The phrase "simulate, then design" may be counterintuitive for those who grew up in a world when CAD replaced paper drawings. Why is it emerging now? One reason is the growth of mechatronics in today’s cars. With engineers adding electronics and microprocessors to mechanical systems to improve performance, complexity grows.
“System complexity is driving the need for model-based system engineering (MBSE),” agrees Paul Weal, Strategic Development Manager for LMS North America. “If we look at automotive as an example, fuel economy is driving design changes. Engineers have done a great job with subsystem performance—there is not a lot of low-hanging fruit left.” Improving fuel economy even more requires developing new power architectures, such as hybrids. With motors, batteries, and engines working together, engineers need to understand the interactions between these subsystems to optimize performance. “[By simulating] the full system, we can [balance] trade-offs in fuel economy, drivability, and passenger thermal comfort,” Weal said.
Weal also broadens the concept. “MBSE equates better to Simulate-Design-Integrate-Validate [than simply Simulate-Design]. I prefer to call it System Simulation because we are simulating more than just the mechanical domain with AMESim,” he said, referring to the multiphysics simulation package that LMS offers. “We are simulating mechanical, electrical, hydraulic, and thermal domains concurrently in one full system model.” One can no longer add controls onto mechanical systems as an afterthought.
Another key point Weal makes is that LMS is simulating systems based on physics and not on geometry. This precludes traditional CAE methods that use meshes and models derived from CAD. In this approach, vehicle engineers assess performance of the full system much earlier. How is this different from traditional approaches? “Engineering today builds models for every major design revision and every context or scenario for performance evaluation,” explains Weal. “We now see the next horizon. First, managing the model variants, the associated data, and context of application, then second, leveraging existing and validated models to quickly synthesize systems, assess the performance, and evaluate architecture choices.”
This approach does not replace traditional CAE models and meshes created from CAD geometry. “Traditional CAE is still important. This is especially true in the detail design phase,” said Weal. He notes that AMESim interfaces with most traditional CAE tools, setting the boundary conditions for them. Once completed, detailed analysis results are sent back to AMESim to improve the fidelity and check the performance at the system level.
Showing its commitment to MBSE, in 2010, LMS purchased Emmeskay, a global provider of simulation for the discipline. The purchase complemented its multiphysics systems simulations, showing how serious LMS was about MBSE.
How will this comprehensive approach to MBSE evolve in the future? “The coverage of physics will widen. Solver accuracy and performance will continue to improve,” predicts Weal. “Managing model and data and tools for leveraging intellectual property more efficiently will improve. These are both areas of focus for LMS—there is certainly interest and pull from our customers.”