Manufacturing for Design, Not the Other Way Around

Good designers and engineers know how to design for manufacturing. Using their experience and guidelines for the various processes, they can design parts that are easy and cost-effective to machine, cast, forge or otherwise get into the shape they want. By some estimates, designers and engineers spend 30-50 percent of their time designing a shape to achieve the desired function and the rest of their time adapting that shape for the manufacturing process that will be used to make it. While this is important, it is frustrating to think how much time is spent designing for manufacturing. 

Additive manufacturing changes that. With AM, we finally have manufacturing for design, because we can make nearly any shape we want with this technology. An AM system does not care, so to speak, whether it is making a solid block or a complex, organic shape; the computer just tells the laser or deposition head where to melt or deposit material in each layer. Yet, as we have discussed in the past, some shapes are easier to manufacture additively than others. For instance, overhanging features printed in metal often need support structures, thin walls can collapse or be damaged depending on their build orientation, and thick sections can tear themselves apart as residual stresses build up inside the part. 

While no AM equipment provider wants to admit it, these considerations do restrict what types of parts can be made easily with the technology. Fortunately, AM still allows ample opportunities to produce geometries that would be impossible to make with conventional processes. Its opportunistic versus restrictive nature exemplifies how design for additive manufacturing (DFAM) is both freeing and constraining at the same time. The extent to which designers think restrictively versus opportunistically depends on how AM is used. I generally think of three use cases: replicate with AM, adapt for AM and optimize for AM. 

Replicate with AM

In this case, the geometry is given and cannot be modified, because the goal is to replicate an existing part exactly. One example is the link and fitting for a U.S. Navy helicopter that I discussed in the November issue (gbm.media/supports). Replicating with AM is often where organizations start, because they want to be able to make an “apples-to-apples” comparison between the AM part and its conventionally made counterpart. The only real benefit here is speed, because the part is being replicated exactly. It will not cost any less (in fact, it will likely cost a lot more, given the cost of material feedstock for AM systems), and it will not perform any better, or be lighter or stronger. In fact, because the part’s geometry was designed to be made with a conventional manufacturing process and not with AM, the restrictive nature of DFAM applies most in this use case (see graphic). 

Adapt for AM

In this use case, the geometry is given, but it can be modified to minimize or avoid the restrictive aspects of AM (overhangs, support structures, thin walls, thick sections and so on). In the piston crown example described in December (gbm.media/functional), we adapted the original design and were able to reduce the support structures (restrictive DFAM) while enhancing the internal geometry with conformal cooling channels (opportunistic DFAM). The adapted design printed faster, required almost no postprocessing and offered improved heat transfer for better performance. Adapting designs for AM reduces the restrictive aspects of the process that can drive up cost and takes advantage of opportunities to enhance performance. So if replicating with AM compares apples to apples, then adapting for AM is like comparing apples to oranges.

Optimize for AM

In this use case, the geometry is designed specifically for AM. This is where generative design tools such as topology optimization, lattice structures, biomimicry and so on come into play and leverage the opportunistic aspects of AM. The lightweight oil and gas part highlighted in January (gbm.media/lattices) is an example of this. By using lattice structures and thinking about part orientation and overhangs during the design process, we were able to additively manufacture a part that was lighter weight, required no support structures and permitted material substitution that led to enhanced part life.

If the first two use cases are akin to comparing fruit to fruit, then this third use case is like comparing fruits to vegetables—you can get the same nutrients out of each, but what you might eat is an entirely different type of food, one that you may not have considered before. Similarly, with AM, we can rethink what we are trying to achieve and not be as constrained by the process that will be used to make it. This is where AM can take us.