Integrating artificial‑intelligence assistance into chemical process control education
Chemical process control is traditionally one of the most mathematically demanding courses in the chemical engineering curriculum. Students need to master ordinary differential equations (ODEs), Laplace transforms, transfer functions and proportional‑integral‑derivative (PID) controller design. At V...
Gespeichert in:
| Veröffentlicht in: | Education for chemical engineers Jg. 54; S. 100492 |
|---|---|
| 1. Verfasser: | |
| Format: | Journal Article |
| Sprache: | Englisch |
| Veröffentlicht: |
Elsevier Ltd
01.01.2026
|
| Schlagworte: | |
| ISSN: | 1749-7728, 1749-7728 |
| Online-Zugang: | Volltext |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| Zusammenfassung: | Chemical process control is traditionally one of the most mathematically demanding courses in the chemical engineering curriculum. Students need to master ordinary differential equations (ODEs), Laplace transforms, transfer functions and proportional‑integral‑derivative (PID) controller design. At Villanova University we developed a pilot project to integrate artificial‑intelligence assistance throughout the process‑control pipeline. Three modules were created: (1) Python coding for ODE simulation, (2) Laplace and inverse Laplace transforms with transfer‑function parameter estimation, and (3) PID controller tuning. For each module, pre-class videos and handouts introduced theoretical concepts and demonstrated example code. Students reviewed these materials before class, and class time was used primarily for answering questions and clarifying concepts. Most project work was completed independently outside of class. Students were instructed to ask a large‑language‑model chatbot (ChatGPT) to explain equations, generate and debug Python code, and interpret the physical meaning of models. Pre‑class handouts and videos prepared students for both in-class and independent practice, and anonymous surveys measured learning impact and students’ overall experience. Fifty-nine students completed all three modules and responded to a 15-question survey, which included 13 multiple-choice rating questions and 2 open-ended questions. The rating questions were grouped into categories measuring students’ confidence in Python coding, understanding of ODE modeling and process control concepts, ability to use AI tools effectively, and perceived applicability of the skills to real-world problems. Responses were scored on a five-point scale, where a rating of 5 indicated the most positive feedback. Statistical analysis showed mean ratings ranging from 3.28 to 4.36, with the highest confidence gain reported in the category assessing the ability to apply the learned skills to real-world problems. Qualitative comments reveal that students valued the novelty of using AI tools but desired clearer instructions and more coding guidance. This study demonstrates a fully integrated approach to AI-assisted education in chemical process control, spanning ODE modeling, Laplace transform, transfer functions, and controller tuning within a single coherent framework. By combining Python-based coding in Google Colab with structured AI interaction strategies, the work bridges theoretical concepts and practical applications across the entire modeling–control pipeline. This project offers an example for AI-based control and engineering education studies.
•Developed a three-module AI-assisted Python learning framework for ODEs, Laplace, and PID tuning in Colab.•Used a quasi-flipped model with pre-class videos, handouts, and AI prompts for guided problem-solving.•Engaged 59 students; 47 reported moderate to high gains in coding, process control, and application skills.•Identified strengths in coding, AI use, and engagement; challenges in AI verification and math-heavy topics. |
|---|---|
| ISSN: | 1749-7728 1749-7728 |
| DOI: | 10.1016/j.ece.2025.10.002 |