Cause of failure and optimization of a V-belt pulley considering fatigue life uncertainty in automotive applications

High accuracy of dimensions and strength in design requirements are required to produce reliable automotive components with consistent strength distribution. For example, a V-belt pulley is widely used to transmit power between rotational mechanical elements. However, due to defects from the manufacturing process and heterogeneity of materials, different kinds of failure damage may occur in pulleys of identical shape and material. Common applications in the automotive industry include crankshafts, water pumps, air-conditioner compressors and power steering pumps. Although the shape and the usage of pulleys are very simple, evaluating the pulley design is difficult because the loading conditions and installation environment are complicated. This paper focuses on the clutch pulley in the A/C compressor system of automotives and cause of failure was investigated. The applied stress distribution of the pulley under high-tension and torque was obtained by using finite element analysis (FEA) and based on theses results, the life of the pulley with variation in fatigue strength was estimated with a durability analysis simulator. The results for failure probabilities of 50% and 1% were compared with the fatigue life. Incidentally, the purpose of this study was to optimize the fatigue life of vehicle components from the stochastic point of view. The fatigue life was obtained by an approximation function, and the optimum design was verified by fatigue tests considering durability and validity. The design optimization of a V-belt pulley was performed using an approximation function, which improved the fatigue life. A new shape optimization procedure was presented to improve the fatigue life of the pulley in automotive applications and the shape control concept was introduced to reduce the shape design variables. Design of experiment (DOE) was employed to evaluate the design sensitivity of fatigue life with respect to shape design variables. (C) 2008 Elsevier Ltd. All rights reserved