열교환기 냉각팬의 베어링 하우징 신뢰성 기반 최적설계

Abstract

Purpose: This study aimed to reduce the weight of the bearing housing for heat exchanger cooling fans and satisfy reliability of the design constraints at 95% through reliability-based design optimization (RBDO). Methods: The design of experiments method was used in conjunction with an orthogonal array to establish the design variables of the bearing housing, and finite element analysis was performed to analyze the sensitivity trends of the performance factors with changes in the magnitudes of the design variables. To minimize epistemic uncertainty arising from the surrogate model generated using the response surface method (RSM), the optimal values obtained from deterministic design optimization process were sequentially added to update the surrogate models. Subsequently, deterministic and reliability-based optimization was performed using the updated RSM. Results: The optimal RBDO-based solution was obtained by considering variations in dimensional tolerance. Consequently, the weight of the bearing housing was reduced by 52.48% compared to that of the initial design, and the reliability of the design constraint was satisfied at 95%. Conclusion: The conservative optimum design of a bearing housing was obtained through RBDO by considering safety factors to avoid overdesign. This design procedure can be applied to various mechanical parts to avoid overdesign.