Time-compressed crash simulation

  • 28-May-2010 12:09 EDT
Altair Engineering's HyperCrash was used to validate the vehicle model prior to submitting for analysis.

Instead of the usual two or more weeks needed to mesh, assemble, and simulate a vehicle crash finite-element (FE) model, a demonstration of the entire CAD-to-crash process was done in a day.

Altair Engineering's HyperWorks' suite of CAE products and embedded process automation provided the framework for completing the full vehicle proof-of-concept crash simulation during a 24-h span in April. A team of five Altair engineers worked in shifts as a relay team to complete the project.

The computing hardware used for the event included 64 central processing units, 32 GB of RAM, Linux 64 bit for meshing, assembly, and computation as well as a Microsoft Windows operating system. The crash setup was done on a 64-bit Windows XP laptop.

More than 1000 assemblies and 2000 parts representing the body-in-white (BIW) and subsystems—including the instrument panel, powertrain, front bumper, seat assemblies, tires, and suspension components—were used for the CAD2CRASH24 project.

"This CAD2CRASH24 initiative is intended to give engineers time to carry out more investigative studies and explore more alternatives relating to weight, materials, cost, and performance while still accelerating the time to market of the final product," said Jeff Brennan, Chief Marketing Officer and Executive Vice President of Global Markets at Altair Engineering, headquartered in Troy, MI.

Brennan also noted that a faster turnaround time provides an opportunity for engineers to perform statistical and stochastic studies in the crash domain, which could further improve the performance and robustness of designs.

For the proof-of-concept demonstration, Ford Motor Co. supplied the BIW CAD model of a 2010 model year Ford midsize vehicle, and the subsystems were morphed to fit a BIW model from a National Highway Traffic Safety Administration (NHTSA) model, according to Pradeep Srinivasan, Senior Technical Specialist at Altair Engineering.

The simulation project involved parallelized batch meshing of the BIW and the subsystems. "This involved firing off groups of parts across multiple compute nodes on Altair's compute cloud to batch mesh the models. As individual finite-element meshes were completed, another part would automatically be sent for meshing," explained Srinivasan.

All components and assemblies were mass trimmed to compensate for the mass differential between the CAE mass and the bill-of-material mass.

Altair's HyperCrash handled mass balancing, dummy positioning, seat deformation as well as seatbelt routing modeling tasks. HyperCrash also validated the model by performing hundreds of different part and model level quality checks. "Altair engineers visually reviewed the state of each check by status colors—red, orange, and green," said Srinivasan.

The model was then analyzed by RADIOSS, the solver in the HyperWorks CAE software suite. "RADIOSS solves for a 100 ms new car assessment program (NCAP) frontal crash test using an FTSS 50th percentile Hybrid III dummy and reference metric (IMM) airbag," noted Srinivasan.

According to Brennan, "RADIOSS was a critical ingredient to the success of this initiative. The speed at which the analysis can be run has a great effect on the ability to turn around this process in only 24 h and in the future perform multiple simulations in the same time frame."

Results were imported to HyperView, a HyperWorks results visualization and post-processing solution, and interactive standard reports were auto-created. "Those standard results included deformation animations, accelerations, energy plots, occupant injury criteria, and so on," Srinivasan said.

Added Brennan, "We envision that someday simulation could, with enough fidelity and advancements to robust design technology, survive on its own without the need for physical testing. However, today automotive design requires at least one physical test to ensure passenger safety."

HTML for Linking to Page
Page URL
Rate It
3.67 Avg. Rating

Related Items

Training / Education
Technical Paper / Journal Article
Technical Paper / Journal Article
Training / Education