中药提取过程在线紫外动态趋势回归分析及终点判定

发布时间：2018-10-21 07:45

【摘要】：随着中药生产行业的转变,中药生产过程现代化成为了必然趋势。而目前由于缺乏有效的在线检测手段,中药生产检测滞后性大,产品质量参差不齐。针对中药提取过程在线监控及终点判定的问题,基于朗伯比尔定律及中药提取动力学方程建立提取液紫外吸光度随时间变化模型,并提出动态分析及终点判断方法,该方法包括紫外吸光度序列拟合、稳健性分析以及终点计算。以千年健提取过程为例,使用本实验室自主研发的中药提取液在线采样系统进行光谱收集,并对离线样本求取其有效固体物质含量作为参考。分析过程中,首先对实时检测的光谱进行插值及滤波处理,计算230.2~400nm紫外谱区的吸光度均值并结合前期数据组成吸光度均值序列;随后,进行了该吸光度序列与提取液固体含量之间的相关分析,得到线性相关系数r~2=0.9828,证明二者存在较强的线性关系;最终对吸光度均值序列进行动态模型稳健回归及提取终点判定。结果表明:回归过程通过稳健性分析能够充分识别出测量异常点,提高了拟合曲线与原始吸光度的复相关系数,使其达到0.99,并通过终点判定将千年健提取时间由最初人工设定的180min缩短至122min。实验证明,紫外在线监控方法实现了提取过程的在线监控及终点判定,对稳定产品质量、提高经济效益具有重要意义。
[Abstract]:With the change of traditional Chinese medicine production industry, modernization of traditional Chinese medicine production process has become an inevitable trend. At present, due to the lack of effective online detection means, the production of traditional Chinese medicine detection lag, product quality is uneven. In order to solve the problem of on-line monitoring and endpoint determination of traditional Chinese medicine extraction process, based on Lambert's law and extraction kinetics equation of traditional Chinese medicine, a model of ultraviolet absorbance variation with time was established, and a dynamic analysis and an endpoint judgment method were put forward. The method includes UV absorbance sequence fitting, robustness analysis and end point calculation. Taking the extraction process of millennium health as an example, the on-line sampling system of traditional Chinese medicine extract developed by our laboratory was used for spectral collection, and the effective solid matter content was obtained from off-line samples as a reference. In the process of analysis, the spectrum of real-time detection is firstly interpolated and filtered, and the absorbance mean value of 230.2~400nm ultraviolet spectrum region is calculated and combined with the previous data to form the absorbance mean sequence. The correlation between the absorbance sequence and the solid content of the extract was analyzed, and the linear correlation coefficient was obtained, which proved that there was a strong linear relationship between the two. Finally, the dynamic model robust regression and the extraction end point of the absorbance mean sequence were carried out. The results show that the regression process can fully identify the abnormal points by robust analysis, and improve the complex correlation coefficient between the fitting curve and the original absorbance. The extraction time of millennium health was shortened from the original manual set 180min to 122 minutes. It is proved by experiments that the method of ultraviolet on-line monitoring can realize the on-line monitoring and determination of the end point of the extraction process, which is of great significance for stabilizing the product quality and improving the economic benefit.
【作者单位】：浙江大学工业控制技术国家重点实验室;
【基金】：国家(863计划)项目(2012AA10A503)资助
【分类号】：O657.3;TQ461

【相似文献】