2025 ASEE Annual Conference & Exposition

The course Chemical Process Control involves ordinary differential equation simulation, and while students typically take process modeling before studying process control, they may forget essential ODE concepts over time. Artificial intelligence tools, such as ChatGPT, offers a novel approach to refresh students' understanding of ODE modeling and improve their ability to simulate ODE models and prepare them more effectively for the study of process control. This study introduces a self-paced, AI-assisted learning project for Chemical Process Control course, with potential applications in other chemical engineering and engineering courses. This pilot test aims to evaluate the transformative potential of AI-driven learning in engineering education.

In this self-paced, ChatGPT-based online learning project, the instructor provided students with a handout containing three ODE models, ranked by difficulty: the car distance model, the classical enzymatic model, and the fed-batch bioreactor model. Example questions were included to guide students in using ChatGPT to understand each term in the ODE models, numerical solutions used to solve ODE models, Python coding in Google Colab to simulate ODE models. Students then applied their models to open-ended, real-world scenarios chosen by them. They were required to share their Google Colab programs with the instructor and submit a Word document summarizing their understanding of the ODE models, their simulation results, and the real-world predictions they derived. All students were connected through Google Classroom to inspire collaborative learning.

The project was tested by six high school volunteers in Summer 2024 to gather feedback and assess its feasibility before its introduction to college students in Spring 2025. The hypothesis was that if high school students with limited exposure to ODE modeling and simulations could complete the project, senior college students would also succeed. The aim was to use student feedback for improvements. All six students (two 10th-graders, three 11th-graders, and one 12th-grader) successfully developed and simulated the three ODE models in Python using Google Colab, only supported by ChatGPT, with an average completion time of 4.8 ± 1.8 hours. They achieved decent accuracy rates in multiple-choice quizzes: 100% on ODE model understanding, 83.3% on numerical methods, 66.7% on Python coding, and 83.3% on applying the bioreactor model. The survey of students’ self-evaluation reported increased confidence in understanding ODE models, using Python for simulations, and utilizing AI tools like ChatGPT. Students reported that ChatGPT was helpful in providing timely assistance, allowing them to progress independently. While interest in STEM showed a slight increase, some challenges were noted. They suggested improvements, such as clearer instructions on AI-assisted learning and an expanded set of ODE models for more practice to enforce their learning. Some students also found it challenging to modify ChatGPT-generated code and struggled with unfamiliar terminology, particularly regarding bioreactors and enzymatic systems.

This pilot study highlights the potential of AI tools like ChatGPT to enhance learning in computational modeling and process control. It may provide a valuable addition to engineering education.