Crash injury assessment for wheelchair users using parametric human models and scalable manual wheelchair finite...

Wheelchairs offer customization options to ensure a proper fit for a user's specific body size and shape, yet crash injury assessments rarely account for geometric variabilities in wheelchairs and their users. This study aimed to develop scalable finite element (FE) models of manual wheelchairs coupled with diverse human body models (HBMs) to investigate protection of wheelchair-seated occupants.

HBMs representing small female, small obese female, midsize male, large male, and large obese male occupants were generated by morphing the midsize male simplified GHBMC model (GHBMC M50-OS v2.3) based on statistical human skeleton and body shape models. A previously validated FE model of a Ki Mobility Catalyst 5 manual wheelchair was used as the baseline. A method was developed to scale the dimensions of wheelchair components to fit the wheelchairs for different occupant sizes following the common practice of wheelchair customization. The scaling method was further validated/calibrated by comparing predicted and actual wheelchair dimensions for 10 wheelchair users. A MATLAB-based algorithm was then developed to automate the wheelchair scaling process, integrate the HBMs with the customed wheelchair model, and fit the seat belt. Thirty frontal crash simulations were then conducted using five HBMs/wheelchairs, three restraint configurations (fixed belt, advanced belt with load limiter, and advanced belt with an airbag), and two wheelchair tiedowns (4-pt strap and UDIG). Injury measures included HIC15, BrIC, DAMAGE, chest deflection, maximum principal strain (MPS) in ribs, femur force, and lower tibia force. Analysis of variance (ANOVA) was performed to test statistical significance.

By comparing to the volunteer data, customized wheelchair models based on occupant size reduced the root-mean-square-error (RMSE) of seat width by 25.5% and seat height by 66.4% compared with using a single wheelchair model for the entire population. ANOVA revealed that occupant effects were the most dominant for femur injury (η² = 0.94, p$\text{<}$ 0.001) and chest injury (η² = 0.61, p$\text{<}$ 0.001), restraint configuration was the second-largest contributor to chest MPS (η² = 0.26, p$\text{<}$ 0.001), while tiedown configuration is less significant for the tested frontal impact condition. Regardless of the restraint or tiedown configurations, the large obese male sustained the highest head and chest injury risks due to the poor belt fit and higher inertia. Female models showed considerably higher chest injury risks than the non-obese male models. Across all HBMs, belts with load limiters reduced the chest injury risks significantly (∼30% on average), and advanced belts with an airbag helped further reduce the injury risks across all body regions.

Scalable wheelchair modeling enables realistic injury risk assessment for wheelchair users with a wide range of size and shape. Crash simulations showed that occupant characteristics and restraint system design dominated injury outcomes, suggesting the need for advanced restraint systems for wheelchair-seated occupants.