Hybrid Approach to Identify Kinematic Points of a Suspension System

The knowledge of the suspension kinematic points are of main interest for many engineers in order to create complex multi body system (MBS) models of road vehicles for benchmarking or in-depth investigation on the suspension. In many cases, those either are known from OEM construction data or are commonly determined with the help of contact giving multi-axis coordinate measurement machines. Another literature-discussed approach is the indirect determination under usage of Kinematics and Compliance (KnC) measurements. The method presented hereinafter has the advantage of an easy integration in the standard KnC measurement process, reducing both time and costs. As a result, the hard point information obtained will transfer the real suspension to the simulation in the same conditions as on the test rig. This will be advantageous for pursuing investigations in case of special consideration of the position of the vehicle relative to the (virtual) test environment. At the Institute of Automobile Engineering of the TU Dresden, a systematic method for the identification of kinematic point positions (x-, y- and z-values) using a hybrid photogrammetric and optimization approach has been developed. In a first step, the kinematic point positions are approximately determined using a high resulting optical measurement system. The suspension is positioned at the respective wheel deflection, typically because of the vehicles empty weight with or without an additional drivers weight. In a second step, the approximately identified hard points are used as initial values of the subsequent iteration process as to find possible spatial positions of kinematic points in a small range. Therefore, the KnC simulation results from an MBS model in ADAMS/Car are compared to the KnC measurements from the Suspension Motion Simulator (SMS) test rig iteratively. The objective is to minimize the errors between the KnC characteristic curves and the simulation. The iteration is realized with a simulation exchange between ADAMS/Car and MATLAB. For the purpose of validating the developed identification method, the kinematic points have been measured by a coordinate measurement machine directly as well. The differences between identified positions of kinematic points and those gathered from the measurement machine show to be sufficient for the desired modelling of the suspension.