Abstract
Keywords: rubber bush, quasi-static, neural network, suspension, vehicle dynamics
Abstract: In the multibody dynamic model of vehicles, bushings are usually modeled as a combination of spring-dampers in five or six directions. With this spring-damper bushing model, the computer simulation time may be quite long due to the high frequency components of bushing characteristics. In this paper, two types of bushing models are introduced.
A quasi-static model is proposed to achieve the numerical efficiency in simulation of vehicle suspension systems. Forces acting on links are resolved and transmitted to attached points with a quasi-static assumption. In the low control arm with bushing elements, one end is connected to the adjacent body by bushing, and the other end is connected by spherical joint. The elastic deformations of bushing elements are determined by minimizing the potential energy function with a quasi-static equilibrium assumption at each time step. Several simulations with a full vehicle model are carried out to compare the efficiency of the developed quasi-static bushing model. From the pulse steer simulation, the computer simulation time is tremendously reduced to the rigid body modelling with spring-damper bushing model.
The other bushing model suggested in this paper is an empirical bushing model using an artificial neural network model. A black-box approach is carried out to model the nonlinear dynamic characteristics of bushings. One-axis durability test was performed to describe the mechanical behaviour of typical elastomeric bushing components. The test results are used to develop an empirical bushing model with an artificial neural network. The back propagation algorithm is used to obtain the weighting factor of the neural network. Since the output for a dynamic system depends on the histories of inputs and outputs, Narendra algorithm of NARMAX form is employed to consider these effects.
To include test results of the bushing learned through the neural network, MATLAB and Simulink was used to construct the empirical bushing module. The interface module was developed to connect MATLAB and ADAMS program. To show the usability of the proposed bushing model, a half car model was employed and simulated. Simulation results from the random input show a good agreement to the experiments.