螺栓法兰连接的数值模拟及接触收敛分析

螺栓法兰连接的数值模拟过程中涉及许多接触对,在处理接触问题时,收敛困难这一问题普遍存在.以使用ANSYS Workbench对齿轮压缩机法兰连接的有限元分析为例,详细分析了影响计算收敛性的主要因素:接触对的建立、接触类型和接触算法的选取,以及罚刚度大小.在模拟过程中,用非对称绑定接触模拟螺母与螺栓的连接,非对称摩擦接触模拟螺母与法兰,以及法兰之间的连接,线性施加预紧力的方式来模拟螺栓拧紧.在数值计算过程中,绑定接触采用罚函数法,摩擦接触采用增广拉格朗日乘子法.初次计算采用一个较小的罚刚度系数,然后根据计算过程中接触迭代计算和N-R残差检查情况,逐步增大罚刚度系数直至得到收敛的计算结果.计算结果收敛性好,接触面摩擦应力分布符合压应力分布规律且仅有少量的滑移,满足工程使用要求.为其他类型接触问题的分析过程中解决收敛困难的问题提供了参考.

The numerical simulation of bolted flange joints and the contact convergence analysis

The numerical simulation process of bolted flange joints involves many contact pairs, and the difficulty in the convergence of numerical simulation commonly exists for contact problem. Taking the finite element analysis for the flange joints of geared compressor by ANSYS workbench as an example, we analyzed the main factors influencing computational convergence in detail, including the construction of contact pairs, the determination of contact type, the selection of proper algorithm and the value of F_KN(contact stiffness factor). In the process of simulation, unsymmetrical bonded contact was applied to the connection between nuts and bolts, and unsymmetrical frictional contact was applied to the contact on two flange surfaces and the connection between nuts and flanges. Moreover, the screwing of bolts was simulated by applying bolt pretension in linear. In the process of the numerical computation, pure penalty and augmented Lagrangian were applied to the bonded contact and the frictional contact, respectively. with a small F_KN in the inifial computation, it was gradually increased according to the result of the contact iteration and N-R residual until the computation converged. The convergence of the example is well, the frictional stress on the contact surfaces conforms to the distribution of compression stress and there is only very small slide on the contact surfaces, so the computation result meets engineering utilization requirements. At the same time, it offers some reference to resolve the convergence problem of other''s contact type.