Milk Piece Of Rapid Extinction Piece Production Process Quality Control Method

To improve the level of Traditional Chinese Medicine (TCM) manufacturing process, promote the technology progress of traditional chinese medicine industry, guarantee the security and effectiveness of TCM product. In this study, Rukuaixiao tablet is taken as an example to make a research of developing a method to quickly controlling the quality of alcohol deposition, drying granulation and coating, furthermore, it can provide a research idea for process on-line quality control.The rapid determination method of alcohol deposit fluid was established by NIRS. Synergy interval partial least squares(SiPLS) was used to build the calibration models with HPLC as a reference method for Tanshinol, Protocatechuic Aldehyde, Hesperidin and Salvianolic acid B. Training and testing sets were partitioned by SPXY algorithm, Pretreatments of spectra were discussed in detail. Correlation coefficients of prediction are 0.9968,0.9901,0.9948 and 0.9960; RMSECVs are 0.0105,0.0017,0.0075 and 0.0241;RMSEPs are 0.0051,0.0011,0.0069 and 0.0198 for Tanshinol, Protocatechuic Aldehyde, Hesperidin, Salvianolic acid B, respectively.A rapid quantitative analysis method of index components was developed for drying granulation. Backward interval partial least squares(BiPLS) was used to build the calibration models with HPLC as a reference method for Tanshinol, Protocatechuic Aldehyde, Hesperidin and Salvianolic acid B. Training and testing sets were partitioned by SPXY algorithm, Pretreatments of spectra were discussed in detail. Correlation coefficients of prediction are 0.8822,0.9766,0.8529 and 0.9829; RMSECVs are 0.4131,0.0235,0.1340 and 0.5983; RMSEPs are 0.2475,0.0421, 0.2410 and 0.4362 for Tanshinol, Protocatechuic Aldehyde, Hesperidin, Salvianolic acid B, respectively.A rapid analysis method of moisture content was developed for drying granulation. Water content of samples were measured by weightlessness method as reference values. First training and testing sets were partitioned by SPXY algorithm, then, smoothing and autoscaling were used to pretreat spectra, and calibration model had been built by backward interval partial least squares(BiPLS). Correlation coefficient of prediction is 0.9625, RMSECV is 0.2385, RMSEP is 0.3739.To establish a rapid and nondestructive method to determine coating thickness of Ruikuaixiao. Partial least squares (PLS) was applied to build the calibration models, and Coating thicknesses were measured by Spiral Micrometer as reference values. Training and testing sets were partitioned by Kernard-stone algorithm. Different calibration methods were compared, spectral pretreating methods, wavenumber selection and factors of NIRS were discussed in detail. Correlation coefficients of prediction, root mean square error of calibration, and root mean square error of prediction obtained by PLSR model were 0.9938,0.0149 and 0.0194 respectively.As those works shown, the NIRS technique could be applied in quality analysis of TCM manufacturing process, and developed to be a new method of monitoring the quality of TCM manufacturing process, it will help to settle some problems in the modernization of TCM.