基于BP神经网络的FDM成型件精度预测  

Precision Prediction of FDM Parts Based on the BP Neural Network

作　　者：王宏松[1] 修辉平[1] 韩燕[1] Wang Hongsong;Xiu Huiping;Han Yan(Jiujiang Vocational and Technical College,Jiujiang,Jiangxi,332007)

机构地区：[1]九江职业技术学院,江西九江332007

出　　处：《九江职业技术学院学报》2018年第4期8-11,共4页Journal of Jiujiang Vocational and Technical College

摘　　要：通过正交试验研究了FDM成型工艺参数对成型件尺寸和几何精度的影响。以FDM成型工艺参数为输入,成型件几何精度和尺寸精度作为输出,构建BP神经网络模型,利用正交试验所获样本对所建立的神经网络进行训练,建立起工艺参数与成型件精度之间的非线性映射关系,并验证神经网络模型的可靠性。在此基础上,对FDM成型件精度进行预测,便于指导生产实践。The influence of FDM molding process parameters on the dimensional and geometric precision of molded parts was studied by orthogonal test.The BP neural network model is constructed with FDM forming process parameters as input and geometrical precision and size precision of molded parts as output.The neural network is trained by the samples obtained by orthogonal test,and the mapping relationship between process parameters and the accuracy of forming parts are established.And the reliability of the neural network model is verified.On this basis,the optimal precision of FDM molded parts can be estimated,which can be used to guide the production practice.

关 键 词：正交试验 FDM BP神经网络 精度 预测 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程] TH164[自动化与计算机技术—控制科学与工程]