论文部分内容阅读
目的:比较不同模型拟合制粒过程参数对肿节风颗粒成型的影响。方法:采用湿法制粒制备肿节风颗粒,研究了物料含水量、润湿剂浓度、物料液固比对颗粒得率和颗粒成型性的影响。采用响应曲面模型和神经网络模型拟合实验数据,验证两种模型的优劣并采用响应曲面模型筛选了最优制粒工艺参数。结果:响应曲面模型与神经网络模型均能较好的拟合实验数据,其中神经网络模型具有较高的准确性和预测能力;制粒过程参数对肿节风颗粒得率与颗粒成型率具有较大的影响,其中物料含水量是关键制粒工艺参数;最优制粒工艺参数为:物料含水量为2.04%,润湿剂浓度为53.48%,物料液固比为17.20%,颗粒得率为(98.18±0.32)%,颗粒成型率为(99.30±0.40)%。结论:应用响应曲面模型和神经网络模型有助于对制粒过程和机制的理解,并能够对制粒工艺参数的优化提供指导。
OBJECTIVE: To compare the effects of different model fitting granulation parameters on the forming of Aegilops leading to wind. Methods: The wet granulation was used to prepare swollenose granules. The effects of water content, concentration of wetting agent and liquid to solid ratio on the yield and granule formation of granule were studied. The response surface model and neural network model were used to fit the experimental data. The advantages and disadvantages of the two models were verified and the optimum granulation process parameters were screened by response surface model. Results: Both the response surface model and the neural network model fit the experimental data well, and the neural network model has higher accuracy and predictive ability. The parameters of the granulation process have a better effect on the particle yield and the rate of particle formation The moisture content of the material is the key granulation process parameters; the optimal granulation process parameters are as follows: the moisture content of material is 2.04%, the concentration of wetting agent is 53.48%, the material liquid-solid ratio is 17.20% (98.18 ± 0.32)%, the rate of particle formation was (99.30 ± 0.40)%. CONCLUSIONS: The application of response surface models and neural network models is helpful in understanding the granulation process and mechanism and can provide guidance on the optimization of the granulation process parameters.