Input Convex Neural Network Modeling for Rheological Performance Evaluation of Non-Newtonian Fluid Residues in Chemical Reactors

ZHANG Dongkuan; CHU Zhiqiang; TIAN Hong; Jeffrey S.CROSS; JI Guozhao

doi:10.7652/xjtuxb202512007

您当前的位置：

首页 >

文章列表页 >

Input Convex Neural Network Modeling for Rheological Performance Evaluation of Non-Newtonian Fluid Residues in Chemical Reactors

Special Topic Pollution Control and Resource Utilization in the“Dual Carbon”Context | 更新时间：2025-12-08

- Input Convex Neural Network Modeling for Rheological Performance Evaluation of Non-Newtonian Fluid Residues in Chemical Reactors
- Journal of Xi'an Jiaotong University Vol. 59, Issue 12, Pages: 80-91(2025)
- 作者机构：
- 作者简介：
- 基金信息：
- DOI：10.7652/xjtuxb202512007
  CLC： TP30
- Received：25 May 2025，
  
  Published：10 December 2025
- 稿件说明：
移动端阅览
ZHANG Dongkuan, CHU Zhiqiang, TIAN Hong, et al. Input Convex Neural Network Modeling for Rheological Performance Evaluation of Non-Newtonian Fluid Residues in Chemical Reactors[J]. Journal of Xi'an Jiaotong University, 2025, 59(12): 80-91.
DOI：

ZHANG Dongkuan, CHU Zhiqiang, TIAN Hong, et al. Input Convex Neural Network Modeling for Rheological Performance Evaluation of Non-Newtonian Fluid Residues in Chemical Reactors[J]. Journal of Xi'an Jiaotong University, 2025, 59(12): 80-91. DOI： 10.7652/xjtuxb202512007.

摘要

针对化工釜残非牛顿流体流变特性建模精度低、异常数据敏感性强的问题，提出了一种输入凸神经网络建模方法，以提升复杂非牛顿流体的黏度预测稳定性与物理一致性。首先，构建具有凸性约束的3层神经网络结构，引入权重非负限制，以确保模型输出的物理合理性；其次，设计指数映射正则项，增强对高黏度区间及异常点的鲁棒性；最后，基于5种典型釜残样本，在0～2500 s

-1

剪切速率范围内采集黏度响应数据，对输入凸神经网络与幂律模型、Carreau模型、线性回归、支持向量回归、决策树、随机森林和前馈神经网络进行拟合性能评估。研究结果表明：输入凸神经网络在多模型对比中展现出更优的拟合效果与预测能力，均方误差低至2.06×10

，拟合优度可达0.987，整体性能在准确性与稳健性之间取得良好平衡；输入凸神经网络的预测曲线更为平滑，能够避免传统模型出现的震荡与失真。该方法为化工釜残等复杂非牛顿流体的流变建模提供了有效技术路径，对输送控制与处理工艺优化具有重要工程参考价值。

Abstract

To address the problems of low accuracy and high sensitivity to outliers for rheological performance modeling of non-Newtonian fluid residues in chemical reactors

an Input convex neural network（ICNN）modeling method is proposed for higher viscosity prediction stability and better physical consistency of complex non-Newtonian fluids.For the purpose of this study

a three-layer ICNN model with convexity constraints was first constructed

and non-negative weight restrictions were introduced to secure physical reasonability of the model；then an exponential mapping regularization item was designed to en

hance robustness to high-viscosity regimes and outliers；finally

a fitting performance evaluation of the input convex neural network was conducted with viscosity response data of 5 typical residue samples collected from chemical reactors at a shear rate of 0—2500 s

-1

using the power-law model

Carreau model

linear regression

support vector regression

decision tree

random forest

feedforward neural network models as references for comparison.The results show that among the models

the ICNN model demonstrates better fitting effect and more accurate prediction

with a low mean square error of 2.06×10

and a fitting performance metric of 0.987

having achieved a good balance between accuracy and robustness；besides

the prediction curve of the ICNN model is smoother

having avoided oscillation and distortion that are common with classical models.The model proposed in this study provides an effective method for rheological performance modeling of complex non-Newtonian fluids such as fluid residues in chemical reactors

and it is of significant engineering reference value for the optimization of transport control and processing technologies.

关键词

Keywords

references

蔡思超 , 周静 , 杜金泽 , 等 . 煤化工酚基精馏釜残资源化利用过程初步分析 [J ] . 化工进展 , 2022 , 41 ( 6 ): 3360 - 3371 .

CAI Sichao , ZHOU Jing , DU Jinze , et al . Process analysis of resource utilization of phenol-based distillation residue from coal chemical industry [J ] . Chemical Industry and Engineering Progress , 2022 , 41 ( 6 ): 3360 - 3371 .

王娇 , 刘吉宝 , 宋云鹏 , 等 . 高含固餐厨垃圾有机组成对其流变特征的影响 [J ] . 环境科学学报 , 2021 , 41 ( 10 ): 4107 - 4115 .

WANG Jiao , LIU Jibao , SONG Yunpeng , et al . Effect of organic composition of high solid food waste on its rheological behaviors [J ] . Acta Scientiae Circumstantiae , 2021 , 41 ( 10 ): 4107 - 4115 .

马媛 , 杜金泽 , 周静 , 等 . 酚基釜残深拔残渣分析及其热转化制备CO 2 吸附材料 [J ] . 化学工业与工程 , 2021 , 38 ( 1 ): 43 - 52 .

MA Yuan , DU Jinze , ZHOU Jing , et al . Chemical analysis of the phenolic tar residue and its potential conversion to CO 2 adsorbents by thermal treatment [J ] . Chemical Industry and Engineering , 2021 , 38 ( 1 ): 43 - 52 .

徐硕涵 , 赵凯 , 白天阳 , 等 . 喷嘴内流影响下的喷油器性能参数敏感性分析 [J ] . 西安交通大学学报 , 2025 , 59 ( 5 ): 189 - 197 .

XU Shuohan , ZHAO Kai , BAI Tianyang , et al . Parameters sensitivity analysis of injector performance with the effect of nozzle internal flows [J ] . Journal of Xi’an Jiaotong University , 2025 , 59 ( 5 ): 189 - 197 .

陈芝 , 刘建忠 , 陈建 , 等 . DMF精馏残渣与煤共成浆特性 [J ] . 动力工程学报 , 2021 , 41 ( 1 ): 71 - 78 .

CHEN Zhi , LIU Jianzhong , CHEN Jian , et al . Coslurry ability of DMF distillation residue with coal [J ] . Journal of Chinese Society of Power Engineering , 2021 , 41 ( 1 ): 71 - 78 .

周储朋 , 钱善华 , 任海栋 , 等 . 基于流体仿真的搅拌器结构参数对搅拌特性影响研究 [J ] . 化学工业与工程 , 2024 , 41 ( 3 ): 124 - 134 .

ZHOU Chupeng , QIAN Shanhua , REN Haidong , et al . Research on the influence of agitator structural parameters on stirring characteristics based on fluid simulation [J ] . Chemical Industry and Engineering , 2024 , 41 ( 3 ): 124 - 134 .

徐家通 , 周子愉 , 廖晴雨 , 等 . 复杂流体屈服应力的测定：理论、方法及应用 [J ] . 高分子通报 , 2025 , 38 ( 5 ): 689 - 717 .

XU Jiatong , ZHOU Ziyu , LIAO Baoliang , et al . Determination of yield stress in complex fluids:theory, methods and applications [J ] . Polymer Bulletin , 2025 , 38 ( 5 ): 689 - 717 .

刘泓志 , 徐善坤 , 郭易东 , 等 . 泥浆流变模型对泥水盾构排浆管道压力损失特性的影响分析 [J ] . 现代隧道技术 , 2024 , 61 ( 1 ): 182 - 189 .

LIU Hongzhi , XU Shankun , GUO Yidong , et al . Analysis of the effect of slurry rheological model on the pressure loss characteristics of slurry discharge pipe of slurry shield [J ] . Modern Tunnelling Technology , 2024 , 61 ( 1 ): 182 - 189 .

何文卓 , 刘晓玲 , 朱鹏娟 , 等 . 基于Carreau流变模型的圆柱滚子轴承热混合润滑分析 [J ] . 润滑与密封 , 2023 , 48 ( 5 ): 52 - 59 .

HE Wenzhuo , LIU Xiaoling , ZHU Pengjuan , et al . Analysis of thermal mixed lubrication of cylindrical roller bearings based on Carreau rheological model [J ] . Lubrication Engineering , 2023 , 48 ( 5 ): 52 - 59 .

DWIER S , GARROUCH A , LABABIDI H . Delineation of the effective viscosity controls of diluted polymer solutions at various flow regimes [J ] . Fluid Phase Equilibria , 2024 , 584 : 114143 .

ISMAIL A , RASHID H M A , GHOLAMI R , et al . Characterization based machine learning modeling for the prediction of the rheological properties of waterbased drilling mud:an experimental study on grass as an environmental friendly additive [J ] . Journal of Petroleum Exploration and Production Technology , 2022 , 12 ( 6 ): 1677 - 1695 .

BEMANI A , MADANI M , KAZEMI A . Machine learning-based estimation of nano-lubricants viscosity in different operating conditions [J ] . Fuel , 2023 , 352 : 129102 .

PAROLINI N , POIATTI A , VENÉ J , et al . Structure-preserving neural networks in data-driven rheological models [J ] . SIAM Journal on Scientific Computing , 2025 , 47 ( 1 ): C182 - C206 .

TAÇ V , LINKA K , SAHLI-COSTABAL F , et al . Benchmarking physics-informed frameworks for datadriven hyperelasticity [J ] . Computational Mechanics , 2024 , 73 ( 1 ): 49 - 65 .

YAN Tongtong , WANG Dong , XIA Tangbin , et al . New shapeness property and its convex optimization model for interpretable machine degradation modeling [J ] . IEEE Transactions on Reliability , 2023 , 72 ( 2 ): 703 - 715 .

张争奇 , 雷志坤 , 张天天 , 等 . 高黏改性沥青制备及高温流变性能 [J ] . 材料科学与工程学报 , 2025 , 43 ( 1 ): 47 - 54 .

ZHANG Zhengqi , LEI Zhikun , ZHANG Tiantian , et al . Preparation and high temperature rheological properties of high viscosity asphalt [J ] . Journal of Materials Science and Engineering , 2025 , 43 ( 1 ): 47 - 54 .

程曦 , 张志勇 . 基于人工神经网络的复杂介质中波的传播不确定性分析方法 [J ] . 电子与信息学报 , 2021 , 43 ( 12 ): 3662 - 3670 .

CHENG Xi , ZHANG Zhiyong . An uncertainty analysis method of wave propagation in complex media based on artificial neural network [J ] . Journal of Electronics& Information Technology , 2021 , 43 ( 12 ): 3662 - 3670 .

TAHERI M , KLIDBARY S H . Generating a large family of nonlinear activation functions (LFNAFs)in neural networks [J ] . The Journal of Supercomputing , 2025 , 81 ( 5 ): 661 .

GUJARATI K R , BATHALA L , VENKATESH V , et al . Transformer-based automated segmentation of the median nerve in ultrasound videos of wrist-to-elbow region [J ] . IEEE Transactions on Ultrasonics Ferroelectrics and Frequency Control , 2024 , 71 ( 1 ): 56 - 69 .

范昌浩 , 李明佳 , 李梦杰 , 等 . 基于可解释机器学习的固体填充床储热装置快速设计研究 [J ] . 西安交通大学学报 , 2024 , 58 ( 11 ): 87 - 97 .

FAN Changhao , LI Mingjia , LI Mengjie , et al . Research on rapid design of solid packed-bed heat storage devices based on interpretable machine learning [J ] . Journal of Xi’an Jiaotong University , 2024 , 58 ( 11 ): 87 - 97 .

徐岸 , 吴永明 , 郑洋 . 基于自监督与蒸馏约束的正则化类增量学习方法 [J ] . 计算机辅助设计与图形学学报 , 2024 , 36 ( 5 ): 775 - 785 .

XU An , WU Yongming , ZHENG Yang . A regularized class incremental learning method based on selfsupervision with distillation constraints [J ] . Journal of Computer-Aided Design & Computer Graphics , 2024 , 36 ( 5 ): 775 - 785 .

李锋 , 彭天波 . 包含Mullins效应的HDR热-超-黏弹性本构模型 [J ] . 力学学报 , 2024 , 56 ( 12 ): 3498 - 3506 .

LI Feng , PENG Tianbo . An HDR thermal-hyper-viscoelastic constitutive model with Mullins effect [J ] . Chinese Journal of Theoretical and Applied Mechanics , 2024 , 56 ( 12 ): 3498 - 3506 .

张卓伦 , 袁帅鹏 , 李铁克 , 等 . 基于两级决策树模型的轧制时间预测方法 [J ] . 计算机集成制造系统 , 2025 , 31 ( 1 ): 197 - 210 .

ZHANG Zhuolun , YUAN Shuaipeng , LI Tieke , et al . Steel rolling time prediction method based on two-level decision tree model [J ] . Computer Integrated Manufacturing Systems , 2025 , 31 ( 1 ): 197 - 210 .

高旺 , 刘裕荐 , 郭雅娟 , 等 . 基于随机森林的GNSS观测粗差拟准检定方法 [J ] . 仪器仪表学报 , 2025 , 46 ( 1 ): 42 - 53 .

GAO Wang , LIU Yujian , GUO Yajuan , et al . Quasiaccurate detection method for GNSS observation gross errors based on random forest [J ] . Chinese Journal of Scientific Instrument , 2025 , 46 ( 1 ): 42 - 53 .

李晓颖 , 单娴 , 张哲硕 , 等 . 基于G-Loss的在线支持向量回归算法 [J ] . 系统科学与数学 , 2025 , 45 ( 2 ): 587 - 602 .

LI Xiaoying , SHAN Xian , ZHANG Zheshuo , et al . Online support vector regression algorithm based on G-Loss [J ] . Journal of Systems Science and Mathematical Sciences , 2025 , 45 ( 2 ): 587 - 602 .

CHEN Zhanfeng , LI Xuyao , WANG Wen , et al . Residual strength prediction of corroded pipelines using multilayer perceptron and modified feedforward neural network [J ] . Reliability Engineering & System Safety , 2023 , 231 : 108980 .

YANG Ao , SUN Shirui , QI Lu , et al . Development of an interpretable QSPR model to predict the octanolwater partition coefficient based on three artificial intelligence algorithms [J ] . Green Chemical Engineering , 2025 , 6 ( 2 ): 193 - 199 .

Views

下载量

CSCD

Alert me when the article has been cited

提交

Tools

Publicity Resources

Drag Reduction in Turbulent Pipe Flows of Aqueous Xanthan Gum Solutions

Centrifugal Blood Pump Model Design and Verification Test

Effects of Expansion Angle Crater Film Hole on the Film Cooling Characteristics of a Double-Wall Structure

Energy Efficiency Optimization of the Cryogenic Air Separation Process for Electronic-Grade Ultra Pure Nitrogen

Bi-Level Multi-Obj ective Optimization of a Wind-Solar Coupled Hydrogen Production System Based on the Synergy of Hybrid Electrolyzer Arrays

Related Author

张根广 1

张鸣远 1

杨万英 1

朱宪然 1

Related Institution

School of Energy and Power Engineering, Xi'an Jiaotong University

College of Water Resources and Architectural Engineering, Northwest Sci-Tech University of Agriculture and Forestry, Yangling

⁰