Data-driven Quality Control Methods For Manufacturing Process Of Traditional Chinese Medicines

With the rapid growth of traditional Chinese medicine(TCM)market,the requirements on the quality of TCM have become more and more strict.Batch-to-batch quality consistency of TCM has always been an important issue for the development of TCM.There is a great need to improve the quality control strategy in TCM manufacturing.In recent years,various kinds of modern analytical techniques have been used for the measurement of raw materials,in-process materials and final products.Computer integrated process systems have been developed to better manage the manufacturing data.However,with the shortage of data analysis techniques,the manufacturing data are still used on a low level,which could have played a key role in the improvement of process quality control strategy.In this work,data-driven methodology has been introduced into the process quality control of TCM manufacturing.The applications of data-driven methodology for critical process parameter(CPP)identification,process modelling,monitoring and optimization have been well demonstrated with several case studies.The main contents and achievements of this dissertation are summarized as follows:1.With multiple process parameters and quality attributes involved in TCM manufacturing,a method integrating multi-targets for screening CPPs and a method using near infrared spectroscopy(NIRS)for rapid screening CPPs were established respectively.Taking the ethanol precipitation process of Danhong Injection as a case study,these two methods were applied successfully,which were shown to be effective strategies for the identification of CPPs.2.With the intricate mechanisms of TCM manufacturing,data-driven process modelling methods were proposed.In lab scale,a rapid process modelling method based on direct analysis in real time mass spectrometry(DART-MS)was proposed.And in industrial scale,data-driven models were established using data from the TCM manufacturing information management system.These two methods were applied successfully in case studies of the chromatographic process and the ethanol precipitation process,which were shown to be useful strategies for process understanding of TCM manufacturing.3.To compensate the shortage of process monitoring in TCM manufacturing,a process monitoring and endpoint determination method based on ultraviolet spectroscopy was established and applied in the chromatographic process of Danshen Multi-phenolic-acids Injection.Moreover,a method for process monitoring and fault diagnosis based on HPLC-MS fingerprints was proposed and applied in the ultrasonic extraction process of Danshen.These methods were helpful to identify and eliminate process deviations in a timely manner.4.To minimize the impact of quality variation of raw materials on the final products,a feedforward control strategy for TCM manufacturing was proposed,in which process parameters were optimized according to the input materials.The method was applied successfully in the ethanol precipitation processes of Danshen Injection and Danhong Injection,which provided an effective strategy to improve the quality of TCM.5.To improve the batch-to-batch quality consistency of TCM,a batch mixing method was proposed,in which different batches of herbal extracts were mixed with a well-designed proportion to meet the quality standard.In order to provide an acceptable efficiency for industrial production,the method uses minimum number of batches of extracts to meet the quality limits.The method can be used in industrial production to reduce the quality variation of TCM.