AI Model Predicts Mass Transfer in L-Shaped Columns
Researchers developed a machine learning model that accurately calculates mass transfer coefficients in complex extraction columns.
Updated on Oct. 26, 2026 in Artificial Intelligence
Live Poll
Do you believe advanced machine learning models effectively increase industrial process reliability?
Scientists have created an artificial neural network to predict mass transfer coefficients in L-shaped pulsed sieve plate extraction columns. This model outperforms traditional regression techniques, providing a more reliable way to analyze non-standard chemical processing equipment.
Why it matters
Predicting mass transfer coefficients is notoriously difficult for non-standard column geometries, which are critical for various industrial processes. By enhancing predictive accuracy, this AI approach could optimize performance and efficiency in complex chemical engineering systems.
The artificial neural network model achieved a coefficient of determination above 0.98 with an average absolute relative error below 6%. The study evaluated these results based on 36 specific operating conditions tested per column section.
The details
Researchers trained artificial neural network, random forest, and support vector regression models using experimental data from an L-shaped pulsed sieve plate extraction column. Sensitivity analysis revealed that dispersed phase holdup and drop size are the most dominant variables for predicting mass transfer across both horizontal and vertical sections.
Timeline
The research findings were officially published on October 26, 2026.
The Tech Race
This study advances the computational fluid dynamics modeling framework by providing a more precise AI-driven method for predicting mass transfer coefficients in industrial equipment. It represents a significant shift from traditional manual calculations toward automated, high-precision predictive modeling.
The improvement in mass transfer predictions allows engineers to design more efficient chemical processing systems with lower energy requirements. As these models move from research to industry, they could eventually reduce costs and environmental footprints for large-scale manufacturing.
The takeaway
Artificial intelligence is increasingly bridging the gap between theoretical models and physical chemical processes. Engineers can leverage these neural networks to optimize complex hardware without the need for extensive trial-and-error experimentation.
Further reading
For more developments in this field, explore our latest reports in Artificial Intelligence.
More information
View the complete findings in the published peer-reviewed research article.
Source note: This article includes information reported by Nature.
Live Poll
Do you believe advanced machine learning models effectively increase industrial process reliability?
