A butterfly valve is a type of flow control device, which is widely used to regulate a fluid flowing through a section of pipe. Currently, analyses and optimization are of special important in the design and usage of butterfly valves. For the analysis, finite element method (FEM) is often used to predict the safety of valve disc, and computational fluid dynamics (CFD) is commonly used to study the flow characteristics of valve. However, it is difficult to obtain accurate results for the optimization of butterfly valve due to the high non-linearities. For this reason, metamodels or surrogate model methods are extensively employed. This paper integrates metamodel with FEM and CFD analysis to optimize a traditional butterfly valve, where the weight of the valve disc is the design objective, and the strength safety of disc and the pressure loss coefficient of valve are constraints. Kriging model is employed as a surrogate model to formulate the objectives and constrains, and the orthogonal array is used as design of experiment to sample the computer analysis. The optimum results with the corresponding variable combinations for the valve disc are obtained easily by this method. Moreover, the structural and fluid analyses with the obtained optimum variable combinations are conducted again to verify the accuracy of the optimization method. The results demonstrate the capability and potential of this method, which integrates the Kriging model with FEM and CFD analysis, in solving the optimization of a butterfly valve.

A numerical dynamic analysis is carried out to simulate a three-dimensional flow and to study the performance of a single-disc butterfly valve with a diameter of 1.35 m in the opening process. The unsteady fluid flow during the opening process is studied by using the moving grid technique. The turbulent separated flow is also studied to understand the unsteady flow better through velocity distributions and pressure distributions on the mid-symmetric plane. The formation of vortexes is also discussed. Then the characteristics of butterfly valves such as the flow coefficient, the dynamic torque coefficient, and the drag coefficient are studied. The curves of these characteristics corresponding with the variational valve opening angle are obtained to investigate the performance of the single-disc butterfly valve.

A large variety of metal forming processes is required in manufacturing for automotive applications. Traditionally, the design process for metal forming tools is based on trialand-error and on the skill of experienced die-makers. This approach results in high development cost and long lead-times. Especially, the drum clutch requires tight dimensional accuracy in inside diameter and gear shape because it is used as the main component for the automatic transmission. The drum clutch investigated in this study is formed in five forming steps, which are first deep drawing, second deep drawing, restriking, embossing, and Grob processes. Dimensional accuracy of the final products greatly depends upon how much more accurate pre-form is manufactured in the previous forming processes before the Grob process. The deep drawing, restriking, and embossing processes in which the pre-form is formed are very important and decisive steps. Thus in some cases, excessive strain by these operations causes dimensional inaccuracy and cracks initiated from the base and wall of the product.Based on the above background, the objective of this study is to optimize the pre-form shape and tooling so that excessive thinning and crack formation are avoided while a sharp corner radius and flatness are obtained. Process variables such as the punch shapes both of first and second deep drawing, and punch angle were selected to evaluate the deformation characteristics. The optimum parameters were determined from forming simulations using commercial finite element method codes, DEFORM-2D, specifically developed for metal forming simulation. Finally, experiments for the whole drum clutch forming processes were carried out to verify the optimized forming parameters and the analytical results.