The effects of seat parameters and restraint systems on lap belt submarining: Response of small female and midsize male finite element models

OBJECTIVE: Lap belt submarining increases risk of abdominal injuries. Seat parameters, restraint systems, crash severity (delta-V), and occupant size all influence submarining outcomes following vehicle crash. The objectives of this research were to 1. Identify which seat and restraint parameters have the greatest effect on lap belt submarining of small female and midsize male models in vehicle crash and 2. Determine if these findings were consistent between human body models and crash dummies. METHODS: 800 simulations were conducted using GHBMC simplified occupant models (M50 THAN-OS v2.3 and F05 THAN-OS v2.3) and THOR ATD models (50 M v1.9.2 and 05 F v1.5.2) representing midsize male and small female occupants. The THAN-OS versions of the human models were sized to match the corresponding THOR dummies. All simulations were performed in a simplified sled environment incorporating an OEM seat model. Seven parameters identified in literature as influential to submarining were systematically varied in a design of experiments. These included seatback recline angle, seat cushion angle, buckle anchor position, knee bolster presence, seatbelt load-limiter, retractor/anchor pretensioner, and delta-V. Submarining was defined as the instant the lower edge of the lap belt passed over the ASIS and was identified through a python-based algorithm which was further verified through visual inspection. Following classification of submarining outcomes, univariate logistic regression was performed to evaluate the association between individual parameters and submarining occurrence. Multivariable logistic regression was then used to assess the independent effect of each parameter while controlling for all others. Odds ratios were calculated to quantify effect size, and statistical significance was assessed at α = 0.05. Wald chi-square statistics derived from multivariable models were used to rank the relative contribution of each parameter. Analyses were performed for all models combined and then stratified by model type and sex. RESULTS: In both univariate and multivariable analyses across all simulations, statistically significant associations with submarining were observed for seatback recline angle, knee bolster presence, seat cushion angle, delta-V, and anchor position. Based on Wald chi-square statistics, parameters ranked from greatest to least contribution to submarining response were anchor position, seat cushion angle, seatback recline angle, knee bolster presence, delta-V, pretensioning load, and load-limiting level. Differences in effect magnitude and significance were observed between model type and sex, indicating model/sex-dependent sensitivity to parameter variations; however, overall parameter influence on submarining remained largely consistent between models. CONCLUSIONS: Anchor position, seat cushion angle, and seatback recline angle were identified as the most influential parameters in altering the likelihood of lap belt submarining across both HBMs and ATDs. Importantly, it was found that both HBMs and ATDs indicated similar parameters that were linked to submarining likelihood. These findings support the design of future experimental tests to capture both submarining and non-submarining responses, and building increased confidence in both computational HBMs and physical surrogates.