Abstract In this work, design parameters of carrier-based dry powder inhalation were studied using surrogate modeling technique. The surrogate models constructed were then used to evaluate the key design parameters independently, which were otherwise difficult to determine based on experimental studies alone. Artificial neural network (ANN) was chosen as the surrogate modeling technique and models were constructed based on experimental data obtained from the literature. Twenty-eight variables describing the carrier size distribution, density, surface characteristics and operating conditions of dry powder inhaler were used as the input variables and emitted dose (ED) and fine particle fraction (FPF) were used as the output variables. Carrier surface characteristics were evaluated by applying image analysis on carrier SEM images. Genetic algorithm (GA) was used for the selection of important variables to be included in the surrogate models. Sensitivity analysis was also performed to determine the key variables affecting ED and FPF. Key design criteria for carrier-based dry powder inhalation were proposed based on the surrogate models constructed.

中文翻译：

---

通过替代模型评估载体基干粉吸入中的载体尺寸和表面形态

摘要 在这项工作中，使用代理建模技术研究了基于载体的干粉吸入剂的设计参数。然后使用构建的替代模型独立评估关键设计参数，否则仅根据实验研究很难确定这些参数。选择人工神经网络（ANN）作为替代建模技术，并根据从文献中获得的实验数据构建模型。描述干粉吸入器的载体尺寸分布、密度、表面特征和操作条件的二十八个变量被用作输入变量，并且发射剂量（ED）和细颗粒分数（FPF）被用作输出变量。通过对载体 SEM 图像应用图像分析来评估载体表面特性。遗传算法 (GA) 用于选择要包含在替代模型中的重要变量。还进行了敏感性分析以确定影响 ED 和 FPF 的关键变量。基于构建的替代模型提出了基于载体的干粉吸入的关键设计标准。