面向装配特征的机械设备评估方法  

Structural Feature-oriented Design Method of Machinery Structure

作　　者：张胜文[1,2] 孙启胜 崔家源 方喜峰 ZHANG Shengwen;SUN Qisheng;CUI Jiayuan;FANG Xifeng(Jiangsu University of Science and Technology,Zhenjiang Jiangsu 212003,China;Jiangsu Provincial Key Laboratory of Advanced Manufacturing for Machinery and Equipment,Zhenjiang Jiangsu 212003,China)

机构地区：[1]江苏科技大学,江苏镇江212003 [2]江苏省船海机械装备先进制造重点实验室,江苏镇江212003

出　　处：《机械设计与研究》2020年第6期125-129,共5页Machine Design And Research

基　　金：国防基础科研基金资助项目(B0720060844)。

摘　　要：针对装配特征分析技术相对缺乏、历史数据挖掘尚不充分等问题,提出了基于神经网络、面向装配特征的机械设备评估技术。针对基础特征给出了形式化描述标准,结合神经网络将机械结构的历史装配数据进行聚类泛化,并以多种相似度比对方法以及层次分析法为判定依据,评估并推理出装配对象以及装配方式,辅助完成机械设备三维模型的结构设计。以制麦皮带运输机为设计对象,通过SolidWorks软件的二次开发,开发了机械结构设计模块,辅助完成了对设备的设计和调整,提高了设计工作的效率。Aiming at the problems of relatively lacking assembly feature analysis technology and inadequate historical data mining,an assembly technology-oriented machinery and equipment evaluation technology based on neural network was proposed.A formal description standard was given for basic features.Combined with neural network,the historical assembly data of mechanical structure was clustered and generalized.Based on multiple similarity comparison methods and analytic hierarchy process,the assembly objects and assembly methods were evaluated and inferred to assist in the completion of the structural design of the three-dimensional model of mechanical equipment.Taking the wheat belt conveyor as the design object,through the secondary development of SolidWorks software,a mechanical structure design module was developed to assist in the design and adjustment of the equipment and improve the efficiency of the design work.

关 键 词：机械结构 装配特征 神经网络 

分 类 号：TP271[自动化与计算机技术—检测技术与自动化装置]