Automation of ATD positioning for crash analysis simulation

Abstract

In alignment with the United Nations target 11.2, regulators enforce minimum safety requirements in different new car markets, at the same time consumer information organizations and car manufacturers strive to make cars safer. Standardized test procedures using anthropomorphic test devices (ATD) are commonly used to evaluate car crashworthiness and occupant protection. Using computer aided engineering (CAE) utilizing finite element (FE) analysis, standardized test procedures can be reconstructed in simulations. Although, safety development using CAE can be complex given the variations found in standardized test procedures in different new car markets. This study developed an automated positioning method using intersection checks and node-to-node distance minimization for ATDs in CAE. The method was developed for the Hybrid III adult ATDs for frontal crash evaluation and was implemented for use at the car manufacturer using the simulation pre-processor ANSA’s application programming interface. Three ATDs were considered and were positioned automatically in car environment models. To evaluate the methods positioning performance, an automatically positioned 50th percentile Hybrid III male model was compared with physical positioning. The driver’s position in US NCAP full frontal rigid barrier 35 mph was used. The automatically positioned Anthropomorphic Test Device was positioned within two standard deviations of physical positioning for both X- and Z- coordinates in five out of six measurement locations. Furthermore, two application studies were performed using sled simulations and a mid-size SUV developed at the car manufacturer. The studies were again assessed using the Hybrid III 50th percentile male and the US NCAP Full Frontal Rigid Barrier. The first study assessed global injury criteria reproducibility as results can vary due to model decomposition, i.e., numerical noise. In the environment and load case studied, global injury criteria showed a Coefficient of Variation regarding reproducibility of 0,65% to 4,22%. The second study assessed global injury criteria sensitivity to variations in hip-positioning. The hip-positions were retrieved from physical positioning. In this study, chest deflections showed variations up to 3.5 mm and femur force variations up to one kN for hip-positions within two standard deviations of data in physical Anthropomorphic Test Device positioning. Generally, injury criteria were more sensitive to H-point variations then model decomposition. These results stress the importance of achieving the correct H-point position, in CAE and in physical crash tests.