挖掘机斗杆结构应力神经网络建模  

Modeling of Stress Neural Network for Excavator Bucket Rod Structure

作　　者：李军 罗承旭 林述温[1] LI Jun;LUO Chengxu;LIN Shuwen(School of Mechanical Engineering and Automation,Fuzhou University,Fuzhou 350108,China)

机构地区：[1]福州大学机械工程及自动化学院,福建福州350108

出　　处：《机械制造与自动化》2024年第4期154-160,共7页Machine Building & Automation

摘　　要：为了研究斗杆结构应力神经网络建模方法,在分析不同典型工况下斗杆结构的应力分布特征基础上,确定斗杆结构的应力特征截面,分析斗杆结构各尺寸参数对各特征截面应力的灵敏度,据此确定斗杆结构应力神经网络模型的输入变量,进行结构应力神经网络建模,并对不同神经网络模型结构进行预测误差对比分析。结果表明:所建立的模型可快速预测斗杆结构特征截面的应力,实训样本预测误差小于10%。To study the modeling method of the lever structure of stress neural network,based on the analysis of the stress distribution characteristics of the bucket lever structure under different typical conditions,determines the stress characteristic section of the bucket lever structure,and analyzes the sensitivity of the size parameters of the bucket lever structure to the stress of each feature section.Accordingly,the input variables of the stress neural network model of the bucket lever structure is defined,the model of the structural stress neural network is established,and the prediction error of different neural network model structures is compared and analyzed.The results show that the established neural network model of bucket structure can swiftly predict the stress of bucket structure section with prediction error less than 10%.

关 键 词：挖掘机 斗杆结构 BP神经网络 应力预测 应力特征截面 

分 类 号：TP391.9[自动化与计算机技术—计算机应用技术]