ML in Adsorption

Project Summary: This work highlights how machine learning has been applied to model, predict, and optimize adsorption processes for various pollutants using a range of biosorbents and synthetic materials.

Key Contributions:

• Lead author of an article combining experimental and ML modeling of methylene blue biosorption using bone meal
• Applied experimental design, ANN, and feature analysis to explain biosorption behavior
• Co-author in a study combining ML, experimental design, and clustering to understand uranium removal using hydroxyapatite and bone meal
• Currently developing a ML-driven framework to predict adsorption of dyes (Reactive Black 5) onto biochars using Random Forest and other algorithms

Publications:

• Biosorption of methylene blue by bone meal: experimental and modeling with machine learning and full factorial design – Front. Environ. Chem. (2024)
• Uranium biosorption by hydroxyapatite and bone meal: evaluation of process variables through experimental design – ESPR (2021)
• Removal of Reactive Black 5 Dye by Biochars: Experimental and Data-Driven Modeling Approaches (Under preparation)

Figure 1: ANN structure and variable importance for methylene blue biosorption by bone meal

Figure 2: Cluster and variable analysis for uranium biosorption using hydroxyapatite and bone meal

Figure 3: Random forest regression results for Reactive Black 5 dye removal (in preparation)