Global sensitivity analysis and multiobjective optimization of bogie suspension

Paper in proceeding, 2016

Multiobjective optimization of nonlinear multibody systems with many degrees of freedom is a burdensome computational challenge. A feasible practical methodology for global sensitivity analysis (GSA) of multibody systems with respect to design parameters is proposed based on the multiplicative dimensional reduction method. The computational efficiency of optimization is significantly improved by restricting the input design parameters only to those identified by the GSA. The methodology is applied for GSA of a railway vehicle dynamics with respect to the bogie suspension characteristics. Several multiobjective optimization problems are then formulated and solved for a railway vehicle model with 50 degrees of freedom using genetic algorithm. The results obtained yield practical information regarding the optimized bogie suspension properties which improve the dynamics behaviour of the vehicle from various perspectives. The proposed algorithm can be used in design optimization of nonlinear multibody systems with different applications.