机器学习预测叶片-网孔管线式高剪切混合器性能

摘要高剪切混合器作为一种新型的过程强化设备，工业应用日益广泛，但其工程设计依然依靠经验放大。利用不同定转子构型的叶片-网孔管线式高剪切混合器的功耗、液-液传质系数和乳化性能等数据，采用反向传播神经网络算法、循环神经网络算法和决策树算法等机器学习算法对数据进行分析建模，为高剪切混合器的设计与优化提供工具。结果表明：反向传播神经网络算法和循环神经网络算法都可以准确预测高剪切混合器性能，但是单个神经网络算法存在过拟合和泛化能力差的问题，通过将不同机器学习模型融合进一步提高了模型精度和稳定性。基于自动机器学习的PyCaret程序能够准确拟合数据，但在数据量较小的情况下，其优化能力较差。

	Service

	把本文推荐给朋友
	加入我的书架
	加入引用管理器
	Email Alert
	RSS
	作者相关文章
	王灵杰
	郭俊恒
	李文鹏
	程芹
	张金利

关键词：高剪切混合器机器学习神经网络功耗液-液传质乳化

Abstract： As a novel type of process intensification equipment, high-shear mixers were increasingly widely used in industry, but their design still relies on experimental scaling-up. Data on power, liquid-liquid mass transfer coefficient and emulsification of in-line high-shear mixers with different stator and rotor configurations were collected. And regression fitting analysis on the collected data was performed using machine learning algorithms such as back propagation neural network, recurrent neural network, decision tree, etc., which provided information for the design and optimization of high-shear mixers. The results show that back propagation neural network and recurrent neural network algorithms can predict high-shear mixer's performance precisely and the accuracy of the model can be further improved by fusing different machine learning models. The PyCaret program based on auto machine learning can accurately fit the data. However, its optimization performance is poor when the amount of data is small.

Keywords： high shear mixers； machine learning； neural networks； power； liquid-liquid mass transfer； emulsification

Received 2021-03-30;

Fund:国家自然科学基金项目（U20A20151；21776179）。

Corresponding Authors: 张金利,教授,E-mail:zhangjinli@tju.edu.cn。 Email: zhangjinli@tju.edu.cn

About author: 王灵杰(1988-),男,硕士研究生,主要从事机器学习在高剪切方向的研究。

引用本文:

王灵杰, 郭俊恒, 李文鹏, 程芹, 张金利.机器学习预测叶片-网孔管线式高剪切混合器性能[J]. 化学工业与工程, 2022,39(2): 1-8

WANG Lingjie, GUO Junheng, LI Wenpeng, CHENG Qin, ZHANG Jinli.Forecast of in-line blade-screen high shear mixer's performance based on machine learning[J]. Chemcial Industry and Engineering, 2022,39(2): 1-8

[1] ZHANG J, XU S, LI W. High shear mixers:A review of typical applications and studies on power draw, flow pattern, energy dissipation and transfer properties[J]. Chemical Engineering and Processing:Process Intensification, 2012, 57/58:25-41
[2] QIN H, ZHANG C, XU Q, et al. Geometrical improvement of inline high shear mixers to intensify micromixing performance[J]. Chemical Engineering Journal, 2017, 319:307-320
[3] ZHOU J, LI Y, LI D, et al. Online learning based intelligent temperature control during polymer composites microwave curing process[J]. Chemical Engineering Journal, 2019, 370:455-465
[4] WU S, KONDO Y, KAKIMOTO M A, et al. Machine-learning-assisted discovery of polymers with high thermal conductivity using a molecular design algorithm[J]. Npj Computational Materials, 2019, doi:10.1038/s41524-019-0203-2
[5] 孙永利, 王华金, 郝丽, 等. 基于神经网络和遗传算法的螺旋折流板换热器性能预测[J]. 化学工业与工程, 2016, 33(4):49-55 SUN Yongli, WANG Huajin, HAO Li, et al. Performance prediction of shell-and-tube heat exchangers with helical baffles using multilayer perception neural networks optimized with genetic algorithm[J]. Chemical Industry and Engineering, 2016, 33(4):49-55(in Chinese)
[6] LI W, XIA F, ZHAO S, et al. Mixing performance of an inline high-shear mixer with a novel pore-array liquid distributor[J]. Industrial & Engineering Chemistry Research, 2019, 58(44):20213-20225
[7] REN C, AN N, WANG J, et al. Optimal parameters selection for BP neural network based on particle swarm optimization:A case study of wind speed forecasting[J]. Knowledge-Based Systems, 2014, 56:226-239
[8] SONG J, GUO Y, GAO L, et al. From deterministic to generative:Multimodal stochastic RNNs for video captioning[J]. IEEE Transactions on Neural Networks and Learning Systems, 2019, 30(10):3047-3058
[9] YUAN X, LI L, WANG Y. Nonlinear dynamic soft sensor modeling with supervised long short-term memory network[J]. IEEE Transactions on Industrial Informatics, 2020, 16(5):3168-3176
[10] CHE Z, PURUSHOTHAM S, CHO K, et al. Recurrent neural networks for multivariate time series with missing values[J]. Scientific Reports, 2018, doi:10.1038/s41598-018-24271-9
[11] YU Z, HAGHIGHAT F, FUNG B C M, et al. A decision tree method for building energy demand modeling[J]. Energy and Buildings, 2010, 42(10):1637-1646
[12] GEURTS P, ERNST D, WEHENKEL L. Extremely randomized trees[J]. Machine Learning, 2006, 63(1):3-42
[13] CHEN T, GUESTRIN C. XGBoost:A scalable tree boosting system[C]//Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining. New York, USA:ACM, 2016
[14] FRIEDMAN J H. Stochastic gradient boosting[J]. Computational Statistics & Data Analysis, 2002, 38(4):367-378
[15] PyCaret.org. PyCaret[EB/OL]. https://pycaret.org/about. PyCaret version 1.0.0,2020-04-30
[16] CHENG Q, XU S, SHI J, et al. Pump capacity and power consumption of two commercial in-line high shear mixers[J]. Industrial & Engineering Chemistry Research, 2013, 52(1):525-537
[17] 秦宏云. 管线型高剪切混合器的几何构型优化[D]. 天津:天津大学, 2018