基于BP神经网络算法的烟草机械塑料齿轮早期故障监测与优化  

Early Fault Monitoring and Optimization of Plastic Gears of Tobacco Machinery Based on BP Neural Network Algorithm

作　　者：王欣[1] 卢俊[2] 徐智[1] 张家海[1] WANG Xin;LU Jun;XU Zhi;ZHANG Jia-hai(School of Mechanical and Electrical Engineering,Sanjiang University,Nanjing 210012,China;China Tobacco Jiangsu Industrial Co.,Ltd.,Nanjing Cigarette Factory,Nanjing 210019,China)

机构地区：[1]三江学院机械与电气工程学院,江苏南京210012 [2]江苏中烟工业有限责任公司南京卷烟厂,江苏南京210019

出　　处：《塑料科技》2021年第2期91-94,共4页Plastics Science and Technology

摘　　要：基于人工神经网络的故障识别方法,对塑料齿轮的早期故障信号进行检测,实时发现齿轮故障的进化状态。推导了人工神经网络在塑料齿轮信号监测中的理论模型,比较了不同激活函数对模型结果的影响。选取了磨损、胶合、裂纹、断齿以及正常齿轮的信号进行研究,结果表明:验证集损失和训练集损失随着迭代次数的增加而一直降低;当模型中神经元数目为240时,模型测试识别的准确率约为98.6%。Based on the fault identification method of artificial neural network,the early fault signal of the plastic gear is detected,and the evolution state of the gear fault is found in real time.The theoretical model of artificial neural network in plastic gear signal monitoring is deduced,and the influence of different activation functions on the model results is compared.The signals of wear,gluing,cracks,broken teeth,and normal gears are selected for research.The results show that:the loss of the verification set and the loss of the training set decrease as the number of iterations increases;when the number of neurons in the model is 240,the model tests The recognition accuracy is about 98.6%.

关 键 词：烟草机械 塑料齿轮 人工神经网络 故障识别 激活函数 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程] TQ320.53[自动化与计算机技术—控制科学与工程]