大型组合结构整体性分析的人工神经网络模型  

Artificial neural network model of the unity of analysis of large composite structure

作　　者：吴生富[1] 聂绍珉[1] 金淼[1] 李树奎[1] 

机构地区：[1]燕山大学机械工程学院,秦皇岛066004

出　　处：《塑性工程学报》2007年第2期77-79,93,共4页Journal of Plasticity Engineering

摘　　要：大型装备中普遍采用组合结构,其整体工作性能取决于预紧元件的预紧参数。确定预紧参数的传统方法不适用于大型结构,有限元法的分析次数又过多,实际上无法实现。文章提出用人工神经网络取代有限元进行整体性分析的方法,并针对某大型压机建立了神经网络模型。结果表明,用该方法可快速完成复杂组合结构的整体性分析,且精度较高。Composite structure is usually adopted in large equipment, which unity of work performance is depended on the preloading parameter of preloading element. Traditional means aren＇t applicable on large structure, and the finite element methods cost too much time, so it can not be achieved in fact. In this paper, artificial neural network is used to replace the finite element methods to run the unity of analysis, and how to build the neural networks is also discussed.

关 键 词：组合结构 有限元 神经网络 

分 类 号：TG315.4[金属学及工艺—金属压力加工]