目的 优化白芍的清炒炮制工艺，确定最佳工艺参数，为中药饮片的标准化生产提供依据。方法 采用HPLC法测定炒白芍中没食子酸、儿茶素、芍药内酯苷、芍药苷、1,2,3,4,6-五没食子酰葡萄糖（1,2,3,4,6-pentagalloylglucose，PGG）和苯甲酰芍药苷的含量。以这6种成分及水溶性浸出物含量的总评归一化值（OD）作为综合评价指标。在单因素考察投药量、炒制温度和炒制时间的基础上，结合CRITIC权重赋权法和Box-Benhnken设计-响应面法（Box-Benhnken design-response surface methodology，BBD-RSM）进行炒白芍炮制工艺的优化试验。最后，使用MATLAB R2020b深度学习工具箱构建人工神经网络，对BBD-RSM得出的最优工艺参数进行验证，并分析投药量、炒制温度、炒制时间与OD值之间的关系。结果 BBD-RSM结合人工神经网络优化试验得出最佳炮制工艺参数为白芍饮片投药量46 g/L，炒制温度143 ℃，炒制时间8.4 min，预测综合评分OD值为0.612 2。在此最优条件下进行3次重复验证，测得没食子酸、儿茶素、芍药内酯苷、芍药苷、PGG、苯甲酰芍药苷的平均质量分数分别为1.78、0.56、6.70、26.95、4.02、0.51 mg/g，水溶性浸出物的平均含量为31.62%。综合评分OD值分别为0.640 9、0.628 4、0.596 2，平均值为0.621 8，RSD为3.71%（＜5%），与预测值接近。这表明建立的炒白芍炮制工艺稳定可行。结论 优化的炒白芍工艺稳定可行，可为炒白芍的质量控制提供参考。

Objective To optimize the stir-frying process for Baishao (Paeoniae Radix Alba, PRA) and determine the optimal process parameters, thereby providing a scientific basis for the standardized production of its processed decoction pieces. Methods HPLC was employed to determine the contents of gallic acid, catechin, albiflorin, paeoniflorin, 1,2,3,4,6-pentagalloylglucose (PGG), and benzoylpaeoniflorin in stir-fried PRA. A normalized comprehensive evaluation value (OD value), derived from the contents of these six components and water-soluble extractives, was established as the overall evaluation index. Based on preliminary single-factor investigations of administration dosage, stir-frying temperature, and stir-frying time, the optimization experiment was conducted using the CRITIC weighting method in combination with Box-Behnken design-response surface methodology (BBD-RSM). Subsequently, an artificial neural network was constructed using the deep learning toolbox in MATLAB R2020b to validate the optimal parameters obtained from the BBD-RSM and to analyze the relationships between the dosage of administration, stir-frying temperature, stir-frying time and OD value. Results The optimal processing parameters obtained from the BBD-RSM combined with artificial neural network optimization experiment were determined to be a administration dosage of 46 g/L, a stir-frying temperature of 143 ℃, and a stir-frying time of 8.4 min. The predicted comprehensive score under these conditions was 0.612 2. Triplicate verification experiments conducted at these optimal parameters yielded average mass fractions of gallic acid, catechin, albiflorin, paeoniflorin, PGG, and benzoylpaeoniflorin of 1.78, 0.56, 6.70, 26.95, 4.02, and 0.51 mg/g, respectively, with an average content of water-soluble extractives of 31.62%. The comprehensive OD scores were 0.640 9, 0.628 4, and 0.596 2, giving an average value of 0.621 8 and a relative standard deviation (RSD) of 3.71% (< 5%), which was close to the predicted value. These results demonstrate that the established stir-frying process of PRA is stable and feasible. Conclusion The optimized stir-frying process for PRA is stable and feasible. This study provides a valuable reference for the quality control of stir-fried PRA.

R283.6

河北省重点研发计划项目（23372503D）；国家中医药管理局科研项目（gzy-kjs-2023-029）；国家中医药管理局科研项目（gzy-kjs-2025-015）；河北省省级科技计划项目（252W2501D）