Simulation and testing continues to be a critical part of product development for global industry. Manufacturing and design engineering teams work together to develop, create and fine-tune procedures to optimize supply chain management so that it’s adaptable to fluctuating market conditions while reducing the amount of capital expenditure needed to make iterative changes to physical prototypes. In order to do this more efficiently, moving more segments of physical testing into digital simulation tools seems like the best avenue to pursue. But like Simcenter 2018 Conference of the America’s—the answer is complicated.
The conference starts with a keynote from Nick Appleyard, Vice President, Simulation and Test Solutions, Siemens PLM Software, who welcomed guests from all over the world to the three-day event.
Professionals from multi-disciplinary product design teams around the world who are serious about building products and services using simulation and testing in their experimentation and design were drawn to the event for a few simple reasons: to learn what’s new in the industry, to develop and maintain new and existing relationships in the industry, and to find out what kind of new tools are coming from Siemens, and how they are prepared as a company to respond to disruptive startups.
Appleyard ended his overview of the event by welcoming Radu Tehyunni, Global Director of General Motors to the stage. He spoke about the role of CAE in propulsion systems design at GM and was followed by Jan Leuridan, Senior Vice President of Simulation and Test Solutions for Siemens PLM to the stage. After these presentations, Mark Vitale, a specialist leader at Deloitte Consulting gave a talk about the digital thread and additive manufacturing.
But to understand the concept of the digital thread, you need to first understand the concept of the “digital twin.”
The World of Digital Twins According to Siemens
The full Simcenter product portfolio discussed at the beginning of the event heavily underscored Siemens’ notion of digital twins.
So, what do they mean by digital twin?
1.) A “product digital twin” is tethered to the physical product prototype, providing real-time information to analyze performance while doing a few things, including: allowing users to virtually execute new design alterations and then simulate the effects those changes will have in the digital. This results in reliable virtual design data for product teams to input into the next iteration of the physical product prototype.
2.) A “production digital twin” is used when teams need help validating the effectiveness of a manufacturing process created for the factory floor, resulting in a pre-tested and crystallized understanding of an operational production methodology. Every machine used for manufacturing can have its own digital twin, resulting in correct predictive data for maintenance, preventing expensive downtime during production runs.
3.) A “performance digital twin” processes massive quantities of data from IIoT products in smart plants and other facilities, analyzing them and producing valuable insights for improving product and production system efficiency. The data is captured and aggregated and allows operational actors to make a feedback loop to creating increasingly accurate virtual models.
Simcenter software allows engineers to combine 3D CAE, system simulations and testing into an oracular telescope that can help engineers see how successful their product and production system will be further into a future that includes the construction of a physical product prototype.
Digital Thread for Additive Manufacturing
Deloitte Consulting and the Department of Defense
A single stream of data that connects different digital twins through a product’s entire lifecycle is known as the “digital thread,” and Vitale’s partner keynote demonstrated how the digital thread from digital twins is helping Deloitte Consulting manage additive manufacturing solutions to problematic constraints in the Department of Defense’s manufacturing and supply chain.
The Department of Defense has a massive number of vehicles to maintain and repair, and in the last decade organizations like the DoD have kept extending the life of assets, which can lead to increasingly complex and therefore expensive maintenance and repair operations.
There are longer lead times, inventory holding costs have increased, and the difficulty in sourcing parts grown higher. Traditional manufacturing methods are not optimal for satisfying the variable demand of the parts necessary to maintain an organization with as many assets from different OEMs in different decades as the DoD. With Deloitte Consulting and Siemens, the DoD is exploring Industrial additive manufacturing might fill the need for flexible and distributed production.
In order to achieve a scalable additive manufacturing production capability for this purpose, a digital thread of data that begins with initial designs to finished part histories is necessary. By weaving together applications processes, printers and their data across organizational parameters, the digital thread collects, manages and analyzes valuable information from every stage of the product life cycle.
Vitale said that sometimes components of ground vehicles, ships and aircraft in use by the US Military need maintenance and repair on parts that are no longer being made by the OEM. When you are trying to build a part to spec using advanced additive manufacturing capabilities in a highly regulated industry like aerospace, you have to get everything right at each stage of the product design lifecycle, which means optimizing for repeated inspection is everything.
The digital thread allows design teams to attain better process control, more informed quality testing and increased repeatability for design production. For industrial AM to be spun into operational agile distributed manufacturing effectively and deliver the right part to the right unit at the right time, the information gathered, analyzed and maintained in the digital thread is critical.
Vitale discussed how Deloitte, Siemens and the DoD used industrial AM to create a replacement part for a military vehicle that was no longer in production. By using load data from a simulation run during the design phase of the part, engineers could run a topology optimization and get geometry primitives to evolve and better understand how to improve their design geometry.
To validate the part using the digital thread, the real-time information provided during the industrial AM process could be tracked and inspected layer by layer. If one layer of the part comes out wrong during additive manufacturing, simulations can be run to test if the imperfection of one layer will affect the overall properties of the part beyond acceptable parameter for use as a replacement part. Pretty amazing.
JOBY Aviation and Simcenter STAR-CCM+
Deryl Snyder kicked off Day Two of the conference with a product update presentation centered around Simcenter STAR-CCM+. Simcenter STAR-CCM+ is a multiphysics software for CFD engineers that connects to CAD and PLM systems, performs automated analysis of design updates and integrates with CAE software and allows for coupled co-simulation with 1D & 3D software such as Simcenter 1D for mechatronic system simulation.
Snyder mentioned how a team from JOBY Aviation used Simcenter STAR-CCM+ to increase the cost efficiency of propulsion testing by creating an opening in typical workflows for exploratory design. Aeronautical engineers were able to verify the aerodynamic performance of their all-electrical-vertical take-off and-landing (eVTOL) aircraft prototype with a component level digital twin. STAR-CCM+ helped them understand the effects of trading off between differing design requirements as they considered different design decisions.
Low-order tools are not robust enough for aeronautical engineers at JOBY, and since they are pioneering a new class of aircraft, useful historic data is scarce, meaning statistical analysis methods are not insightful enough as they normally are in aerospace. Building and making iterations on their flight-tested eVTOL prototype is expensive business, even with USD 100 million in investment capital to spend.
Digital twins of products and processes help reduce the cost of physically producing them by taking them as far as they can go in design exploration. Simulating, predicting and optimizing the digital twin of both the product and the production system with data analytics tools, multi-physics simulations and machine learning capabilities give engineers invaluable feedback to explore digital designs in a realistic manner, with realistic design constraints. This has the effect of reducing development time of the product and the production system as well as maintaining a very high level of quality.
By leveraging sophisticated simulation tools from Siemens to produce the highest quality data sets, JOBY aeronautical engineers will continue to build, develop and execute the most efficient physical testing possible.
The presentational and conversational threads running through the conference centered around various practical comparisons of digital simulation tools against their effectiveness when measured against the empirical data gathered by testing physical prototypes in real environments.
Whether its validating the designs of a v-bat ducted propeller design, using multiscale multiphysics simulation and design to optimize an electric delivery truck or simulating rocket booster separation from a UAV, one thing is clear: the nature of testing physical prototypes from data gathered via digital simulation tools has to help product design teams continuously predict how to overcome expected and unexpected design challenges arising in different environments—or it isn’t really valuable. And don’t underestimate the importance of physical testing. If you were a fly on the wall at the event, the number one concept you would overhear in most in conversations is that simulation tools are improving all the time, but replacing physical testing with simulation not a good strategy.
Hands on with Digital Thread
Day three of the conference was for attendees to take a hands-on approach and learn to use some of the tools discussed during days one and two. Attendees interesting in CFD could learn about advanced post-processing in Simcenter STAR-CCM+ or best practices in modeling heat transfer. If you wanted to learn about modeling and simulation for autonomous systems or understand system simulation applications for electric vehicles, you could sit in on a workshop. Those in attendance could also get their hands dirty in the world of generative design with regard to how topology optimization is utilized in part innovation and other areas.
And if you wanted to know more about testing, you could learn about the fundamentals of sound, rotating machinery dynamics or modal analysis and FEA test correlation.
Bottom Line
From experimental research in situ and in laboratories to predictive maintenance in automation cells of manufacturing facilities, Simcenter 2018 was also about sharing best practices in simulation by tweaking product development approaches and streamlining complicated efforts of baking mechanical functions in with electronics, controls and software.
Like any complex domain, there are obstacles and challenges within the world of design, simulation and testing. They run the gamut from annoying minutia to gargantuan catch-22’s that are hard to see a way around. The importance of building working relationships and drawing inspiration from speakers who may be working on problems in areas that seem totally unrelated to one’s own was not lost on attendees and presenters alike.
The mantra about the value of digital twins, at Simcenter 2018 and elsewhere, is that physical prototypes are costly to make and even more costly to make changes to after they’re constructed. The more eventual physical changes to a product or process that could be anticipated digitally through simulation, the most cost effective the product design lifecycle becomes. But don’t even think about replacing testing with simulation, because it’s not going to happen.