Old-fashioned CAD not doing the trick anymore, industry envisions cheaper, more flexible design tools
When computer-aided design (CAD) came on the scene it revolutionized the way that electronics designers worked. Now, however, many of these tools are proving insufficient to the task of complex military/aerospace designs. Scientists at the Palo Alto Research Center (PARC) Systems Sciences Laboratory, a Xerox company, are working at the cutting edge of digital design and manufacturing. They see a future where design tools perform a myriad of tasks: creating 3D models; calculating the degree of manufacturability; pointing to probable imperfections in the design and estimating the cost-effectiveness of design alternatives.
“Design is where it all starts,” said Ersin Uzun, vice president and director of PARC. “A lot of the cost of manufacturing a new product is locked in at the design stage. Most designers don’t have a good idea of what those costs will be.”
This research was driven by defense and aviation designers who need more visibility into the manufacturability of their designs. Especially in aerospace and defense, anything that can enhance performance in terms of vibration loads, structural and load bearing characteristics or reduce system weight is hugely useful, said Sai Nelaturi, a scientist who manages Computation for Automation in Systems Engineering at the PARC System Sciences Lab.
“For example, hypersonic aircrafts can benefit a lot from additive manufacturing and design technology,” he added.
Shifting design and materials choices can translate into fuel savings as well. “Humans don’t have the natural ability to design the optimal solution,” Nelaturi said. “By feeding the computer the constraints, it can work around them and still satisfy the functional requirements.”
PARC began working with the Department of Defense (DoD) eight years ago to develop technologies that accurately predict the cost and performance of a design earlier in the process. Now, they are part of the DARPA TRAansformative DESign (TRADES) program, along with Siemens, Georgia Institute of Technology (GT), and Michigan State University (MSU). The goal of the multimillion-dollar project is “to develop a new digital modeling technology that will expand existing CAD software to design complex objects with superior functional properties that can still be manufactured with current manufacturing processes.”
Image courtesy: DARPA>
“This project will have significant impact for high-performance products in specialized industries such as aerospace, defense, automotive, energy production, manufacturing tooling, medical devices and consumer products,” said Dr. Suraj Musuvathy, senior scientist at Siemens Corporate Technology.
“Siemens’ and GT’s experience in CAD systems provide unique insight into how typical users think about and react to different methodologies,” said Musuvathy. “Michigan State University provides a strong background in multidisciplinary optimization. Siemens and PARC are long-time experts in, and users of, additive manufacturing and composite technologies, providing the team access to engineers and material scientists with extensive experience in commercial applications.”
Often, the cost of building a single tank or fighter jet might triple from the original estimate to actual manufacturing using traditional technologies, Uzun added. “In that sector, the reality is that when they start to create a design, whether it is a helicopter, airplane or subsystem, they often don’t know how much it is going to cost and even whether it is manufacturable,” said Uzun. “They wanted software to estimate cost and manufacturability, which became the much harder problem of how to completely simulate the manufacturing process to accurately answer that question.”
Today’s process, using traditional tools, depend upon an iterative prototyping and redesign process that is both costly and time-consuming. Instead of taking months and even years to troubleshoot a sophisticated design, by spotting the potential need for tweaks and redesigns prior to physical prototyping, it might reduce the time to production to weeks or even days, Uzun said.
“The potential return on investment (ROI) is significant,” he added. “You can have all the feedback that you need at the design stage rather than gathering data about manufacturability, performance, and cost manually.”
A combination of technologies, including artificial intelligence (AI), integrated planning and control, and additive manufacturing, will likely address these challenges. Further, these technologies support easier product customization. “This will allow designers to simulate the whole manufacturing operations for a product to understand the cost, how much deviation there is by using a particular manufacturing method, and whether it will meet the parts-per-million quality specs after manufacturing,” said Uzun.
PARC’s Cyber-Physical Systems group is working with manufacturing clients on both digital and physical components to create more sophisticated design tools. The emerging science of computational design invents new representations and algorithms to design, analyze, and plan the manufacture of highly-complex structures that state-of-the-art technology can manufacture.
In the future, PARC hopes that it will be possible to drag and drop a 3D CAD file into a tool that would then create a handful of potential manufacturing recipes that take into consideration the tools and machines that the organization has.
PARC researchers have developed software tools and algorithms that represent digital twins of their designs from the nano-scale to the macro-scale. The result is that a design can be analyzed with any design element tweaked to improve product performance, lower costs, and address manufacturing concerns prior to building a physical model.
Advanced computational capabilities make it possible for the system to potentially offer a variety of design suggestions, from using multiple materials or changing the shape of a design, said Nelaturi. “State-of-the-art additive manufacturing can create non-intuitive complex shapes with material distribution that a human can’t come up with and that is a step forward,” he added.
The first fruits of their work were introduced through a PARC partner at the International Manufacturing Technology Show (IMTS) 2018 in September. This first iteration takes into account raw materials costs, labor, tooling and machine wear and tear in calculating the cost of general manufacturing.