This thesis describes the development of a sled-to-sled side impact sub-system test setup and two mathematical models to simulate a side impact. The sled-to-sled setup involves a bullet sled, a target sled and a simulated door. The bullet sled strikes the initially stationary target sled and the simulated door at a velocity of 13 m/s. The door then impacts a test subject placed on a seat attached to the target sled. The velocity profile of the door is currently programmed to simulate a mid-size passenger car subjected to a side impact by a moving deformable barrier.

Two cadaveric tests were conducted using this device. In one cadaver test, there was 102 mm of space, but no padding, between the rigid door and the test subject. In the other cadaver test, there was a soft 102-mm pad, but no space between the simulated door and the test subject. Autopsy data showed that both of the cadavers had an AIS 4 chest injury.

Seven SID dummy tests were also conducted using this sled-to-sled setup. Experimental results showed that:​ (1) ARCEL pads provided the lowest TTI (77 g's);​ (2) For all of the tests, ASA exceeded the human tolerance limit of 30 g's;​ (3) TIT and ASA increased dramatically if padding was not used;​ (4) TIT and ASA were the highest if there were no shoulder engagement and no padding.

The rigid body MADYMO model matched all the existing cadaveric data at Wayne State University with an average error of 12%. A finite element model of a human occupant was also developed. The twelve ribs were modeled by shell elements. The visceral contents were modeled as an elastic solid accompanied by an array of discrete dampers. Bone condition factors were obtained after autopsy to provide material properties for the model. A total of four runs to simulate Heidelberg type sled tests, eight runs to simulate pendulum tests and two runs to simulate sled-to-sled tests were performed. Chest band data from the latest sled-to-sled tests were compared with the calculated deformable chest contours. It is found that there is a fair agreement between model results and experimental data. The model can be used to estimate the current chest injury parameters of an elderly occupant involved in a side impact.