Grad Seminar: Using Machine Learning to Evaluate the Potential Environmental and Human Toxicity of Solvents Proposed for use in Post-Combustion Carbon Capture
Abstract
Many industrial activities, such as energy generation from fossil fuels, cement production, and steelmaking, yield carbon dioxide as a byproduct. Since the Industrial boom, CO2 concentrations in the atmosphere have been increasing. One way to offset carbon emissions is to capture carbon dioxide in flue gases. Fossil fuel plants can be retrofitted with Post Combustion Carbon Capture (PCCC) units. The most mature technology for PCCC is amine absorption. Different amines have been proposed for use. In this study, machine learning models were trained to predict indicators of potential environmental toxicity: volatility, lipophilicity, mutagenicity, and neuroactivity. 151 proposed amine solvents are compared using their predicted properties, to determine which classes are likely to be environmentally toxic, and which are likely to be safe.
Presenter
Fatima Ghiasi, MASc candidate in Systems Design Engineering
Join online