Interpretable prediction of polymer material viscosity using machine learning models incorporating shear rate, salinity, concentration, and temperature effects

Abstract Viscosity is a fundamental property of polymer materials and solutions that strongly influences their flow behavior, processing, and performance in various industrial and engineering applications. Polymer materials and polymer solutions are vital components in many industrial and engineering processes, in which viscosity is an important characteristic determining their flow behavior and performance. This paper presents predictive modeling of logarithmic viscosity (LV) of polymer solutions using Artificial Neural Networks (ANN) and Adaptive Neuro-Fuzzy Inference System (ANFIS). 420 samples were collected through a public Kaggle database. Logarithm Shear Rate (LS), Polymer Concentration (PC), NaCl Concentration (NA) and Temperature (TE) are considered as input variables. The collected dataset is randomly divided into training (70%), validation (15%) and test set (15%). From the results, the error histogram revealed that the majority of the prediction errors are centred around −21.35, while the ANN model performed optimally during validation with MSE of 67,522.49 at epoch 20. A representative ANFIS prediction yielded an LV value of 11.7 cP for an input combination of LS of 55.8 s −1 , PC of 0.175 wt%, NA of 2.05 wt%, and TE of 57.5 °C. ANFIS was found to perform exceptionally well compared to ANN, based on an RMSE of 31.50, an MAE of 6.49, and an R2 value of 0.9846. In addition, the logarithm shear rate of 0.69, polymer concentration of 0.21, and NaCl concentration of 0.11 are established as highly influential factors based on SHAP analysis. Logarithmic viscosity increases drastically with an increase in polymer concentration. Overall, the findings confirm that the ANFIS is an effective method to predict the viscosity of polymer solutions through reliable, accurate, and comprehensible means. Possible applications of this technique include polymer processing, formulation, chemical processing, enhanced oil recovery, and flow control systems in industry.

Authors

Publication Details

Journal
Journal of Materials Science Materials in Engineering
Published
2026-10-11
DOI
https://doi.org/10.1186/s40712-026-00609-4
Primary Topic
Rheology and Fluid Dynamics Studies
Type
article
Field-Weighted Citation Impact
0.00
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
article

Interpretable prediction of polymer material viscosity using machine learning models incorporating shear rate, salinity, concentration, and temperature effects

Vamsi Krishna Kudapa, Vijayakumar Sivasundar, Adina Srinivasa Vara Prasad, Amaleswari Rajulapati et al.
Journal of Materials Science Materials in Engineering
Rheology and Fluid Dynamics Studies
article

Interpretable prediction of polymer material viscosity using machine learning models incorporating shear rate, salinity, concentration, and temperature effects

Vamsi Krishna Kudapa, Vijayakumar Sivasundar, Adina Srinivasa Vara Prasad, Amaleswari Rajulapati, Movva Naga swapna sri, Kibrom Hagos Gebrehiwot, Sandeep Kumar, N. Dhasarathan
article en

Abstract

Abstract Viscosity is a fundamental property of polymer materials and solutions that strongly influences their flow behavior, processing, and performance in various industrial and engineering applications. Polymer materials and polymer solutions are vital components in many industrial and engineering processes, in which viscosity is an important characteristic determining their flow behavior and performance. This paper presents predictive modeling of logarithmic viscosity (LV) of polymer solutions using Artificial Neural Networks (ANN) and Adaptive Neuro-Fuzzy Inference System (ANFIS). 420 samples were collected through a public Kaggle database. Logarithm Shear Rate (LS), Polymer Concentration (PC), NaCl Concentration (NA) and Temperature (TE) are considered as input variables. The collected dataset is randomly divided into training (70%), validation (15%) and test set (15%). From the results, the error histogram revealed that the majority of the prediction errors are centred around −21.35, while the ANN model performed optimally during validation with MSE of 67,522.49 at epoch 20. A representative ANFIS prediction yielded an LV value of 11.7 cP for an input combination of LS of 55.8 s −1 , PC of 0.175 wt%, NA of 2.05 wt%, and TE of 57.5 °C. ANFIS was found to perform exceptionally well compared to ANN, based on an RMSE of 31.50, an MAE of 6.49, and an R2 value of 0.9846. In addition, the logarithm shear rate of 0.69, polymer concentration of 0.21, and NaCl concentration of 0.11 are established as highly influential factors based on SHAP analysis. Logarithmic viscosity increases drastically with an increase in polymer concentration. Overall, the findings confirm that the ANFIS is an effective method to predict the viscosity of polymer solutions through reliable, accurate, and comprehensible means. Possible applications of this technique include polymer processing, formulation, chemical processing, enhanced oil recovery, and flow control systems in industry.

Journal of Materials Science Materials in Engineering
Openalex Percentile: Top 22%
Rheology and Fluid Dynamics Studies
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.