that facilitated the splint usability.