Chemometric Analysis of NMR Spectra and Machine Learning to Investigate Membrane Fouling

Abstract

Efficient membrane filtration requires the understanding of the membrane foulants and the functional properties of different membrane types in water purification. In this study, dead-end filtration of aquaculture system effluents was performed and the membrane foulants were investigated via nuclear magnetic resonance (NMR) spectroscopy. Several machine learning models (Random Forest; RF, Extreme Gradient Boosting; XGBoost, Support Vector Machine; SVM, and Neural Network; NN) were constructed, one to predict the maximum transmembrane pressure, for revealing the chemical compounds causing fouling, and the other to classify the membrane materials based on chemometric analysis of NMR spectra, for determining their effect on the properties of the different membrane types tested. Especially, RF models exhibited high accuracy; the important chemical shifts observed in both the regression and classification models suggested that the proportional patterns of sugars and proteins are key factors in the fouling progress and the classification of membrane types. Therefore, the proposed strategy of chemometric analysis of NMR spectra is suitable for membrane research, which aims at investigating comprehensively the fouling phenomenon and how the foulants and environmental conditions vary according to the filtration systems.

An evaluation of homeostatic plasticity for ecosystems using an analytical data science approach

“…For example, in a practical wastewater treatment process, 31 P NMR methods were applied to model biofilm phosphorus release performance at different pH [98] . Moreover, 1 H NMR and machine learning extreme gradient boosting (XGBoost) has also been used to understand the mechanism of membrane fouling (reduction in membrane permeability), which is a problem in membrane bioreactors (MBRs), an essential part of wastewater treatment operations [99] . Changes in the nature of membrane fouling material as a function of conditions have been investigated using NMR data [100] .…”

A Review on Membrane Biofouling: Prediction, Characterization, and Mitigation

“…Various studies demonstrated the successful application of different AI techniques for the prediction of membrane biofouling; for instance, Yokoyama et al [48] combined NMR spectroscopy and several ML models to predict the maximum transmembrane pressure (TMP), analyze the chemical compounds causing fouling based on a chemometric analysis of NMR spectra, as well as determining their effects on fouling progress. Out of the tested models, random forest (RF) exhibited the highest accuracy in the analysis of the NMR spectra; in addition, the analysis revealed that among the bacterial-EPS components, polysaccharides contributed the most to membrane biofouling.…”

Deconvolution and Analysis of the 1H NMR Spectra of Crude Reaction Mixtures

“…Prior approaches to the automated analysis of NMR spectra employed machine learning models to analyze the spectra. The majority of current models analyze the spectrum of a pure sample of an unknown compound to determine its identity. − These models can be employed only on spectra that contain a single compound, but full workflows capable of analyzing multicomponent spectra are becoming more commonplace. − Although powerful for their trained tasks, these machine-learning models require a database containing spectra of each potential component and cannot identify novel compounds.…”