No abstract

No abstract

A robust suspension system design optimization, which takes into account the kinematic behaviours influenced by bush compliance uncertainty, is presented. The design variables are the positions of the joints, and the random constant is the bush stiffness with uncertainty. The design goals for these kinematic behaviours are typically represented as deviations over the wheel movements. It can be very difficult to evaluate the analytical design sensitivity because the deviation is defined by using the maximum and minimum values over the parameter interval. To avoid the difficulty, this study introduces a metamodel technique. The sample variances for the design goals are approximated from metamodels. In addition, a sequential approximation optimization technique is used to solve a robust design problem for the suspension system. The robust design problem has 18 design variables and 18 random constants with uncertainty. The proposed approach required only 189 evaluations until it converged. The selected design reduced the maximum deviations in the toe and camber angles by 72 per cent and 50 per cent respectively, and their variances by 90 per cent, while satisfying the constraints of changes in the toe angle, camber angle, and front-to-rear change in the wheel centre.

This study presents the robust design optimization process of suspension system for improving vehicle dynamic performance (ride comfort, handling stability). The proposed design method is so called target cascading method where the design target of the system is cascaded from a vehicle level to a suspension system level. To formalize the proposed method in the view of design process, the design problem structure of suspension system is defined as a (hierarchical) multilevel design optimization, and the design problem for each level is solved using the robust design optimization technique based on a meta-model. Then, In order to verify the proposed design concept, it designed suspension system. For the vehicle level, 44 random variables with 3% of coefficient of variance (COV) were selected and the proposed design process solved the problem by using only 88 exact analyses that included 49 analyses for the initial meta-model and 39 analyses for SAO. For the suspension level, 54 random variables with 10% of COV were selected and the optimal designs solved the problem by using only 168 exact analyses for the front suspension system. Furthermore, 73 random variables with 10% of COV were selected and optimal designs solved the problem by using only 252 exact analyses for the rear suspension system. In order to compare the vehicle dynamic performance between the optimal design model and the initial design model, the ride comfort and the handling stability was analyzed and found to be improved by 16% and by 37%, respectively. This result proves that the suggested design method of suspension system is effective and systematic.

No abstract