| 引用本文: | 孙旺,朱平,严宏鑫.基于BP和RBF神经网络对静电纺丝工艺参数的优化研究[J].材料科学与工艺,2023,31(3):56-62.DOI:10.11951/j.issn.1005-0299.20220294. |
| SUN Wang,ZHU Ping,YAN Hongxin.Optimization of electrospinning process parameters based on BP and RBF neural networks[J].Materials Science and Technology,2023,31(3):56-62.DOI:10.11951/j.issn.1005-0299.20220294. |
|
| 摘要: |
| 针对静电纺丝在制备过程中易受到如聚合物含量、电压、推进速度和接收距离等工艺参数影响的问题,提出一种静电纺丝工艺参数的优化方法,以提升纳米纤维制备效率。以聚乳酸纳米纤维膜为研究对象,采用纤维直径为性能评价指标,设计实验获得训练和测试样本,借助BP(Back Propagation)和RBF(Radial Basis Function)神经网络构建不同工艺参数下的预测模型。结果表明:BP和RBF神经网络模型均能较好的对纤维直径进行预测,但RBF神经网络模型预测精度更高,其平均绝对误差(MAE)为12.125 nm,相对误差不超过7%。RBF神经网络建立的预测模型具有更高的稳定性,模型泛化能力更好,综合预测性能更加优越。所建立的模型可以帮助研究人员制备具有确定纤维直径的静电纺丝纳米纤维膜,实现对工艺参数的优化。 |
| 关键词: 静电纺丝 纳米纤维 RBF神经网络 纤维直径预测 工艺参数优化 BP神经网络 |
| DOI：10.11951/j.issn.1005-0299.20220294 |
| 分类号:TS102；TQ340.64 |
| 文献标识码:A |
| 基金项目:装发快速扶持项目第二阶段(61409220157). |
|
| Optimization of electrospinning process parameters based on BP and RBF neural networks |
|
SUN Wang,ZHU Ping,YAN Hongxin
|
|
(School of Instrument and Electronics, North University of China, Taiyuan 030051, China)
|
| Abstract: |
| Electrospinning is easily affected by process parameters such as polymer content, voltage, propulsion speed and receiving distance,which results in low preparation efficiency. To improve the preparation efficiency of nanofibers,an optimization method of electrospinning process parameters is proposed. Taking the polylactic acid (PLA) nanofiber membrane as the research object, the fiber diameter is used as the evaluation index of the membrane quality. And the training and test samples are obtained through the experiments. Finally, using BP (Back Propagation) and RBF (Radial Basis Function) neural networks to establish the prediction models under different process parameters. The results indicate that both BP and RBF neural network models can predict the fiber diameter well. But the RBF neural network model has higher prediction accuracy, with the mean absolute error (MAE) of 12.125 nm and the relative error of less than 7%. Therefore, the RBF neural network model can help researchers to prepare electrospinning nanofiber membranes with defined fiber diameters and optimize the process parameters. |
| Key words: electrostatic spinning nanofibers RBF neural network fiber diameter prediction optimization of process parameters BP neural network |
