(577b) Teaching Sustainability to Undergraduate Students: Predicting Life Cycle Inventory Data for Environmental Impact Assessment

Advances in the chemical industry have led to the discovery of new compounds that do not have documented properties. Due to their novelty, access to their Life Cycle Inventories (LCIs), environmental impacts, and other sustainability metrics can be challenging as there is little available information in the literature. Furthermore, there are existing chemicals without such data available. Predicting these metrics are essential in assessing the sustainability of novel and existing chemicals. Experimentation or computationally intensive large-scale computations such as Density Functional Theory (DFT) and Molecular Dynamics (MD) are time and energy intensive which prevents facilities from performing Life Cycle Assessments. Since LCI data are required to perform any environmental impact assessment, it is imperative to find ways for its estimation to enable the development of greener processes.

In this work, we use Machine Learning (ML) methods to predict LCIs readily from molecular descriptors and thermodynamic properties. We first construct a database containing the properties of existing chemicals, such as density, molecular weight, boiling point, functional groups, number of atoms, toxicity, and many more. Overall, 23 and 200 thermodynamic and molecular properties were gathered respectively. We then use the Ecoinvent database from SimaPro® and the National Renewable Energy Laboratory’s United States Life Cycle Inventory database to gather LCI data for the existing chemicals. The gathered data is then used to build a supervised learning model using the gradient boosted trees algorithm in python. This model will be used to predict several different impact categories, human health, ecosystem quality, climate change, and resource requirement. Human health impact is measured in Disability Adjusted Life Years (DALY), ecosystem quality impact is measured in the Potentially Disappeared Fraction of a species in a given area per year (PDF*m²*year), climate change is measured in kilograms of carbon dioxide equivalent (kg CO₂ eq), and resource requirement is measured in megajoules (MJ) .

In this work we explore the teaching of sustainability assessment and design through the Engineering Clinic course offered by the Henry M. Rowan College of Engineering at Rowan University. The Engineering Clinic is a 2-credit course offered every semester with 3-hour meetings scheduled twice a week. Engineering Clinic teams are comprised of undergraduate students, graduate students, and faculty advisors. The faculty advisors’ network with industries and research institutions and obtain grants for this work from federal and regional organizations as well as industries. In addition, the faculty advisors define the overall goals, tasks, and learning outcomes of the project. The graduate students are responsible for the dissemination of the tasks defined by the faculty and mentorship of the undergraduate students. The undergraduate students meet twice weekly with the graduate students and faculty advisors to share their work progress and receive feedback and directions for the next steps. The deliverables for the undergraduate team include a final comprehensive report and technical presentation towards the end of the academic semester. Using a well-established rubric for final grading, the students are assessed based on their overall performance in the project in regards to initiative, clear problem definition, identification of solution methods and application, technical presentations and final report. During the presentation, students are evaluated on various categories such as project introduction effectiveness, organization and clarity of presentation slides, visual aids for effective communication, and overall handling of discussions. Furthermore, many students also participate in regional and national conferences to present their work, thereby enhancing their professional and networking skills.

This undergraduate-graduate-faculty synergy allows for sufficient technical input from the faculty advisors, an opportunity for the graduate students to interact with stakeholders, and the undergraduate students a chance to apply their classroom knowledge to real-world engineering problems.