Abstract: Robust design of powertrain mounting system is researched to improve the robustness of the system under uncertain stiffness property of the mount. Global optimization algorithm and parallel computation technique are employed to improve the computational efficiency. The motion equations of the system are built in crankshaft axis system, torque roll axis system and vehicle axis system. Modal formulation is presented for each excitation in a specific axis system. A design procedure is proposed. In this procedure global optimization algorithm is employed to search feasible region first; then the feasible region is divided and distributed into each computation core; finally, parallel computation technique is used to carry out robust evaluation. A robust optimization model is formulated based on a powertrain mounting system and optimization is carried out according to the proposed design procedure. Results indicate that the robust mean value of decoupling ratio has a little variation from the deterministic one, but the robust standard deviation is obviously smaller than the deterministic one. Especially for higher uncertain level, the robustness of the system is significantly improved.

Key words: vehicle engineering, powertrain mounting system, robust design, modal formulation, global optimization, parallel computation