目的 建立一种基于声发射技术的养胃颗粒流化床制粒过程在线检测方法。方法 以10批养胃颗粒流化床制粒过程为研究对象，在制粒过程中在线采集声发射信号并取样，离线分析颗粒含水量和平均粒径（D₅₀）作为参考值，采用偏最小二乘回归法建立声发射信号同各项质量指标之间的校正模型。结果 颗粒含水量和D₅₀的预测决定系数（R_p²）分别为 0.859 6、0.834 4，预测误差均方根（RMSEP）分别为0.61%、40.6 μm。结论 基于声发射技术的分析方法可用于对养胃颗粒流化床制粒过程中多指标的在线检测，为该工艺过程的实时反馈控制提供技术支持。

Objective To establish an on-line detection method for fluid bed granulation process of Yangwei Granules (养胃颗粒) based on acoustic emission technology. Methods Fluid bed granulation processes of 10 batches of Yangwei Granules were operated for study. Acoustic emission signals and samples were collected on-line during granulation processes and reference values of moisture and mean granule size (D₅₀) were obtained off-line. Partial least-square regression (PLSR) method was used to establish calibration models between acoustic emission signals and indicators. Results The coefficient of determination (R_p²) for moisture and D₅₀ of validation set were 0.859 6, 0.834 4 and root mean square error of prediction (RMSEP) were 0.61%, 40.6 μm, respectively. Conclusion The method using acoustic emission technology can be applied for on-line monitoring of quality indicators in fluid bed granulation process for Yangwei Granules, which provides technical support for the real-time feedback control of this process.

R283.6

国家科技重大专项“重大新药创制”（2018ZX09201011-002）