【摘要】：采用SLA 3D打印機(jī)打印不同參數(shù)的同一模型,測(cè)量成型件模型尺寸參數(shù),并利用改進(jìn)的LSSVM模型對(duì)不同參數(shù)的成型件尺寸誤差進(jìn)行預(yù)測(cè)。首先分析主要影響SLA 3D打印模型質(zhì)量的原因,確定四個(gè)主要因素:疊層厚度,模型擺放角度和支撐密度,接觸點(diǎn)大小。設(shè)計(jì)試驗(yàn),采用SLA 3D打印機(jī)在此參數(shù)下打印,再對(duì)打印成型件進(jìn)行測(cè)量確定成型件尺寸信息及尺寸誤差,基于已有數(shù)據(jù)建立改進(jìn)的LS-SVM模型對(duì)不同打印參數(shù)下的成型件的尺寸誤差進(jìn)行預(yù)測(cè)。結(jié)果表明模型預(yù)測(cè)正確率達(dá)到92.6471%,改進(jìn)的LS-SVM相較于原尋優(yōu)方法及BP神經(jīng)網(wǎng)絡(luò)對(duì)SLA 3D打印尺寸誤差預(yù)測(cè)有良好的效果。
[Abstract]:SLA 3D printer is used to print the same model with different parameters to measure the dimension parameters of the model, and an improved LSSVM model is used to predict the dimension error of the model with different parameters. Designing experiment, using SLA 3D printer to print under this parameter, then measuring and determining the size information and size error of the printed parts. Based on the existing data, an improved LS-SVM model is established to predict the size error of the formed parts under different printing parameters. The results show that the prediction accuracy of the model reaches 92. 6471%. Compared with the original optimization method and BP neural network, the improved LS-SVM has a good effect on the prediction of dimensional error of SLA 3D printing.
【作者單位】： 上海交通大學(xué)機(jī)械與動(dòng)力工程學(xué)院;
【基金】：上海市科委項(xiàng)目(15111102203;16111106102) 上海交通大學(xué)醫(yī)工(理)交叉基金資助(YG2014MS04;YG2015MS09)
【分類號(hào)】：TP18;TP334.8
，

本文編號(hào)：2216718