Development and Validation of a Dynamic Abdominal Pressure Twin Sensor Finite Element Model

Some anthropomorphic test devices (ATDs) currently being developed are equipped with abdominal pressure twin sensors (APTS) for the assessment of abdominal injuries and as an indicator of the occurrence of the submarining of an occupant during a crash event. The APTS is comprised of a fluid-filled polyurethane elastomeric bladder which is sealed by an aluminum cap with an implanted pressure transducer. It is integrated into ATD abdomens, and fluid pressure is increased due to the abdomen/bladder compression due to interactions with the seatbelt or other structures. In this article, a nonlinear dynamic finite element (FE) model is constructed of an APTS using LS-PrePost and converted to the LS-Dyna solver input format. The polyurethane bladder and the internal fluid are represented with viscoelastic and isotropic hypoelastic material models, respectively. The aluminum cap was considered a rigid part since it is significantly stiffer than the bladder and the fluid. To characterize the APTS, dynamic compression tests were conducted on a servo-hydraulic load frame under displacement control and held at the peak compression to allow for stress relaxation prior to slowly releasing the compression amount. The initial peak pressures and loads were 15–17% above the level observed at a 10-second hold period with 50% of the decay occurring within 300 ms. The material properties are identified using an inverse method that minimizes the difference between measured and predicted load and pressure time histories. Further, the bio-fidelity static specifications of the APTS manufacturer are used as a basis to identify the quasi-static material parameters. This approach resulted in a reasonable match between physical test data and model-simulated data for dynamic compressions of 10 mm and 15 mm (~50% compression). Additional compression tests are conducted at two compression levels (5 and 10 mm) and at four load offset configurations for use in the model validation. The FE model was used to predict peak pressure responses within approximately 10% error at full-load capacity and achieved CORA ratings >0.99 for the pressure time history. The proposed inverse method is expected to be generally applicable to the component characterization of other models and sizes of APT sensors.