In this paper, a [mite element analysis is utilized to investigate the material properties such as yielding stress, strength coefficient and strain hardening exponent effects on forming load and maximum effective stress. The adductive network was then applied to synthesize the data set obtained from the numerical simulation. The predicted results of the maximum forging load and maximum equivalent stress of bevel gear forging from the prediction model are consistent with the results obtained from FEM simulation quite well. After employing the predictive model can provide valuable references in prediction of the maximum forging load and maximum equivalent stress of bevel gear forging under asuitable range of material parameters.
2. Basic theory
2.1 Finite element modeling
where is the volumetric strain rate, Jr is functional of the total energy and work, and k, a penalty constant, is a very large positive constant.where (J is the effective stress, e is the effective
strain-rate,F, represents the surface tractions and, Uj isthe velocity components. The variational form for finite-elementdiscretization is given by:5C and 5 e, are the variations in effective strainrate and volumetric strain rate. Eq. (1) and Eq. (2) are the basic equation for the [mite element formulation. A commercial FE code DEFORM-3D [8] is adopted to analyze the plastic deformation of the near net-shape bevel gear forging from a sintered metal billet. The iteration methods adopted for solving the nonlinear equations are Newton-Raphson and the direct iteration methods. The direct iteration method is used to generate a good initial guess for NewtonRaphson method, whereas Newton-Raphson method is used for speedy [mal convergence. The ,convergence criteria for the iteration are the velocity error norm 116VII/llvll:o; 0.01 and the force error norm 11M'11/IIFII:o; 0.1 , where livII is defined as
The [mite element method has been applied to simulate the plastic flow ofmetal materials during the forming process. For the bevel gear forging process of a plastic deformation problem, the governing equations for the solution of the mechanics in plastic deformation for metal materials involve equilibrium equations, yield criterion, constitutive equations and compatibility conditions. The duality of the boundary value problem and the variation problem can be seen clearly by considering the construction ofthe function
2.2 Adductive network synthesis and evaluation
where h is the number of hidden units in the network. Combining Eqs. (4) and (5), the general polynomial function in a polynomial functional node can be expressed as:
In the abductive network, a complex system can be decomposed into smaller, simpler subsystems grouped into several layers using polynomial functional nodes. The polynomial network proposed by Ivakhnenko [9] is a group method of data handling (GMDH) techniques. Theses nodes evaluate the limited number of inputs by a polynomial function and generate an output to serve as an input to subsequent nodes of the next layer. The structure of polynomial network is shown in Fig. 1 [10]. It consists of sigma (summation) units in the hidden layer and pi (product) units in the output layer. Output of a sigma unit is a weighted sum of its inputs, and output of a pi unit is a product of its input. Let the k th input pattern to the network be specified by X and let the weight associated with connection from input unit i to hidden unit j be wij' Then, the output ~'k ofther sigma unit is given by where h is the number of hidden units in the network. Combining Eqs. (4) and (5), the general polynomial function in a polynomial functional node can be expressed
翻译：液压系统
液压传动和气压传动称为流体传动，是根据17世纪帕斯卡提出的液体静压力传动原理而发展起来的一门新兴技术，1795年英国约瑟夫•布拉曼(Joseph Braman,1749-1814)，在伦敦用水作为工作介质，以水压机的形式将其应用于工业上，诞生了世界上第一台水压机。1905年将工作介质水改为油，又进一步得到改善。 液压系统的英文文献及中文翻译(7):http://www.751com.cn/fanyi/lunwen_22554.html