中药制造炼蜜过程中水分的近红外光谱在线检测方法研究  

Development on near infrared spectroscopy method of in-process quantitative analysis of moisture content during honey refining

作　　者：戴平 曾敬其 胡小艳 王静 韩燕雨[2,3] 王晓萌 张晓梦 杜菁 戚武振[2,3] 林羽 吴志生[1,2,3] DAI Ping;ZENG Jing-qi;HU Xiao-yan;WANG Jing;HAN Yan-yu;WANG Xiao-meng;ZHANG Xiao-meng;DU Jing;QI Wu-zheng;LIN Yu;WU Zhi-sheng(College of Pharmacy,Fujian University of Traditional Chinese Medicine,Fuzhou 350122,China;School of Chinese Materia Medica,Beijing University of Chinese Medicine,Beijing 102488,China;Engineering Research Center of Chinese Medicine Production and New Drug Development,Ministry of Education,Beijing 102488,China;National Engineering Research Center for R&D of TCM Multi-ingredient Drugs,Institute of Science,Beijing Tongrentang Co.,Ltd.,Beijing 100079,China)

机构地区：[1]福建中医药大学药学院,福建福州350122 [2]北京中医药大学中药学院,北京102488 [3]中药制药与新药开发教育部工程研究中心,北京102488 [4]北京同仁堂股份有限公司科学研究所中药复方新药开发国家工程研究中心,北京100079

出　　处：《中草药》2023年第17期5522-5529,共8页Chinese Traditional and Herbal Drugs

基　　金：国家优秀青年科学基金资助项目(82022073);国家自然科学基金资助项目(82274110);中央高校基本科研业务费北京中医药大学揭榜挂帅项目(2022-JYB-JBZR-018);中央高校基本科研业务费北京中医药大学揭榜挂帅项目(2022-JYB-JBZR-019);国家医学攻关产教融合创新平台——中药智能制造工程(90010062820031)。

摘　　要：目的建立一种在中试规模稳定的中药制造炼蜜过程水分在线检测方法,提高炼蜜产品的质量均一性。方法采用旁路外循环策略构建中试规模炼蜜过程中近红外光谱(near infrared spectroscopy,NIRS)的在线测量装备,开发减压工艺炼蜜过程水分的NIRS在线检测方法。引入灰色关联度作为NIRS在线检测模型的性能评价指标,采用系统优化策略,分别对模型的光谱预处理方法、建模波段和多元校正算法进行优化。结果炼蜜原料批次间差异是影响模型预测性能的关键因素,建模波段和多元校正算法优化可以降低炼蜜原料批次间差异对模型的影响。最终,最优模型参数为光谱预处理方法选择傅里叶变换,建模波段1880~2040 nm,多元校正算法选择支持向量回归,模型校正集与预测集样品预测结果的相对偏差均小于5.00%。结论所建方法快速、无损且准确可靠,可以监测炼蜜过程中水分的动态变化,提高炼蜜产品质量一致性。Objective To establish an online moisture detection method for the honey refining procedure of traditional Chinese medicine manufacturing at a stable pilot production scale,improving the uniformity of refined honey.Methods A bypass external circulation strategy was adopted to construct online near-infrared spectroscopy(NIRS)measurement equipment for the pilot-scale refining process,and a method for online detection of moisture during the decompressed honey refining process using NIRS was developed. The grey correlation degree was introduced as the performance evaluation index of the NIRS online detection model, and a system optimization strategy was adopted to optimize the spectral preprocessing method, modeling bands, and multivariate correction algorithm. Results The variations between batches of the refining materials were found to be a key factor affecting the model prediction performance, and optimization of the modeling bands and multivariate correction algorithm could reduce the impact of the variations between batches of the refining materials on the model. Finally, the optimal parameters of the detection method were obtained as follows: the spectral preprocessing method was Fourier transform, the modeling band was 1880—2040 nm, and the multivariate correction algorithm was support vector regression. The relative deviations between the reference value and the predicted value of all samples from the calibration set and prediction set were all less than 5.00%. Conclusion The method had the advantages of quickness, non-destructiveness and precision, which could be applied to monitor the dynamic change of moisture content during honey refiningand to improve the consistency of refined honey.

关 键 词：中药制造 炼蜜 近红外光谱 水分 在线检测 灰色关联度 

分 类 号：R283.6[医药卫生—中药学]