基于ANN-MC方法进行弹性机构的动态强度可靠性分析  

Dynamical Strength Reliability Analysis of a Flexible Mechanism Using Neural Networks

作　　者：杨鑫[1] 袁茹[1] 王莉[1] 

机构地区：[1]西北工业大学机电学院,西安710072

出　　处：《机械科学与技术》2008年第4期462-465,共4页Mechanical Science and Technology for Aerospace Engineering

基　　金：陕西省自然科学基础研究计划项目(2005E228)资助

摘　　要：弹性机构强度可靠性和运动可靠性是工业机器人、空间操作臂等领域最为关注的问题,而精确实用的可靠性分析方法是解决此问题的关键。本文在弹性机构动力学有限元分析的基础上,引入构件尺寸和运动副摩擦等随机变量,获得弹性机构随机动态应力的时间历程;将蒙特卡罗法(Monte-Carlo,MC)与人工神经网络结合(artificial neural network,ANN),形成了基于小样本对弹性机构动态强度可靠性进行计算的ANN-MC方法;将该方法用于曲柄滑块机构的动态强度可靠性计算,获得其动态强度可靠性为99.10%。本文方法不仅节省了计算时间,而且提高了动态可靠度求解精度。The dynamical strength reliability and motion reliability of a flexible mechanism are very important in the field of industrial robot and space operation. Precise and practical reliability analysis method is a key to solve such problems. This article introduces some random parameters such as dimension of the mechanism＇s components and friction of motion joint to analyze the maximum dynamical strength of the flexible mechanism using finite element method, and the dynamical strength proceeding that changes by time is obtained. The MC （Monte-Carlo）method and ANN （ Artificial Neural Network） method are combined to make an analysis of the dynamical strength reliabili- ty. With such a method, only a small sample is needed. ANN-MC method is also used to compute the dynamical strength reliability and the computed reliability of the flexible mechanism is 99.10 percent. The method not only decreases the computing time, but also enhances the computing precision of the strength reliability.

关 键 词：人工神经网络 蒙特卡罗 动态可靠度 弹性机构 

分 类 号：TH112[机械工程—机械设计及理论]