Injury prediction in a side impact crash using human body model simulation

Area of Science:

• Biomechanics and injury biomechanics
• Computational modeling and simulation
• Automotive safety engineering

Background:

• Understanding occupant loading in real-world crashes is crucial for injury prevention and vehicle design.
• Finite element models offer a promising approach for reconstructing crash-related injuries.

Purpose of the Study:

• To develop a robust methodology for reconstructing real-world crash injuries using vehicle and human body finite element models.
• To validate the predictive capabilities of the Total Human Model for Safety (THUMS) in simulating occupant injuries.

Main Methods:

• A near-side impact crash from the Crash Injury Research and Engineering Network (CIREN) database was selected.
• Vehicle finite element models were used to reconstruct the crash, with parameters optimized to match crush profiles.
• The Total Human Model for Safety (THUMS) was employed to predict occupant injury risk using established injury criteria.

Main Results:

• THUMS predicted a >90% risk of AIS3+ thoracic injuries and a 70% risk of AIS4+ injuries, consistent with the case occupant's outcome.
• The model accurately predicted the absence of pelvic injury, with low calculated risks for AIS2+ and AIS3+ injuries.
• Injury prediction metrics showed high confidence, with a maximum 95% confidence interval width of only 5%.

Conclusions:

• The study successfully demonstrated a variation study methodology for injury prediction.
• Human body models, such as THUMS, can be reliably used to robustly predict injury probability in real-world crash scenarios.
• This approach supports advancements in vehicle safety design and injury prevention strategies.