神经网络优化膳食营养补充剂胶囊支架3D打印工艺  

Optimization of 3D printing technology of nutraceuticals capsule scaffold using neural network

作　　者：陈虹竹 张良[2] 姚佳 胡小佳[2] 刘倩楠[2] 刘伟[2] 孙钦秀 胡宏海[2] 刘书成 Chen Hongzhu;Zhang Liang;Yao Jia;Hu Xiaojia;Liu Qiannan;Liu Wei;Sun Qinxiu;Hu Honghai;Liu Shucheng(College of Food Science and Technology,Guangdong Ocean University,Guangdong Provincial Key Laboratory of Aquatic Product Processing and Safety,Guangdong Province Engineering Laboratory for Marine Biological Products,Guangdong Provincial Engineering Technology Research Center of Seafood,Key Laboratory of Advanced Processing of Aquatic Product of Guangdong Higher Education Institution,Zhanjiang 524088,China;Key Laboratory of Agro-products Processing,Ministry of Agriculture and Rural Affairs,Institute of Food Science and Technology,Chinese Academy of Agricultural Sciences,Beijing 100193,China;School of Biomedicine,Beijing City University,Beijing 100083,China)

机构地区：[1]广东海洋大学食品科技学院,广东省水产品加工与安全重点实验室,广东省海洋生物制品工程实验室,广东省海洋食品工程技术研究中心,水产品深加工广东普通高等学校重点实验室,湛江524088 [2]中国农业科学院农产品加工研究所,农业农村部农产品加工综合性重点实验室,北京100193 [3]北京城市学院生物医药学部,北京100083

出　　处：《农业工程学报》2021年第4期18-23,共6页Transactions of the Chinese Society of Agricultural Engineering

基　　金：中央级公益性科研院所基本科研业务费专项(S2019XK01);农业农村部财政经费项目;广东普通高等学校海洋食品绿色加工技术研究团队(2019KCXTD011)。

摘　　要：为了确定熔融沉积成型3D打印技术制备个性化膳食营养补充剂胶囊支架的工艺条件,打印9通道聚乳酸(Polylactic Acid,PLA)胶囊支架结构,采用Box-Behnken设计,以支架通道的面积均方误差(Mean-Square Error,MSE)作为响应指标,通过神经网络模型模拟温度、打印头孔径和速度3个因素对胶囊支架打印效果的影响。结果表明:3×4×1结构3层神经网络模型能够较好拟合3D打印过程(R2=0.998),当孔径为0.3 mm时,随着打印温度和打印速度的增大,MSE均低于10%。打印温度过高或打印头孔径过大时,胶囊支架内部会出现拉丝现象。综合考虑打印能耗和效率,确定当打印头孔径为0.3mm,打印速度25~30mm/s,打印温度介于173~180℃之间,MSE范围为1%~2%时,胶囊支架打印精度和微观结构较好。研究结果为3D打印精细化结构及打印参数优化提供理论参考。Nutraceuticals are herbal and food products with high nutritional values,which are commonly used to treat or prevent diseases.In personalized nutrition,there is an inter-individual response to nutraceuticals intervention,due to the fact that a sub population can benefit more than others.The traditional production of dietary supplements or functional food has very limited flexibility of fabrication to achieve personalized customization with nutrients or functional factors.The oral dosage forms can inevitably lead to insufficient or excessive intake.A type of 3 D printing technology,fused deposition modeling(FDM)has been selected to create complex and accurate shapes using food-derived thermoplastic materials.Particularly,polylactic acid(PLA)has been successfully used to produce the scaffold of oral capsule.In this study,a 9-channel PLA capsule scaffold structure and Box-Behnken design were used to explore the effects of printing temperature,nozzle diameter,and printing speed on the FDM 3 D printing accuracy.Meanwhile,the printing channel area error(MSE)was used as the response index during the optimization.A 3×4×1 three-layer neural network(NN)model was established.The coefficients of determination(R2)were above 0.998,indicating an excellent fitness between the actual and predicted MSE in the NN model.The nozzle diameter was the most influential factor.The MSE remained almost unchanged with the increase of printing temperature and speed at the nozzle diameter of 0.3 mm,where the MSE of the capsule channel area was less than 10%.In terms of energy consumption and efficiency,a low level of MSE was obtained with a lower temperature and faster speed.No significant difference was observed between the prediction values of the NN model and the experimental ones.The results indicated that the NN model can be used to predict the process of FDM 3 D printing.An optimal combination of parameters was achieved,where the nozzle diameter was 0.3 mm,and the printing speed ranged from 25 mm/s to 30 mm/s,while the printing tempera

关 键 词：神经网络 工艺优化 3D打印 熔融沉积成型 胶囊支架 

分 类 号：TS201.4[轻工技术与工程—食品科学]