Advances in technologies such as cloud-based high-performance computing, AI, and machine learning are leading us closer to a “push-button manufacturing.” future, but despite these advances, the fact remains that design and manufacturing are largely disconnected. Autodesk's Diego Tamburini explains how to break down the siloes.
The need for closing the gap between design and manufacturing teams is overwhelmingly compelling; traditionally, by the time a product has been designed, only 8% of the product budget has been spent, but 80% of the cost of the product has been locked in.
Some design solutions (such as CAD, or add-ons to CAD) offer some ‘design for manufacturability’ (DFM) capabilities, but they are very limited. This is mainly because DFM is very difficult to achieve, but also because of the history behind software companies who produce these tools. But this is changing; the leading software companies in the industry are building or acquiring capabilities to cover the entire product development lifecycle. So, the incentive (and ability) to better connect its various phases is growing.
Let’s explore three fundamental approaches to software solutions for connecting design and manufacturing:
Check the design against rules
With this approach, the design is checked against a set of rules, and violations are highlighted for the designer to see and alter without leaving their familiar CAD environment. Since these rules are commonly focused on feasibility, solutions that implement this rule-checking approach are called Design for Manufacturability (DFM) solutions.
DFM solutions can identify feasibility issues early on, and capture some of the best practices and manufacturing tribal knowledge floating around the company; but they are also very limited.
Firstly, they focus almost exclusively on machining, sheet metal, casting, and injection molding. Yet more importantly, rules can only encapsulate a limited subset of manufacturability considerations appropriately. Rules are OK for making pass/fail determinations based on geometric properties such as distances, thicknesses, clearances, draft angles, radii, and interferences, but are inadequate for anything more complex.
For example, running simulations, checking the design against the manufacturing capabilities of a given machine, suggesting alternative designs or manufacturing processes, and optimising the design against multiple goals. As such, organisations should instead look to move to checking design against manufacturing capabilities.
Check the design against manufacturing capabilities
With this method, designers check their designs against the actual manufacturing capabilities available to the company (internally or via suppliers). This could be achieved, in increasing level of sophistication, by:
- enabling a closer collaboration between designers and manufacturing engineers
- arming the designer with tools to check for manufacturability by themselves
- having the system automatically check for manufacturability as the design progresses
- having a system that not only checks whether the design can be made, but that proactively suggests alternatives to improve its manufacturability. This could include encouraging reusability or modularity, identifying opportunities to reduce the number of parts or fasteners, proposing a better material or manufacturing technology, or a better sequence of operations.
Advances in AI, machine learning, and cloud-based high-performance computing are making this possible and it’s a future we can start to develop today.
Some Manufacturing-as-a-Service (or cloud manufacturing) companies are already offering some degree of design for manufacturability validation via this approach. For example, ProtoLabs offers DFM analysis for 3D printing, CNC machining, and injection molding and Shapeways provides automated and manual checks, as well as some geometry fixing tools.
Generate a manufacturable design
A step further is ‘generative design’, which mimics nature’s evolutionary approach to design. Designers or engineers input design goals into generative design software, along with parameters such as materials, manufacturing methods, and cost constraints. Then, using cloud computing, the software explores all the possible permutations of a solution, quickly generating design alternatives.
Following this approach, manufacturability is baked into the parameters used to generate the design. So, instead of defining the geometry of a part and then checking it for manufacturability, the system generates a design that is manufacturable in the first place. If it cannot be made, it’s not proposed as a design alternative. This effectively blurs the lines between design and manufacturing and gets us closer to the “push button manufacturing” nirvana.
Design for additive manufacturing is particularly suited for this approach because of the organic, curvy geometries often produced by generative design algorithms. In fact, there are already a few real-world examples of generative design used to produce parts that are 3D-printable, such as Airbus A320 partition, the Hack Rod performance car, and the Under Armour 3D-printed shoes.
We’re reaching a tipping point that allows us all to take a fresh look at the problem and come up with new and innovative ways to connect design and manufacturing. Key technologies are reaching their maturity and are becoming more applicable, while industry forces are pushing for a closer integration between design and manufacturing.
As an industry, we spend tremendous amounts of time and resources figuring out how we are going to manufacture our designs. We cannot afford this waste anymore – and, crucially, we don’t have to.
