Application of Machine Learning-Based Models to Understand and Predict Critical Flux of Oil-in-Water Emulsion in Crossflow Microfiltration

Abstract

Random Forest (RF) and Neural Network (NN), respectively, were employed to understand and predict the critical flux (J crit) of oil-in-water emulsions in crossflow microfiltration. A total of 223 data sets from various studies were compiled, with nine operational parameters and one target variable of critical flux. RF indicated crossflow velocity (CFV) as the most dominant parameter in determining critical flux, outweighing surfactant and oil variations. Exceptions were found in specific cases when casein concentration was the most dominant, since the smaller sizes of casein significantly decreased J crit. The NN model predicted the best when all nine input parameters were integrated and the worst when CFV was the sole parameter used for model development, even though CFV was identified as the most dominant. The results here demonstrate the usefulness of machine learning tools to enhance the understanding on and prediction of critical flux without any governing equations.

Understanding Single-Protein Fouling in Micro- and Ultrafiltration Systems via Machine-Learning-Based Models

“…In our previous study, NN and RF models were used to predict critical flux and the significance of each parameter during MF with oil-in-water emulsion solution, respectively. 102 The results showed good predictive ability of these models with low error value and R 2 above 0.9. ML methods have also been implemented for organic solvent nanofiltration 103,104 and a wastewater treatment plant, 105 which attests to the usefulness of such tools in the membrane filtration field.…”

Review of Machine Learning for Hydrodynamics, Transport, and Reactions in Multiphase Flows and Reactors

“…They presented a detailed prediction study of flow field parameters including the evaporated liquid vapor fraction, temperature, pressure, velocity, and spray penetration and Sauter mean diameters in the liquid phase with reasonable accuracy (ranging from 0.05 to 13.5%). Chew and co-workers 152,154,206,213,244,245 directed efforts toward a deep understanding and an enhanced prediction of fast gasparticle riser flow characteristics assisted by several common ML methods including ANN, RF, and self-organizing map (SOM).…”

“…Persistent lack of physical comprehension continuously stymies preferable prediction performance of the key parameters in multiphase flow and reactor systems, although scientists have made systematic contributions to experimentally formulated correlations throughout the past decades. − The correlations of the key parameters in multiphase units are commonly expressed by gas/liquid/solid phase properties, operating conditions (e.g., phase concentration, velocity, and temperature), devices configurations (e.g., height and diameter), or a combination of them in dimensionless forms such as Archimedes, Froude, Nusselt, Reynolds, Sherwood, and Weber numbers. However, the prediction discrepancies between the existing empirical correlations of key parameters such as the particle entrainment and minimum fluidization velocity in gas-particle riser flows can reach several orders of magnitude. , Fortunately, the advanced research and development of flexible ML tools have the potential to complement the incomplete knowledge to boost the prediction ability of key multiphase field parameters such as mass flow rate/flux, − minimum fluidization velocity, , mixing rate/index, , overall/local hold-up, − pressure/pressure drop, − velocity, ,− temperature, − and other parameters − in multiphase/particulate flows and reactors. Note that interested readers may be referred to a relatively comprehensive list of the existing literature summarized in Table S4.…”

“…They presented a detailed prediction study of flow field parameters including the evaporated liquid vapor fraction, temperature, pressure, velocity, and spray penetration and Sauter mean diameters in the liquid phase with reasonable accuracy (ranging from 0.05 to 13.5%). Chew and co-workers ,,,,, directed efforts toward a deep understanding and an enhanced prediction of fast gas-particle riser flow characteristics assisted by several common ML methods including ANN, RF, and self-organizing map (SOM). In their essential study on the determination of the relative dominance of the riser flow parameters, it was suggested that the radial position is of key importance to the local solids flux and elutriation, and the overall solids flux has the most dominant effect on the local solids hold-up.…”

Machine Learning and Data Science in Chemical Engineering

“…This special issue of Industrial & Engineering Chemistry Research presents an excellent collection of articles from internationally renowned researchers from all around the world to showcase the application of machine learning and data science in the aforementioned chemical engineering problems. − We truly appreciate the efforts from all contributing authors to make it happen. We hope these articles provide new insights and perspectives as to how machine learning can be used in a wide variety of chemical engineering problems, and stimulate more creative solutions to existing and future challenges.…”