基于BP神经网络的机电产品清单数据智能补全算法研究  

Research on Intelligent Completion Algorithm of Mechanical and Electrical Product List Data Based on BP Neural Network

作　　者：丁永新 李涛[2] 郭志威 张洪潮[3] Ding Yongxin;Li Tao;Guo Zhiwei;Zhang Hongchao(School of Mechanical Engineering,Dalian University of Technology,Dalian,Liaoning 116081,China;Institute of Major Equipment Design,Dalian University of Technology,Dalian,Liaoning 116081,China;Department of Industrial Engineering,Texas Tech University,Lubbock,TX76001,USA)

机构地区：[1]大连理工大学机械工程学院,辽宁大连116081 [2]大连理工大学重大装备设计研究所,辽宁大连116081 [3]德克萨斯理工大学工业工程系,美国德州TX76001

出　　处：《机电工程技术》2023年第1期142-145,共4页Mechanical & Electrical Engineering Technology

基　　金：国家重点研发计划项目(编号:2020YFB1711603)。

摘　　要：生命周期评价是一种分析机械产品的全生命周期环境影响的有力工具,但是其前期的清单收集工具极其繁琐且耗时,为了简化机电产品生命周期评价的步骤中清单数据收集工作,降低清单数据的数据填充难度,提出了一种基于BP神经网络的清单数据智能填充算法。该算法基于文本相似度和数据相似度算法分别计算该零部件的各项参数与案例库中已知零部件的各项参数信息之间的相似度,并把这些参数的相似度信息作为神经网络输入,经输出层计算出零部件之间整体的相似度,分别求取所有零部件和缺失数据的零部件之间的相似度信息并排序,从而求得零件库中与缺失信息的零部件之间最相似的零部件,用求出的最相似的这个零部件的信息完成缺失数据的零部件信息的智能填充。Life cycle assessment is a powerful tool for analyzing the environmental impact of mechanical products in the whole life cycle, but its early inventory collection tool is extremely cumbersome and time-consuming. Because of the difficulty of data filling in data, an intelligent filling algorithm for inventory data was proposed based on BP neural network. The similarity between the various parameter information, and the similarity information of these parameters was used as the input of the neural network, the overall similarity between the parts was calculated through the output layer, and all parts and parts with missing data were obtained respectively. The similarity information between them was sorted, so as to obtain the most similar parts in the parts library to the parts with missing information, and the obtained information of the most similar parts was used to complete the intelligent information of parts with missing data filling.

关 键 词：机电产品清单数据 BP神经网络 相似度 智能补全 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]