In fluidized bed drying processes, accurate monitoring of granule moisture content is critical for quality control. Traditional detection methods primarily rely on manual periodic sampling (e.g., every 5-10 min) for offline measurement, resulting in significant time lag and insufficient real-time performance, which makes it difficult to capture dynamic changes in moisture content. To address this issue, a TCN-GRU approach with near-infrared spectroscopy-assisted label expansion is proposed. This approach uses process parameters-specifically, inlet air temperature, outlet air temperature, and material temperature-as input variables to predict moisture content. Firstly, feature parameters that strongly correlate with moisture content are identified through correlation analysis between process parameters and moisture content during the drying process. To overcome the limitations of traditional feature extraction in capturing time-series dynamics, feature engineering techniques such as sliding window statistical, temperature change rate, temperature ratio, and temperature difference are introduced. Finally, a Temporal Convolutional Network-Gated Recurrent Unit (TCN-GRU) model is constructed. The TCN component efficiently extracts local temporal features via parallel processing, while the GRU captures longterm dependencies with lower computational complexity compared to LSTMs. Experimental results on a Cassia Twig drying dataset demonstrate the effectiveness of the proposed method, achieving a coefficient of determination (R2) of 0.9966, a mean absolute error (MAE) of 0.1903, and a mean squared error (MSE) of 0.0516. This method provides a cost-efficient "soft sensor" solution for the digital transformation in pharmaceutical manufacturing.

In this paper, a method for rapidly determining the content of chlorogenic acid, neochlorogenic acid, cryptochlorogenic acid, gardeniside, and strychnoside in Reduning Injection(RI) was established based on near-infrared spectroscopy(NIRS), midinfrared spectroscopy(MIRS), and spectral fusion technology. Six pretreatment methods and five variable screening methods were investigated, and the best method was selected to establish a partial least square(PLS) model of two single spectra. At the same time,the NIRS and MIRS were fused with equal weights and characteristic bands, and the PLS model was established. The prediction effect of the four models on the quality control components was compared: NIRS>characteristic band fusion>MIRS>equal weight fusion. The relative standard error of prediction(RSEP) of the NIRS models on the five quality control components was less than 2. 5%, and the ratio of performance to deviation(RPD) was greater than 9. 5. The results show that the single spectrum model of NIRS is the best quantitative detection method, and the model of NIRS combined with the PLS algorithm can be used for the rapid detection of Reduning Injection.

Objective To establish a rapid method for predicting the hardness of Rhodiola grandiflora Tablets（RGT） based on near infrared spectroscopy（NIRS）. Methods The NIRS of 600 production samples and self-made samples were collected. The partial least square（PLS） algorithm model was established by comparing the advantages and disadvantages of the models under different spectral pretreatment methods and different characteristic variable screening conditions. The NIRS of 120 samples were collected for external verification of the model to predict the hardness of RGT. Results Among the PLS models for plain tablet hardness, the full band model was the best for the spectrum without pretreatment. The correlation coefficient of training set(R cal ) and the correlation coefficient of verification set(R pre ) of the prediction model were 0.971 9 and 0.988 7, respectively. The root mean square error of prediction（RMSEP） is 2.03 N, the ratio of performance to deviation（RPD） is 6.68, and the relative standard to deviation（RSEP） is 4.24%. The average relative prediction error of the internal validation of the model is 2.82%, and the average relative prediction error of the external validation is 4.59%, both of which are less than 5%. The detection rate of unqualified tablets is as high as 97.33%.Conclusion The hardness prediction model of RGT established by near-infrared spectroscopy combined with partial least squares algorithm has good model performance and prediction ability. This study provides a new method for nondestructive testing of the hardness of RGT.

The mixing process is a critical link in the formation of oral solid preparations of traditional Chinese medicine. This paper took the extract powder of Guizhi Fuling Capsules and Paeonol powder as research objects. The angle of repose, loose packing density, and particle size of the two powders were measured to calibrate discrete element simulation parameters for the mixing process. The discrete element method was used to calibrate the simulated solid density of Paeonol powder and extract powder of Guizhi Fuling Capsules based on the Hertz-Mindlin with JKR V2 contact model and particle scaling. The Plackett-Burman experimental design was used to screen out the critical contact parameters that had a significant effect on the simulation of the angle of repose. The regression model between the critical contact parameters and the simulated angle of repose was established by the Box-Behnken experimental design, and the critical contact parameters of each powder were optimized based on the regression model. The best combination of critical contact parameters of the extract powder of Guizhi Fuling Capsules was found to be 0.51 for particle-particle static friction coefficient, 0.31 for particle-particle rolling friction coefficient, and 0.64 for particle-stainless steel static friction coefficient. For Paeonol powder, the best combination of critical contact parameters was 0.4 for particle-particle static friction coefficient and 0.19 for particle-particle rolling friction coefficient. The best combination of contact parameters between Paeonol powder and extract powder of Guizhi Fuling Capsules was 0.27 for collision recovery coefficient, 0.49 for static friction coefficient, and 0.38 for rolling friction coefficient. The verification results show that the relative error between the simulated value and the measured value of the angle of repose of the two single powders is less than 1%, while the relative error between the simulated value and the measured value of the angle of repose of the mixed powder with a mass ratio of 1∶1 is less than 4%. These research results provide reliable physical property simulation data for the mixed simulation experiment of extract powder of Guizhi Fuling Capsules and Paeonol powder.

Hard capsules of traditional Chinese medicine(TCM) have different degrees of hygroscopicity, which affects the stability and efficacy of drugs. In this paper, 30 kinds of commercially available TCM capsules were used as the research object. The hygroscopicity curves of capsule contents, capsule shells, and capsules were tested respectively, and the first-order kinetic equation was used for fitting. The results show that during the 24 h hygroscopicity process, the capsule shell can reduce the weight gain caused by the hygroscopicity of the contents by 0.80%-53.0% and the hygroscopicity rate of the capsule contents by 1.74%-91.3%, indicating that the capsule shell has a strong delay effect on the hygroscopicity of the contents of the TCM capsules. Seven physical parameters of the contents of 30 kinds of TCM capsules were determined, and 14 prescription process-related parameters were sorted out. A partial least squares model for predicting the hygroscopicity rate of the contents of TCM capsules(with shell) for 24 h was established. It is found that the hygroscopicity rate of the capsule shell is positively correlated with the hygroscopicity of the contents of TCM capsules(with shell), suggesting that the capsule shell with a low hygroscopicity rate is helpful for moisture prevention. In addition, the pre-treatment process route of the preparation and the type of molding raw materials affect the hygroscopicity. A larger proportion of the extract in the capsule content and a smaller proportion of the fine powder of the decoction pieces indicate stronger hygroscopicity of the capsule content. The 24 h hygroscopicity rate of 15% was used as the classification node of hygroscopicity strength, and the hygroscopicity rate constant of 0.58 was used as the classification node of hygroscopicity speed. The classification system of hygroscopicity behaviors of TCM capsules was established: the varieties with strong and fast hygroscopicity accounted for about 6.67%, while those with strong and slow hygroscopicity accounted for about 33.3%; the varieties with weak and fast hygroscopicity accounted for about 26.7%, while those with weak and slow hygroscopicity accounted for about 33.3%. The classification system is helpful to quantify and compare the hygroscopicity behavior of different TCM capsules and provides a reference for the quality improvement, moisture prevention technologies, and material research of TCM capsules.

目的 探究水-醇双提工艺下中药复方粉末影响颗粒流动性的关键物性参数.方法 以水-醇双提工艺下杏贝止咳颗粒(Xingbei Zhike Keli,XZK)、桂枝茯苓胶囊(Guizhi Fuling Jiaonang,GFJ)以及参乌益肾片(Shenwu Yishen Pian,SYP)3个中药复方品种的制粒前粉末与制粒后颗粒为研究对象,采用多元统计分析方法,绘制粉末物理指纹图谱,结合Pearson相关系数评价粉末质量一致性;采用主成分分析(principal component analysis,PCA)结合因子分析评价颗粒流动性;并构建以松装密度(Da)、振实密度(Dc)、休止角(a)、豪斯纳比(IH)、粒径＜50 μm百分比(Pf)、均匀性(HG)、均齐度(UN)、粒径(D10、D50、D60、D90)、分布宽度(span)、分布范围(width)、比表面积(SSA)、孔隙率(Ie)、卡尔指数(IC)、含水量(HR)、吸湿率(H)为自变量、以颗粒流动性总因子得分(TFS)为因变量的正交偏最小二乘法-判别分析(orthogonal partial least-squares discrimination analysis,OPLS-DA)模型,用以辨识粉末关键物性参数.结果 分别构建了 3个品种各15批粉末的物理指纹图谱,结合Pearson相关系数结果显示3个品种粉末质量一致性良好,其中SYP颗粒优于XZK颗粒再优于GFJ颗粒;计算得颗粒流动性TFS,整体来看流动性XZK颗粒优于SYP颗粒再优于GFJ颗粒;OPLS-DA模型优化后辨识出H、UN、SSA、HR为影响颗粒流动性的关键物性参数,置换检验结果表明模型有效可靠.结论 基于OPLS-DA模型辨识出影响水-醇双提工艺下中药复方粉末影响颗粒流动性的关键物性参数为H、UN、SSA、HR.

Objective Taking Guizhi Fuling Capsules(GFC,桂枝茯苓胶囊)and Tianshu Capsules(TC,天舒胶囊)as research objects,a rapid method for detecting the moisture content of two preparation intermediates was established by combining near-infrared spectroscopy(NIRS)technology with machine learning algorithms.Methods The NIRS of GFC total mixed particles and TC total mixed particles were collected.The effects of different preprocessing methods,variable screening methods and algorithms on the model were investigated.The optimal modeling conditions were selected to establish a universal NIRS quantitative model for moisture content of two intermediates.Results The generalized path seeker(GPS)algorithm was superior to the partial least squares(PLS)algorithm in establishing quantitative models for the same intermediate.Compared with the PLS universal model,the GPS universal model had higher predictive performance,with the relative standard errors of prediction(RSEP)decreasing from 3.17％to 3.03％,and the ratio of performance to deviation(RPD)increasing from 4.83 to 5.05.The GPS universal model could be used to predict the moisture content of intermediates,and there was little difference in prediction accuracy between GPS and that of the independent models.Conclusion The universal quantitative model established by GPS algorithm combined with NIRS technology could quickly and accurately determine the moisture content of two preparation intermediates.

The pharmaceutical manufacturing model is gradually changing from intermittent manufacturing to continuous manufacturing and intelligent manufacturing. This paper briefly reviewed the supervision and research progress in continuous pharmaceutical manufacturing in China and abroad and described the definition and advantages of continuous pharmaceutical manufacturing. The continuous manufacturing of traditional Chinese medicine(TCM) at the current stage was summarized in the following three terms: the enhancement of the continuity of intermittent manufacturing operations, the integration of continuous equipment to improve physical continuity between units, and the application of advanced process control strategies to improve process continuity. To achieve continuous manufacturing of TCM, the corresponding key technologies, such as material property characterization, process modeling and simulation, process analysis technology, and system integration, were analyzed from the process and equipment, respectively. It was proposed that the continuous manufacturing equipment system should have the characteristics of high speed, high response, and high reliability, "three high(H~3)" for short. Considering the characteristics and current situation of TCM manufacturing, based on the two dimensions of product quality control and production efficiency, a maturity assessment model for continuous manufacturing of TCM, consisting of operation continuity, equipment continuity, process continuity, and quality control continuity, was proposed to provide references for the application of continuous manufacturing technology for TCM. The implementation of continuous manufacturing or the application of key continuous manufacturing technologies in TCM can help to systematically integrate advanced pharmaceutical technology elements and promote the uniformity of TCM quality and the improvement of production efficiency.

目的 应用衰减全反射中红外光谱(mid-infrared spectroscopy,MIRS)技术建立桂枝茯苓胶囊(Guizhi Fuling Capsules,GFC)浓缩过程中没食子酸、芍药苷、苯甲酸、苯甲酰芍药苷及挥发油桂皮醛和肉桂酸的定量分析模型,实现GFC浓缩过程的质量控制.方法 以HPLC检测值为参照,采集GFC浓缩过程的MIRS,结合偏最小二乘(partial least square,PLS)法分别建立6种指标性成分的定量模型.结果 没食子酸、芍药苷、苯甲酸、肉桂酸、苯甲酰芍药苷及桂皮醛的校正集相关系数(rcal)分别为 0.992、0.977、0.986、0.985、0.974、0.980,验证集相关系数(rpre)分别为 0.985、0.961、0.988、0.993、0.978、0.975,校正均方根误差(corrected root mean square errors,RMSEC)分别为 0.132、0.771、0.042、0.044、0.075、0.185,预测相对偏差(relative standard error of prediction,RSEP)和相对误差均小于10％.结论 MIRS技术具有快速方便、结果可靠的优点,可以应用于GFC浓缩过程中挥发油桂皮醛和肉桂酸及其他指标性成分的测定,为GFC浓缩过程的在线监控提供了一种新方法.

目的 基于决策树算法,深入挖掘热毒宁注射液金银花青蒿醇沉过程(金青醇沉)数据并探究潜在生产规律,提升该过程质量控制水平.方法 依托数字化中药提取工厂数据平台收集205批金银花和青蒿浸膏(金青浸膏)历史数据并整合成数据矩阵.将数据集随机划分为训练集和测试集后分别采用分类与回归树(classification and regression tree,CART)、随机森林(random forests,RF)和TreeNet算法建立金青醇沉过程模型,比较各模型性能并基于历史数据划分关键变量控制范围.结果 RF和TreeNet模型性能较好且性能接近,综合各模型分析结果得出醇提罐料液比及金银花浓缩收率为重要的影响因素,对重要变量进行依存度分析并优选批次,并以优选批次的金银花分配浸膏质量及加醇量做控制图,密度为1.11 g/cm3的金银花浸膏分配控制范围为557.92～639.62 kg,加醇量的控制范围为3.370～3.828 m3;密度为1.12 g/cm3的金银花浸膏的控制范围为540.4～616.9 kg,加醇量的控制范围为3.317～3.859 m3.结论 决策树算法建立的金青浸膏醇沉过程模型能够有效地挖掘潜在的生产过程规律,为生产过程的质量控制提升提供技术支撑.

The purpose of the study is to present a nondestructive qualitative and quantitative approach of hard-shell capsule using near-infrared (NIR) spectroscopy combined with chemometrics. The Yaobitong capsule (YBTC) was used for demonstration of the proposed approach and the NIR spectra were collected using a handheld fiber probe (FP) without the damage of capsule shell. By comparing the differences and similarities of the NIR spectra of capsule shells, contents and intact capsules, a preliminary conclusion can be drawn that the NIR spectra contained the information of the contents. Characteristic variables were selected by competitive adaptive weighted resampling (CARS) method, and least squares support vector machine (LSSVM) method based on particle swarm optimization (PSO) algorithm was applied to the construction of quantitative models. The relative standard error of prediction (RSEP) values of five saponins including notoginsenoside R1, ginsenoside Rg1, Re, Rb1, and Rd were 3.240%, 5.468%, 5.303%, 5.043%, and 3.745%, respectively. In addition, for qualitative model, three different types of adulterated capsules were designed. The model established by data driven version of soft independent modeling of class analogy (DD-SIMCA) demonstrated a satisfactory result that all adulterated capsules were identified accurately after an appropriate number of principal components (PCs) were chosen. The results indicated that although the NIR spectra collection was affected by capsule shell, sufficient content information can be obtained for quantitative and qualitative analysis after combining with chemometrics. It further proved that acquired NIR spectra do contain the effective component information of the capsule. This study provided a reference for the rapid nondestructive quality analysis of traditional Chinese medicine (TCM) capsule without damaging capsule shell.

目的 基于近红外光谱(near infrared spectrum,NIRS)技术,建立一种快速预测天舒片崩解时间的方法.方法 采集39个批次共468个样品的NIRS,对比分类和回归树(classification and regression trees,CART)算法与偏最小二乘(partial least-square,PLS)算法2种模型的预测效果,建立天舒片崩解时间预测模型.结果 经基线校正处理后建立的CART模型性能最优.与PLS模型相比该模型将相对校正均方根偏差(relative root mean square error of correction,RRMSEC)由7.43％降低至4.94％,相对预测均方根偏差(relative root mean square error of prediction,RRMSEP)由7.84％降低至7.66％.结论 NIRS技术结合CART算法预测天舒片崩解时间是可行的,为天舒片崩解时间快速无损检测提供了一种新方法.

目的 比较不同算法对桂枝茯苓胶囊内容物吸湿性预测模型性能的影响,确定最优建模算法.方法 以54个物理性质参数为输入,胶囊内容物吸湿性为输出,对比偏最小二乘算法(partial least squares,PLS)、决策树算法(classification and regression tree,CART)、多元自适应回归样条算法(multivariate adaptive regression splines,MARS)和广义路径追踪算法(generalized path seeker,GPS)对建立吸湿性预测模型性能的影响.结果 MARS算法建立的预测模型性能最佳,预测能力最强,模型的校正集决定系数(R2c)为0.843,预测集决定系数(R2p)为0.808,校正集均方根误差(root mean square error of calibration,RMSEC)为0.391,预测集均方根误差(root mean square error of prediction,RMSEP)为0.472,平均相对预测误差为2.69％,小于5％.结论 MARS算法建立的吸湿性预测模型更适合桂枝茯苓胶囊的生产应用,该算法可嵌入在线控制系统,为生产过程的质量控制智能化提供技术支持.

目的 旨在通过近红外光谱(NIRS)信息与腰痹通胶囊(Yaobitong Capsules,YC)中间体中值粒径(D50)的关系分析,探讨对该品种生产过程中4种中间体建立D50近红外通用定量模型的可行性.方法 采集YC生产过程中的原料细粉、干燥颗粒、整粒颗粒和总混颗粒4种中间体的NIRS,考察不同预处理方法对模型的影响,并采用间隔偏最小二乘法(iPLS)、组合间隔偏最小二乘法(siPLS)和移动窗口偏最小二乘法(mwPLS)优选NIRS波段,采用偏最小二乘法(PLS)对4种中间体建立1个D50通用定量模型.结果 通用模型的交叉验证均方根误差(RMSECV)为3.918 μm,预测均方根误差(RMSEP)为2.832μm,预测相对偏差(RSEP)为2.26％,小于5％,性能偏差比(RPD)为4.60,大于3,该模型可以用于定量测定,且与独立模型比,预测准确性相差不大.结论 NIRS通用定量模型可用于YC 4种中间体D50的测定.

目的 应用在线中红外光谱技术(MIR)对热毒宁注射液金银花Lonicerae Japonicae Flos、青蒿Artemisiae Annuae Herba(金青)醇沉过程中新绿原酸、绿原酸、隐绿原酸、异绿原酸A、异绿原酸B、异绿原酸C和断氧化马钱子苷的含量进行在线监控,提高该过程在线质量控制水平.方法 利用衰减全反射中红外光谱技术采集9个批次金青醇沉过程样本的光谱信息,组合间隔偏最小二乘法(synergy interval partial least squares,SiPLS)找到最佳波段,偏最小二乘法(PLS)建立该过程7种指标成分的定量模型,以决定系数(R2)、校正集误差均方根(RMSEC)、交互验证集误差均方根(RMSECV)、预测相对偏差(RSEP)为指标评价模型性能.结果 经模型优化,7种指标成分模型的R2值均大于0.96,RMSEC和RMSECV值均小于0.2 mg/mL,RSEP值和相对误差均小于10％.结论 在线中红外光谱分析技术可用于定量测定金青醇沉过程7种指标成分的含量,结果准确可靠.

Objective: To establish a method for determination of 6gingerol, 8gingerol and 10gingerol in dried ginger, processed ginger and ginger. Methods: Gingerol, 6-gingerol, 10-gingerol, 6-gingerol and 8-gingerol were determined by RP-HPLC. Results: the contents of 6-gingerol, 8-gingerol and 10-gingerol were 7.06%, 1.88% and 2.52% respectively. The linear correlation coefficient r of the standard curve was 0.9997, and the RSD of peak area was 0.43% after 5 consecutive injections; The maximum relative deviation of three detection results of the same sample is 4.7%. Conclusion: RP-HPLC method for the determination of gingerol in dried ginger, ginger and processed ginger is simple and accurate.

To control the risks of powder caking and capsule shell embrittlement of Guizhi Fuling Capsules, a predictive model for hygroscopicity of contents in Guizhi Fuling Capsules was built. A total of 90 batches of samples, including raw materials, intermediate powders and capsules, were collected during the manufacturing of Guizhi Fuling Capsules. According to the production sequence, 47 batches were used as the calibration set, and the properties of raw materials and the four intermediate powders were comprehensively characterized by the physical fingerprint. Then, the partial least squares(PLS) model was developed with the content hygroscopicity as the response variable. The variable importance in projection(VIP), variance inflation factor(VIF) and regression coefficients were used to screen out potential critical material attributes(pCMAs). As a result, five pCMAs from 54 physical parameters were screened out. Furthermore, different models were built by different combinations of pCMAs, and their predictive robustness of 43 batches was evaluated on the basis of the validation set. Finally, the tap density(D_c) of wet granules obtained from wet granulation and the angle of repose(α) of raw materials were identified as the critical material attributes(CMAs) affecting the hygroscopicity of the contents of Guizhi Fuling Capsules. The prediction model established with the two CMAs as independent variables had an average relative prediction error of 2.68% for samples in the validation set, indicating a good accuracy of prediction. This paper proved the feasibility of predictive modeling toward the control of critical quality attributes of Chinese medicine oral solid dosage(OSD). The combination of the continuous quality improvement, the industrial big data and the process modeling technique paved the way for the intelligent manufacturing of Chinese medicine oral solid preparations.

Lonicerae Japonicae Flos and Artemisiae Annuae Herba(LA or Jinqing) alcohol precipitation has various process parameters and complex process mechanism, and is one of the key units for manufacturing Reduning Injection. In order to identify the critical process parameters(CPPs) affecting the weight of the extract produced from the alcohol precipitation process, 259 batches of historical production data from 2017 to 2018 were collected, with a total of 829 318 data points. These data showed characteristics of large data, such as a large data volume, a low value density, and diverse sources. The data cleaning and feature extraction were first performed, and 48 feature variables were selected. The original data points were reduced to 9 936. Then, a combination of Pearson correlation analysis and grey correlation analysis were used to screen out 15 potential critical process parameters(pCPPs). After that, the partial least squares(PLS) was used in prediction of the weight of the extract, proving that the performance of predictive model based on 15 pCMAs is equivalent to that of predictive model based on 48 feature variables. The variable importance in projection(VIP) index was used to identify 9 CPPs, including 2 alcohol precipitation supernatant volume parameters, 4 initial extract weight parameters and 3 added alcohol volume parameters. As a result, the number of data points was 1 863, accounting for 0.28% of the original data. The big data analysis approach from a holistic point of view can effectively increase the value density of the original data. The critical process parameters obtained can help to accurately describe the quality transfer mechanism of the Jinqing alcohol precipitation process.

目的:探讨近红外光谱分析技术应用于检测天舒片包衣薄膜厚度的可行性.方法:采集9批天舒片包衣过程样品的近红外漫反射光谱,运用Kennard-Stone算法将样本集划分为校正集和验证集.优选预处理方法,并采用组合间隔偏最小二乘法(siPLS)和移动窗口偏最小二乘法(mwPLS)优选光谱区间,建立测定包衣厚度的偏最小二乘法(PLS)定量模型.结果:在标准正态变量变换+一阶导数+ Norris Derivative平滑对光谱进行预处理并结合siPLS的优选区间建立的模型中,校正集预测值与实测值的相关系数0.966,验证集预测值与实测值的相关系数0.991,表明预测值与实测值的相关性较好.校正均方根误差(RMSEC)0.198％,预测均方根误差(RMSEP)0.062％,表明模型的预测性能良好.结论:近红外光谱分析技术用于天舒片包衣薄膜厚度的测定具有很高的准确性,能够为中药片剂生产过程中包衣厚度的在线检测提供技术支持.

In this paper, a real time release testing(RTRT) model for predicting the disintegration time of Tianshu tablets was established on the basis of the concept of quality by design(QbD), in order to improve the quality controllability of the production process. First, 49 batches of raw materials and intermediates were collected. Afterwards, the physical quality attributes of all materials were comprehensively characterized. The partial least square(PLS) regression model was established with the 72 physical quality attributes of raw materials and intermediates as input and the disintegration time(DT) of uncoated tablets as output. Then, the variable screening was carried out based on the variable importance in the projection(VIP) indexes. Moisture content of raw materials(%HR), tapped density of wet masses(D_c), hygroscopicity of dry granules(%H), moisture content of milling granules(%HR) and Carr's index of mixed granules(IC) were determined as the potential critical material attributes(pCMAs). According to the effects of interactions of pCMAs on the performance of the prediction model, it was finally determined that the wet masses' D_c and the dry granules'%H were critical material attributes(CMAs). A RTRT model of the disintegration time prediction was established as DT=34.09+2×D_c+3.59×%H-5.29×%H×D_c,with R~2 equaling to 0.901 7 and the adjusted R~2 equaling to 0.893 3. The average relative prediction error of validation set for the RTRT model was 3.69%. The control limits of the CMAs were determined as 0.55 g·cm~(-3)<D_c<0.63 g·cm~(-3) and 4.77<%H<7.59 according to the design space. The RTRT model of the disintegration time reflects the understanding of the process system, and lays a foundation for the implementation of intelligent control strategy of the key process of Tianshu Tablets.