What Is Generative Design, and How Can It Be Used in Manufacturing?
- What Is Generative Design?
- How Is Generative Design Used in Manufacturing?
- 3 Key Benefits of Using Generative Design in Manufacturing
- How Generative Design Goes Beyond Topology Optimization
- What Is the Future of Generative Design in Manufacturing?
Generative design represents a paradigm shift in how things are designed, with machine learning and cloud computing as your design partners. It’s about design exploration, innovation, and advanced computation.
What Is Generative Design?
Generative design is a form of artificial intelligence (AI) that takes an engineering challenge that you define and presents a wide range of appropriate solutions to choose from, which you can then refine according to your needs.
Instead of trying to keep constraints and parameters in mind while drafting a product, part, or tool, designers tell the software what the limits and possibilities are for the end-use criteria they’ve identified. Types of constraints include:
You essentially tell the generative design algorithm that you don’t know the solution, but you do know the requirements.
At the speed of modern cloud computing, generative design’s self-directed process of determining possible combinations based upon your requirements can present many pre-validated iterations of a finished design—thousands, if necessary.
Think of it like designing from the outside in: The various design attributes are established and help arrive at the best result, rather than the traditional method of applying prototypes to performance criteria for further refinement. It’s a design discipline that can be used for anything from basic, common design challenges to very high-precision designs.
Generative design can be used for a wide range of manufacturing industries, including:
- Consumer products
- Industrial machinery
- Building products
How Is Generative Design Used in Manufacturing?
The primary use case of generative design in manufacturing is to automatically trigger design options that are pre-validated to meet the requirements you’ve established. That can be especially important for efficient manufacturing. Sometimes a part or tool must fit into an entrenched workflow or pipeline—methodologically or physically—as part of a larger device or process.
Retooling that entire workflow to suit a new piece can be disruptive and expensive. The parameters around how it has to fit can be extremely narrow. And though a human designer has the expertise to experiment within those tight restrictions, there are still countless variations to explore using generative design software—such as Autodesk Fusion 360—to minimize weight or material costs or to maximize performance metrics.
Many existing processes—including additive manufacturing, CNC machining, and casting—work better when generative design is introduced. It can be used to improve product performance, reduce cost, and explore innovative design concepts.
3 Key Benefits of Using Generative Design in Manufacturing
There are significant benefits to adopting generative design in manufacturing processes. Using generative design makes it easier to reduce a part’s weight without compromising functionality. It also yields sustainability benefits, reducing raw materials use and environmental impacts while increasing performance and cutting costs.
1. Improve Product Performance
Materials and design are partners in manufacturing, with different substances exhibiting different properties. Using a different material can upend a design completely.
When a new material enters the market or economic factors restrict access to the typical material, it can send designers back to the drawing board in a very expensive process of retooling an entire workflow. Generative design can get them back on track faster, taking into account all of the materials available for a project. Whether they select specific materials or parameters around flexibility, rigidity, weight, or response to environmental factors, the generative design algorithm has a global dataset—the cloud—from which to virtually test alternatives much faster than a team of design engineers.
It’s simple physics: If something is heavier, it’s harder to propel, lift, or otherwise move. And when it comes to fields such as mass transport, one of the clearest paths to fossil-fuel reduction is to make cars, buses, or planes weigh less so they need less fuel to move.
Lightweighting has sparked an interest in futuristic substances like graphene or carbon fiber to replace heavier industrial-era materials like iron and steel.
Decreasing a piece of material’s mass yields many savings, such as:
- Logistics: Shipping lighter products increases volume per unit of shipping power used.
- Manufacturing: Whether it’s on the factory floor or the living-room table, making things is faster and easier.
Using generative design to reduce the amount of material on a global, societal basis could mean a greater ability to produce what’s needed to support rising populations using less raw material. Studies show it’s possible to reduce the amount of material used by up to 40%.
Sustainability is a key benefit of lightweighting—lighter products mean less raw material is needed. Generative design allows users to explore different designs based on different materials. And with that, there is potential to use a more sustainable material.
For example, using generative design, aerospace company Airbus created a cabin partition that separates passengers from the galley area in an A320 aircraft, a successful proof of concept that resulted in some arresting numbers. Every kilogram saved onboard the A320 saves 106 kilograms of fuel, and each partition weighs about 30 kilograms. Airbus calculated that if every partition throughout the cabin were made the same way, it would save more than 1,000 pounds of weight in every aircraft, which would cut CO2 emissions by 166 tons per year for every plane.
Imagine using generative design to scale that up to tray tables, seats, even—given the correct materials—controls or the airframe, and the environmental gains become clear.
In an effort to increase performance and reduce weight, Japanese auto parts manufacturer DENSO overhauled the engine control unit (ECU), which manages the amount and timing of fuel delivered by a vehicle’s electronic fuel-injection system. Using generative design, DENSO created the optimal shape to make the ECU 12% lighter while maintaining the heat-dispersing capacity of the original ECU.
Another lightweighting example comes from Canadian motorcycle parts provider MJK Performance. The company identified a dedicated niche of Harley-Davidson motorcycle enthusiasts who wanted to restyle their rides to perform like European racing bikes. Using mass design and manufacturing techniques, the cost of servicing such a narrow customer base would normally be out of reach. Using generative design, MJK was able to take a traditionally bulky part—the triple clamp—and turn it into a lightweight, unique design that was so successful that the company put the part into production by the end of that same week.
The democratization of additive prototyping and one-off production gave MJK the means to address a thriving market; it found generative design to be the perfect partner to deliver the necessary performance metrics while the characteristic visual aesthetic created its own publicity.
Generative design is also about materials science. If there’s a component designed with a particular material but there’s a more sustainable alternative, all one would need to do is insert it into the digital model; the rest of the geometry will adjust so it still adheres to the performance specifications.
The building blocks of tomorrow’s tools, devices, and instruments are also advancing, and the power of the cloud that drives generative design opens a world of cutting-edge research and expertise. Innovations that don’t traditionally get much exposure in your sector can percolate. For example, if a manufacturer has been using aluminum for years, maybe a vulcanized rubber will increase elasticity. Maybe a polymer junction will absorb more load and allow for using cheaper, lighter materials throughout the rest of the structure.
2. Reduce Cost
Cost reduction is one of the most attractive benefits to the manufacturing sector adopting generative design. Every square millimeter of volume saved in a geometry represents a cost savings, and if that savings is scaled up to the global manufacturing and logistics pipeline, the amount can be staggering. Typically, generative design is credited with about a 20%–40% materials reduction.
But it’s not just about sweeping changes or changing the endpoint of the way things are made.
Think back to the example of Airbus changing a single air cabin partition as a proof of concept. With low-cost development techniques such as 3D printing prototypes, reimagining a single tool or fixture in a production line can make an incremental improvement to an established workflow.
Get that right, and it might encourage designers to move onto the next, slightly bigger challenge. It could potentially transform an entire business.
And the tools to help make cost decisions are baked in at the outset of the generative-design process. Fusion 360 contains a reporting tool that estimates costs of a geometry based on different materials, plotting them in a graph for an easy view of how they change according to quantity.
But how can generative design help reduce costs in existing technologies such as additive manufacturing, parts consolidation, casting, and CNC milling?
Generative Design and Additive Manufacturing
Generative design and additive manufacturing go together like peanut butter and jelly. Extremely complex shapes can be made at minimal incremental cost. Because of that, additive manufacturing lends itself well to generatively designed parts.
In fact, the way to get the most out of additive manufacturing is to use generative design, which can create intricate, high-performing shapes. They’re usually the lightest, strongest shapes. And the only way to manufacture those types of geometries is with additive manufacturing methods.
Traditional lathe and CNC processes often can’t wrangle detail at the minute levels such structures require, which makes additive processes perfect for generative design for manufacturing. Likewise, design optiond modeled in classic CAD environments don’t take advantage of the fine-grained detail that additive manufacturing offers.
For example, a 3D printer can manifest a digital model from any number of inexpensive materials for prototyping or presentation. Refine and reprint it as many times as needed, and when the product is ready for production, the same digital assets can seamlessly transpose to the live production workflow.
Generative Design for Parts Consolidation
Two centuries of industrial design and manufacturing have given the manufacturing sector quite a collection of bits, pieces, tools, instruments, devices, parts, fragments, and components used to make into other stuff. There’s a lot of intellectual expertise in that fleet; sampling it for the best pieces to put together instead of designing a new geometry for every project has been the fastest and cheapest way forward.
However, with low-cost prototyping and manufacturing methods such as 3D printing, the cost of introducing a new structure instead of combining existing designs is decreasing—and the result can vastly outperform traditional “puzzle-piece” approaches.
For example, using generative design technology in Fusion 360, General Motors engineers were able to redesign a standard auto part—the humble seat bracket—to be one stainless-steel piece instead of eight. The software generated more than 150 designs for the seat bracket, which secures seat-belt fasteners to seats and seats to floors. The newly designed part is now 40% lighter and 20% stronger than the previous bracket.
With no preconceived notions about what already exists and only knowledge of performance criteria, generative design can get as creative as necessary. You can end up with a part that’s one single piece, whereas previous methods might have seen engineers construct it from countless others.
Generative Design and Casting
Another area where additive still isn’t advanced enough to dislodge traditional methods is in metal casting (pouring hot metal into a negative space to create a specific shape)—particularly when it comes to large and/or heavy pieces of alloy. Additive manufacturing has actually found a home in the metal casting process, but currently it’s used only for prototyping, not production.
Generative design can offer metal casting big advantages in cost reduction and lightweighting. Saving on raw materials is key for casting, as manufacturers typically produce higher part volumes, where savings can add up or the raw materials are very expensive.
For example, Autodesk teamed up with Michigan-based foundry Aristo Cast to develop an ultralightweight aircraft seat frame. The team used generative design, 3D printing, lattice optimization, and investment casting to ultimately create a seat frame that weighs 56% less than typical current models. For a 615-seat Airbus A380 plane, that would mean saving $100,000 in fuel per year, as well as 140,000-plus fewer tons of carbon in the atmosphere.
Additive techniques can also produce the actual metal casting molds, resulting in far more complicated geometries than those from casting, without the steeper set-up times and added costs.
Generative Design and CNC Machining
There has been a flurry of interest in desktop 3D printers during the past decade, but despite the staggering uptick, additive manufacturing is nowhere near toppling the vastly entrenched infrastructure of machining techniques in heavy manufacturing.
Among the new principles generative design introduces to manufacturing is a deeper level of detail in more complicated structures than engineers have ever had at their disposal. So what happens when traditional CNC manufacturing doesn’t allow for it?
Such a constraint is generative design’s bread and butter. The working method is to input every design, performance, weight, material, and manufacturing parameter into the software to generate the best possible suite of design ideas, as well as the limits or guardrails industrial subtractive machining and milling impose upon manufacturing (such as tool diameter or length).
Where a 3-axis CNC device uses common tools, it’s less suitable for the narrow cavities in which a comparable 5-axis device can work. But while the 5-axis device is more precise, it needs more set-up time. Like in every other business process, there’s a trade-off.
But at the design end of the process, you can simply ask the generative-design algorithm to optimize the designs it presents for the manufacturing process you’re using. If you have a 3-axis CNC device, it won’t give you the excessively curved, lattice-like structures additive processes do so well or the smaller apertures suitable for 5-axis machines.
For example, Evolve MH Development Engineering, based in the UK, wanted to design a lighter, more cost-efficient component for its electric hypercar. Weight reduction is important for electric cars to meet performance and range targets. The Evolve team also required the part to be suited for 2.5-axis CNC milling.
The team of engineers used Fusion 360 to enter the part requirements and parameters: things like strength, stiffness, and performance. Ultimately, the component the team created for the electric hypercar is 40% lighter than their original design and was completed in record time.
3. Expand Innovation by Exploring New Design Concepts
Of all the potential benefits of generative design, perhaps the most pervasive and uncelebrated is embedding a culture of innovation into the manufacturing sector overall. Many chief technical officers, engineers, or designers who start out skeptical generative design’s benefits end up being the most outspoken evangelists of what it can do.
Because generative design operates in the cloud, it will put designers and engineers in touch with other ideas and techniques that traditional methods might never have exposed them to.
Industrial engineers can collaborate more with digital designers. CTOs can start to understand how force is exerted on a material. Factory-floor managers can see how a computer can yield a better first step (or 10 or 100 options) than a person might have created on their own. Innovation and collaboration go hand in hand. With generative design, the world of possibility is greatly expanded.
Chicago-based bike-parts manufacturer and innovation lab SRAM makes parts for a dedicated customer base of knowledgeable enthusiasts—in this case, for off-road cyclists. The company produced a prototype of a complicated shape for a bicycle crank arm that could be made only using the combined processes of additive manufacturing and generative design. It was an ideal test case to see what generative design might do for other bicycle parts. The experience inspired the company to explore new opportunities and innovations in design, materials, and cost. SRAM says that generative design plays a major role in how the company conceptualizes its business.
Another example is Elevate, a joint venture between Hyundai’s CRADLE division and design studio Sundberg-Ferar. Looking like a friendly insect from a sci-fi movie, Elevate is a car with wheels that are situated at the end of four articulated “legs” that can move using a combination of any four modes—two driving styles and two walking styles—which can theoretically take it anywhere for mass transit, disaster relief, and more.
Elevate’s myriad parts were designed from scratch for the concept, many of them with extremely lightweight properties thanks to generative design.
But perhaps outer space may be the ultimate test for any materials science, contending with heat, cold, radiation, the onslaught of high-velocity particles, the dizzying inertia of acceleration, and other environments not found anywhere on earth. And when removing just a few grams saves potentially hundreds of thousands of dollars, space exploration represents the ultimate dance of maintaining performance while minimizing mass.
Few institutions work with such rigorous limits more than NASA’s Jet Propulsion Laboratory, which made generative design the ideal partner to develop a concept extraplanetary lander with lower mass and improved performance.
How Generative Design Goes Beyond Topology Optimization
It’s easy to confuse generative design with other design-optimization approaches, such as topology optimization, but there’s an important difference. Any design made the traditional way is simply an engineer’s best guess about how to solve a problem. Topology optimization can only weigh in on the process from that point, taking what a human has made and making it incrementally better.
Generative design begins earlier in the process. It’s a co-designer that presents ideas based on the constraints around what you need your design to do—not on what you think will work (however experienced you may be). Put even more simply, topology optimization removes material from an original solution, where generative design doesn’t need an original solution—it does the exploratory work to help you come up with one.
The optimization pathway from there doesn’t need to dovetail from a single starting point; you can go back to any other starting point and generate countless other options.
What Is the Future of Generative Design in Manufacturing?
Generative-design technology continues to advance—through research, start-ups, and other roads to innovation. These advances will also transform how humans interact with technology and each other. There will be more collaboration across disciplines—engineers with CTOs and designers—and more innovation for the manufacturing industry.
Faster, Smarter Technologies
A study that covered automotive, aerospace, and sporting goods found generative design reduced cost by up to a fifth and both mass and development time by up to half.
But aside from reducing cost, materials, development time, and environmental impacts across the global manufacturing industry, specific advances are trickling down from labs, dedicated start-ups, and research facilities.
One example is generative design that takes advanced fluid dynamics into account. There are plenty of industrial parts that have to operate in environments where the movement of liquid or gas can vastly affect performance.
When the performance benchmarks and design parameters are entered into the generative-design algorithm, in the future, one parameter might be the way the immediate environment will behave, using properties like:
- Will air pressure increase cause a rise in temperature?
- Will it be super-cooled rapidly?
- Will the movement of air impact one particular surface versus another?
In addition to fluid dynamics, technologies such as small-scale additive and 3D printing will dramatically lower the barrier to entry to new players who’ll come into fields armed with new ideas and no preconceived notions about what can and can’t be done.
While the technology continues to advance, one of the most important changes generative design will bring will be the impact on humans. Generative design, as a form of AI in manufacturing, will fundamentally change the tasks facing manufacturing personnel for the better. With shorter development times, designers and engineers will have time to expand their professional—and creative—horizons.
This article has been updated. It originally published in November 2021.