Simulation of ride comfort and durability demands a reliable model for characterizing tire-road interface. Considering of computational efficiency and accuracy, semi-empirical contact models are widely utilized for the calculation of tire forces operating on uneven roads. A subsequent challenge is effective parameterization for a correct prediction of tire performances. However there exist few open publications relating to parameter tuning techniques and parameter value ranges based on a credible database. In this paper, 46 passenger vehicle tires from various tire manufactures, with different sizes, usages and season types were investigated. As a first step toward understanding of the tire-road contact behavior, static stiffness tests, footprint tests and drum-cleat tests were conducted, all of which were vital for the following model parameterization.

Test findings reveal tires radial deformation properties, contact patch and quasi-static enveloping characteristics at different operation conditions. Secondly a tandem-cam model was applied to reconstruct real profiles because tires geometry and deformability acts as a geometric filter to road unevenness. With regard to the influence of load variation on tire contact length when rolling over obstacles, an improvement of the original model was made by implementing a load dependent coefficient to the tandem base length, which meant that this parameter value was updated at each integration step of the simulation. Supported by the measurements above, all model parameters were identified with the Particle Swarm Optimization algorithm in MATLAB.

The effectiveness of the proposed parameterization method was validated through a complete comparison (for all the investigated tires) between cleat test measures and numerical outputs. A further assessment of the simulation quality was conducted by introducing an evaluation criterion, namely the root mean square error. Results showed that on average, the model accuracy approached to 84% for the prediction of the vertical spindle force and was around 71% to the longitudinal spindle force. At the end, the upper and lower bound of each parameter in this enveloping model based on the 46 tire datasets were listed, which provides a statistical insight into the model-based tire characterization as well as a reference for tire and vehicle simulation.

Dipl.-Ing. Yun Pang, Technische Universität Dresden, GERMANY; Dipl.-Ing. Tobias Schramm, Technische Universität Dresden, GERMANY; Dipl.-Ing. Jan Kubenz, Technische Universität Dresden, GERMANY; Prof. Dr.-Ing. Günther Prokop, Technische Universität Dresden, GERMANY