Occupant Simulation Database and FE-Model based on Honda Accord 2024 Simplified Passenger Model and SOBOL Sampling with 8,192 samples and HIII05F, HIII50M, HIII95M

Description

Abstract (English)

The FE-model is based on the Honda Accord 2014 model incl. interior, published by NHTSA in 2019. The FE-models feature the passenger with the restraint system components belt, PAB, and CAB. As human surrogates, the HIII-05F, HIII50M, HIII95M were used. The original FE-model was incrementally simplified. The models from each step are documented in Plaschkies, F. (2024). FE-model Simplification Chain of Honda Accord 2014 for Occupant Safety Assessment in Front Load Cases. Zenodo. https://doi.org/10.5281/zenodo.13998073. The database was filled using the last model (990_Carpet_Rigid).

Five factors were varied using Sobol-sampling. The factors were: scaling of 1DOF pulse, angle of pulse-vector, PAB mass-flow, PAB opening time of adaptive vent, force of belt load limiter. Each sample was repeated with the HIII-05F, HIII50M, HIII95M as FE-simulation.

The sampling was generated by:
g = scipy.stats.qmc.Sobol(d=5, scramble=False, seed=None, optimization=None)
doe = g.random_base2(m=13)
Python 3.11, SciPy 1.14.0

The file structure was documented in the content.info file. The database has 8,192 samples (equals 24,576 FE-simulations).

The models are a supplement to a dissertation by F. Plaschkies with the estimated publication in 2025.

Technical info (English)

|-- contents.info
|       File structure and content overview
|
|-- fe_models.zip
| |     LS-Dyna FE-models with set parameters according DOE
| |     Unit system of FE-models is mm-t-s, all simulations in v12, mpp
| |-- Assemblies
| | |      Assembling key files
| | |-- sobol_sampling_m13_f5
| | |-- V*
| | | |-- V*.key
| |-- Pool
| | |     Sub-models
| | |     Dummy models:
| | |         HIII05 (outer layer only!) from  Guha, Sarba; Bhalsod, Dilip; Krebs, Jacob (2011): LSTC Hybrid III 5th Fast Dummy Positioning & Post-Processing. Dummy Version: LSTC.H3_5TH_FAST.111130_V2.0. LSTC Michigan.
| | |         HIII50 (outer layer only!) from  Guha, Sarba; Bhalsod, Dilip; Krebs, Jacob (2011): LSTC Hybrid III 50th Fast Dummy Positioning & Post-Processing. Dummy Version: LSTC.H3_50TH_FAST.111130_V2.0. LSTC Michigan.
| | |         HIII95 (outer layer only!) accordingly
| | |-- *.k
|
|-- databases.zip
| |     Extracted data from FE-simulations
| |-- summary_statistics
| | |     Data distribution of selected injury criteria and related signals
| | |     of Top10 highest values
| | |-- *.pdf
| |-- results_info.json
| |       File overview
| |-- dropped_ids.json
| |       ID to be dropped due to likely numerical issues
| |-- doe.parquet
| |       DOE setting in parquet file format (with index if read by pandas)
| |       Index is SIM_ID (id of particular FE-simulation)
| |       Factors are
| |           PAB_M_SCAL = Inflator Mass Flow In Scaling [1]
| |           PAB_vent_T = PAB Mass Flow Out Scaling [s]
| |           Pulse_X_Scale = Translational Scaling of Pulse [1]
| |           Pulse_Angle = Z Rotation of Pulse (analogous buck rotation on sled track) [DEG]
| |           SLL = Retractor Load Limiter B0 [kN]
| |           PERC = Dummy percentile (5, 50 95)
| |-- sim_id_2_id.parquet
| |       Parser file in parquet file format (with index if read by pandas)
| |       Index is SIM_ID (id of particular FE-simulation)
| |       Columns are
| |           ID = Sample ID (all factors apart from PERC are same)
| |           PERC = Dummy percentile (5, 50 95)
| |-- channels.parquet
| |       From FE-simulations extracted channels as parquet file format (with index if read by pandas)
| |       Indices are
| |           TIME = Time stamp in ms, sampled at 10kHz
| |           ID = Sample ID (all factors apart from PERC are same)
| |           PERC = Dummy percentile (5, 50 95)
| |       Columns are channels named in extended ISO MME Format
| |           Units: acceleration (AC) [g], velocity (VE) [m/s],
| |                  displacement (DS) [mm], force (FO) [kN],
| |                  moment (MO) [Nm], volume (VO) [L],
| |                  pressure (PR) [kPa], temperature (TP) [K],
| |                  mass total, in, out (MT, MI, MO) [g]
| |-- injury_criteria.parquet
| |       From channels calculated injury criteria as parquet file format (with index if read by pandas)
| |       Indices are
| |           ID = Sample ID (all factors apart from PERC are same)
| |           PERC = Dummy percentile (5, 50 95)
| |       Columns are Injury criteria with units as defined by e.g. EuroNCAP