Enabled by the convergence of High Power Computing (HPC) and Artificial Intelligence (AI), Generative Design (GD) can dramatically shorten the product design process by reducing the number of iterations needed, bringing products to market faster, and making them lighter and cheaper. Using this technology, Airbus was able to cut 45% or 30kg of the weight of an A320 partition part, helping reduce its carbon footprint by saving more than 3,000kg of fuel per year.
What is Generative Design?
The same way autonomous driving is set to revolutionize transportation by allowing passengers to only specify their destination, generative design or GD can autonomously create optimal designs from a set of system requirements. Users can interactively specify the functional needs and goals of their design, including preferred materials, engineering constraints, and manufacturing process (milling, casting, additive manufacturing etc.); the engine will then automatically produce a manufacture-ready design with limited input from the user.
What are the benefits?
In one example General Motors and Autodesk engineers worked together using generative design to reconceive a small vehicle component: the seat bracket where seat belts are fastened. Autodesk’s Fusion 360 produced more than 150 valid design options based on parameters the engineer set, such as required connection points, strength, and mass. A design was chosen with an organic structure no human could have conceived that was 40% lighter and 20% stronger than the original part. The eight different components making up the bracket were also consolidated into only one that was fully 3D printed.
Generative design enables companies to:
- Free engineers from repetitive design tasks to focus on higher-value decisions like maximizing part performance
- Increase productivity by speeding time to market and limiting scrap
- Reducing weight and cost, while retaining required strength
- Reduce rework by limiting the number of prototypes designed and tested, and to eventually eliminate prototyping entirely once the technology matures
- Leverage advanced manufacturing methods/processes such as additive manufacturing
- Drive innovation and provide competitive differentiation of products at minimal cost
- Improve the quality of product and reduce warranty claims
- Allow for domain knowledge to be easily democratized across the organization
Why does it matter?
Currently, engineering design of components, devices, and machines is performed by people with extensive technical expertise in a given area. Most commonly, a degree in that domain, as well as extensive experience, is required to be apt at performing such a task. In addition, the design process often involves multiple groups with diverse competencies to create the most optimal design based on the required specifications. This includes design engineers, simulation specialists, structural engineers, manufacturing professionals, and fluid experts among others, depending on the use case. It’s an expensive, time consuming and costly process that is prone to errors. However, product design is evolving, driven by a combination of macroeconomic trends and emerging technologies.
- Macro trends impacting engineering design and manufacturing companies:
- Worker shortage: Despite the number of graduating engineers remaining high, the number that finds roles in engineering design firms is decreasing. The appeal of high tech VC firms often is more enticing to graduating engineers than traditional design firms. This is contributing to the industrial skills gap. Similarly, subject matter experts who have been in the industry for decades are increasingly reaching the age of retirement. According to the Manufacturing Institute, the median age of workers in the supply chain is 50. As such there’s a growing threat of lost domain knowledge as the gap between new hires and retirees widens.
- Productivity: The manufacturing industry at large has seen productivity plateau for the past several years, which has a direct impact on local and global GDP. According to the Bureau of Labor Statistics, productivity increased by 1.2% between 2007-2017, compared with 2.6% from 2000-2007
- Competitive landscape: Manufacturers are looking for new, innovative ways to remain relevant in their field, partly driven by fear of being displaced by other firms
- What this means for design engineering firms:
- They must close the skilled labor gap to be able to continue to deliver high-quality products to market effectively by democratizing knowledge across the enterprise. The barriers to entry will dramatically reduce since the generative design will get democratized and will not require highly qualified labor
- They must improve their processes to be more efficient and capture market opportunity by speeding time to market. Reducing iterations and prototyping will help companies bring products from ideation to market faster and more efficiently
- They must differentiate their products and services to compete effectively in a global market to drive design innovation while reducing material consumption and waste, and overall production cost
At the same time, technology trends are disrupting the design process and are creating new opportunities. New technology trends that could disrupt markets include: 5G, robotics (cobots), cloud, 3D printing (additive manufacturing), advanced materials, artificial intelligence, digital design, simulation, high-performance computing, Internet of Things, 3D interfaces (AR and VR). Most of these technology trends can have an impact on the engineering design process and are changing the way products are designed and manufactured.
How does it work, and what technologies are required?
Generative design for industrial products is a relatively new concept. Previously it was not possible to derive true generative optimizations due to limitations of computing power, cloud accessibility and artificial intelligence. With barriers to computing power lower (performance, cost, and accessibility) and the ability to automatically manufacture parts, generative design is an emerging design trend.
- Cloud computing is the on-demand availability of data storage and computing power without direct active management by the user. That ability allows users to perform all types of calculations without the need to own a powerful machine to be able to perform the given computations. The relevance of this technology is the ability to test hundreds if not thousands of potential design possibilities and present design alternatives for consideration allowing the user to choose the most optimized version. In addition, the cloud allows users to gain access to a variety of computing types depending on the given task (CPU, GPU, HPC, etc.)
- Artificial Intelligence is the use of technology and mathematics to create systems that perceive, learn, and act to enhance or replace human capabilities. Integral to the generative design is the goal of augmenting design expertise and capabilities by optimizing designs for multiple objectives simultaneously and providing a designer with several design alternatives, enabling companies to substantially reduce engineering cycles. It also allows the creation of a constraint-driven design that would be non-intuitive to a human or that normally requires deep expertise to optimize.
Long term, the combination of cloud computing and AI will allow a generative design system to automatically consider outside parameters to help optimize not only the design of the product but also how the entire supply chain could affect and optimize the design. This includes cost and availability of materials, geographic location, temperature/humidity dependencies, manufacturing process etc.
Eventually, the generative design will democratize design since it will not require users to have any engineering background or to have an extensive understanding of structures, mechanics and materials. All that will be needed is an understanding of the usage of the part being designed and its purpose; the system will then take care of the rest.
How does generative design fit within the broader Industrial IoT initiative?
By connecting products and assets, operating either in the field (like a truck) or in a factory (like a machine), data gathered can be used to provide more accurate information to help the generative design system further improve and optimize future design iterations.
The current state of the market
Generative design is still in its early stages. End users are mostly experimenting to understand how they can fit and help improve the current design process. Applications are currently limited to prototypes and small parts.
Players in the market are split between incumbent computer-aided design and simulation companies like Autodesk’s Fusion 360, Siemens’ NX, Dassault Systeme’s Catia, PTC’s Creo and Altair; and recent, smaller companies that hope to disrupt the market such as nTopology, ParaMatters, and Desktop Metal. In many cases whilst generative design solutions can provide significant benefits as a standalone, the real value is achieved when the technology is used to augment existing CAD products. As such IHSM expects to see ongoing consolidation in this market, as recently exemplified by the acquisition of Frustum by PTC in November 2018.