SimulationAI Speeds up Innovation by Streamlining Simulation Workflows Using AI
Startup Enables Faster and More Efficient Testing Early in the Development Cycle
Stuart Schaefer, serial entrepreneur and co-founder of SimulationAI, has started five companies in the last sixteen years. His most recent, SimulationAI, is a cutting-edge structural engineering software firm and Schaefer’s second venture based on artificial intelligence (AI).
“Artificial intelligence is the future,” said Schaefer. “SimulationAI leverages machine learning to accelerate structural simulations. Our AI-based technology can significantly reduce the operational and computational costs associated with simulating complex engineering problems. By reducing runtime and streamlining simulation workflow, engineering teams can innovate faster and more efficiently ahead of their competition.”
Unique AI Capabilities Create SimulationAI’s Secret Sauce
SimulationAI is commercializing patent-pending machine learning technology from The Ohio State University (OSU). Company co-founder Dr. Soheil Soghrati, professor of Mechanical and Aerospace Engineering at OSU, led the team of researchers who developed the proprietary AI deep learning model.
“Soheil is a world-class expert in finite engineering analysis and structural engineering,” Schaefer said. “Our software can accelerate structural simulations to test the structural strength of just about anything.”
That “anything” could include the body of a car for crash resistance, micro-scale semiconductors, or the materials used to build an airplane, a hydropower dam, or a 100-story skyscraper. The company primarily focuses on applications in the automotive and aviation industries.
An AI-based Surrogate for Finite Element Method.
A straightforward way to consider the potential impact of SimulationAI’s artificial intelligence capabilities is to consider the modeling process today that is required to test the strength of an airplane spar (the main structural beam of an airplane wing).
“It can take a mechanical engineering team a week to set up the mathematical problem,” and then another week for a computer to run and come up with the answer,” said Schaefer.
“For the first time ever,” he said, “using AI, we have a different way to calibrate the effect of forces on physical parts. Our patent-pending process trains a giant model to solve complicated equations in one-tenth a fraction of the time without a lot of setup by engineers. We are able to do testing that wasn’t physically possible before. The way our model learns is our secret sauce.”
Next Steps
SimulationAI has a working prototype that can predict structural strength with the same level of accuracy as legacy models. The company’s strategy is to complement existing software, not replace it.
“With SimulationAI, engineers can test structural strength very early in development cycles,” said Schaefer. “They can test every week, or for that matter, every day. The idea is to help engineers make better designs faster by finding problems earlier in the innovation cycle.”
Schaefer invites companies that do complex engineering to contact SimulationAI. “When they can test more times and earlier, they can engineer closer to spec and avoid costs in time and materials created by over-engineering,” said Schaefer.
“The worst feeling an engineer can have is that they just wasted a year of their life on something that doesn’t work well. We are going to help them change course faster,” he said.
