Overview

Process simulation is a core skill in chemical engineering education, yet many students struggle to translate design concepts into software-specific operations. In tools such as Aspen Plus, learners must manage thermodynamic model selection, unit operation configuration, stream connectivity, convergence settings, and iterative troubleshooting before they can fully engage with the underlying engineering principles.

This webinar introduces an LLM-powered Copilot framework for Aspen Plus as a new approach to supporting chemical process design education. Rather than replacing the engineer or the student, the Copilot acts as a transparent assistant that helps translate high-level engineering intent into structured simulation actions. Students and instructors remain responsible for defining objectives, interpreting results, questioning assumptions, and making final design decisions.

The talk will present how large language models, the Model Context Protocol, modular Aspen Plus tools, and guided “Skills” resources can be combined to support human-AI collaborative process design. Particular attention will be given to how this approach may be used in teaching environments to reduce the operational barrier of simulation software while strengthening conceptual learning, reflective problem solving, and engineering judgement.

The webinar will use three Aspen Plus case studies as teaching-oriented examples:

1. Water–ethanol binary distillation: Demonstrating how students can move from a natural-language design objective to a working Aspen Plus model, including stream definition, column configuration, convergence, and purity tuning.
2. Pressure-swing distillation for ethanol–water separation: Showing how an AI-Copilot can support thermodynamic reasoning by identifying feasibility limitations and encouraging students to distinguish between software execution and process constraints.
3. IPA extractive distillation reconstruction from literature: Demonstrating how students can reconstruct a published process flowsheet, identify missing equipment, initialise recycle streams, troubleshoot convergence, and compare simulation results with literature data.

These examples illustrate how AI can support active learning: students can test design hypotheses, inspect AI-generated actions, evaluate simulation feedback, and discuss why certain design choices succeed or fail.

Speaker

Jia-Lin Kang, Associate Professor, National Chung Cheng University

Jia-Lin (Conlin) Kang, Ph.D. is an Associate Professor in the Department of Chemical Engineering at National Chung Cheng University (CCU), Taiwan. Dr Kang received his doctorate from National Tsing Hua University and has developed research expertise in process systems engineering, process control, physics-informed AI models, computational fluid dynamics, industrial HAZOP analysis, and digital twins.

His recent work focuses on bridging chemical engineering knowledge with artificial intelligence, particularly deep learning, large language models, and human-AI collaborative process design. He has developed an LLM-powered Copilot framework for Aspen Plus by integrating the Model Context Protocol with structured simulation tools. In addition to its research contribution, this work aims to support the next generation of chemical engineers by reducing the operational barrier of simulation software and enabling students to focus more deeply on process understanding, design reasoning, and engineering decision-making.

The material presented has not been peer-reviewed. Any opinions are the presenter’s own and do not necessarily represent those of IChemE or the Education Special Interest Group. The information is given in good faith but without any liability on the part of IChemE.

Time

09:00–10:00 BST.

Software

The presentation will be delivered via Microsoft Teams. We recommend downloading the app from the Microsoft website, rather than using the web portal.