A model describing a vehicles dynamics in front and rear impact

Abstract:

As a mechanical engineer student at Luleå university of technology this master thesis has been done in close collaboration with Gestamp Hardtech in Luleå. Gestamp Hardtech is today well known for their press hardening technique. One of their largest product volumes are bumper systems for cars. Development of bumper system is an iterative process where every bumper system is designed with computer simulations to have the ultimate crash performance. If short calculation time is required full car simulation is not an option and often a full car model is not provided by the car manufacturer. One way to represent the car in a simplified way a mass node is rigidly connected to the bumper system. The mass node only allows translation in one direction. Due to dynamic properties of the car, the mass node representation of the car has been proven not to be accurate enough to yield a trustworthy bumper design.
A mass, spring and damper system has been developed to handle dynamic behaviour in the primary moving direction of the car. During impact a car sometimes dives or rises due to a height wise offset between the bumper system and the crash object. In this thesis the analogy from the translational model will be used to create a rotational model which reproduces the crash behaviour of the car. To determine the parameters in both the translational and rotational model an optimization routine is used. The routine guesses a set of parameters and then compares the response of the dynamic model with a measured response in a real crash test or a full car simulation. A rotational dynamic model uses torque as input. To calculate the torque a length of the lever, llever has to be known. In this thesis different methods to determine llever are evaluated. A method where the rotation centre is determined by interpolating a point that has the smallest uz is chosen.
Correlation of the dynamic model in finite element analysis is done by studying the pitch and the torque acting on the car during the crash compared to real car crash with equal set up. The modelling technique could resolve the behaviour partly, but the correlation is not quite as good as expected. Several aspects that could contribute to miscorrelation are discussed. The fact that the centre of mass is on a different height than the impact point which will lead to a different input torque is believed to have a large influence. This aspect is to some extended evaluated but should be further investigated.
In summary the model improves the dynamic behaviour of the simplified model but still more investigation is needed on the rotational modelling to ensure correct results. When further validation has been made with other car models a method to scale parameters between car models should be established.