Anderson noted that the transition to virtual design and development (of both automobiles and tires) will be as much a physical challenge as an emotional one—and that perhaps the latter cultural hurdle will be the toughest: convincing customers that the first vehicle designed completely virtually by GM "will be good enough to sell."
"We will be judged by customers on a physical and emotional basis, and we need to shift our priorities toward speed—I love going fast—but I am talking about speed as it relates to engineering," he said. "Iterating things on your own, on your own dime, can be pretty expensive."
Anderson recalled his grandfather's advice of "measuring twice and cutting once" as he discussed the virtues and vices of going completely virtual in design and development.
In many ways, Anderson said, the company's work with ventilators during the pandemic accelerated this line of thinking.
"To incite a change this big, cost reduction and speed are critical," Anderson said. "And if speed is the objective, it is all about learning earlier to drive the cost of discovery down. You may have the fastest car, but if you're late to the race track, it doesn't matter.
"Airplanes, rockets, nuclear products—SpaceX is not launching a rocket to see if it works, it is launching to prove that (virtual design) works."
When companies are "running at the speed of design," Anderson noted, the adaptation of existing and workable design methods should be encouraged, but not at the expense of speed.
"We need to focus on technologies that increase the velocity of learning," he said. "And when I say 'fast,' I mean days or weeks for design, not months or years."
Physical validation can ensure quality, and there are inherent physical, geometric and material properties in any vehicle design—every part needs to be inspected and measured, Anderson said.
"The digital twin is good at exploring variation, and physical properties are almost never nominal," he said, referring to the notion of "built-in quality."
Virtual testing can allow for much earlier learning in the design process, the discovery of "knowing what you know and knowing what you don't know."
The most untenable of the options—not knowing what one does not know—is most likely to occur with physical testing, "which has secrets that can't be known" and costs that can be much higher in discovering them.
"The collateral impacts of deviation are not known (with physical testing), and this takes way longer with more than 3,000 parts sometimes to consider and assemble," Anderson said. "The measure of success is replication and correlation.
"Our focus has to be on how to exploit our fundamental abilities, attempting to visualize the unobservable."
One way to ensure the quality of virtual design and testing is to look at past virtual successes, essentially virtual benchmarking.
As GM attempts to implement this "culture of change," as Anderson defines the challenge to "moving to virtual," one of the most powerful strengths of virtual design is in detailing the "sensitivities and interactions" between parts and the design.
"We can use variation ... to circumvent what we don't know and what we don't need to know," Anderson said. "This is a potent instrument to building credibility."
This understanding is especially critical in developing tire models, he said.
"There are many roads on the journey to 100-percent virtual modeling, but tire models are critical," he said. "Assessing and controlling tire models virtually—through lane-keeping assistance, noise variations (especially as it relates to EVs), rolling resistance—all of these things require tire models."
Non-elastomer portions of automotive development are not immune to the high cost and time required for physical design and testing, and that includes "occupant thermal comfort," or keeping the inside environment of a vehicle comfortable for passengers.
"In a mid-size EV, driving comfort and range are in fierce competition. This is not the case with ICEs," Anderson said.
In addition, transitioning to 100-percent virtual design processes will require a compliance to safety standards regarding emissions, crash testing and other avenues.
Will regulatory bodies accept virtual standards and results? The short answer is "yes," Anderson said.
"Imagine the time and money to run these crash tests," Anderson said. "The virtuality of these things saves lives. No one would sign up for some of these safety tests, as there are extremely complex designs in the safety domain.
"This is a big challenge, but an area where we are making good progress."
Indeed, consumer trepidation exists with the move to virtual design—fear of failure, fear of too much risk for manufacturers and fear of an uncertain future.
"Some fear this move. But the threat of missed opportunity far outweighs these fears," Anderson said. "We need to set up designs to learn progressively with these things."
Virtual design and engineering will allow the observation of the previously unobservable, but for that to occur auto and tire makers need to make "the subjective into the objective."
"This may take us to possibilities that we do not necessarily foresee today. (Virtual design and engineering) should be seen as an opportunity for everyone, not a threat or a limitation," Anderson said. "Going forward, it is going to be about the learning velocity. We cannot rely on correlated standard work. The only way to get to 100-percent virtual by 2025 is to work ahead.
"Everyone needs to be engaged, and everyone needs to be fully committed."
The virtual edition of the 40th Tire Society Conference on Tire Science and Technology was viewed by 180 of the society's 200 members, according to the group.