基于BP神经网络的覆膜砂选择性激光烧结件精度预测  

Precision prediction for SLS of Resin Coated Sand Based on BP Nneural Network

作　　者：刘兆平[1] 王宏松[1] 修辉平[1] 

机构地区：[1]九江职业技术学院,江西九江332007

出　　处：《热加工工艺》2016年第21期91-93,共3页Hot Working Technology

摘　　要：采用正交试验与反向传播(BP)神经网络结合的方法对覆膜砂选择性激光烧结的工艺进行优化。运用BP神经网络建立尺寸精度(xy向收缩)与选择性激光烧结工艺参数之间的预测模型,利用正交试验样本对所建立的神经网络进行训练,形成输入与输出之间的高度映射关系,并验证神经网络模型的可行性。在此基础上,预测各因素在不同水平下覆膜砂选择性激光烧结件的最优精度。Orthogonal experiment and back propagation （BP） neural network were adopted to optimize the technology ofselective laser sintering for resin coated sand. BP neural network was used to build the prediction model between dimensionalaccuracy （shrinkage in x and y direction） and process parameters of selective laser sintering, and orthogonal test samples wereused to train the established neural network to format input-output mapping relationship, and verify the feasibility of the neuralnetwork model. On this basis, the optimal precision of different factors under different levels of selective laser sintering forresin coated sand was forecasted.

关 键 词：正交试验 BP神经网络 覆膜砂 选择性激光烧结 精度 

分 类 号：TG242[金属学及工艺—铸造]