CHEN E4580 Artificial Intelligence in Chemical Engineering
Instructor: Venkat Venkatasubramanian
Credits and Meeting Times: 3 credits; Mon-Wed 1:10 – 2:25 pm
Prerequisite/Co-Requisite: Background in chemical engineering and Python/MATLAB
programming, or instructor permission.
Background: The practice of chemical engineering routinely involves decision-making
to solve challenging problems that often arise in process/product design, fault diagnosis and control,
optimization, and process operations and safety. To accomplish this, historically, chemical engineers have
adapted modeling concepts and techniques from different disciplines such as applied mathematics and
operations research over the decades. To this, Artificial Intelligence (AI) methods have increasingly been
incorporated in the last three decades. Contrary to common impression, AI in chemical engineering is not
new – it has a 35-year-long history and a body of literature comprising of a few thousand papers in
chemical engineering1-2. It has become clear that AI-based approaches need to be part of the modeling
arsenal of well-educated chemical engineers. This course is offered to address this important need.
Description: This course is aimed at chemical engineering students to teach them how to incorporate
AI-based modeling in combination with first-principles-based models derived from the understanding of
the physics and chemistry (and biology) of our products, processes, and systems. To accomplish this, this
course will combine the traditional symbolic AI with the more recent data-driven AI. In this regard, this
course is different from the standard data science course which typically focuses only on the data-driven
aspect.
Topics would include: Chemical engineering modeling approaches, knowledge representation, symbolic
reasoning and inference, knowledge-based systems, statistical data analysis, and machine learning
methods such as clustering, neural networks, random forests, Bayesian networks, and directed evolution.
All these will be discussed using chemical engineering case studies in process monitoring, diagnosis, and
control, process/product design, scheduling, optimization, and process hazards analysis. In particular, the
course will address the development of hybrid models that combine first-principles knowledge of the
underlying physics & chemistry with data-driven techniques and the development of causal mechanism-
based models, in contrast with typical AI courses taught in computer science departments.
Grading: Homework 40% and Project 60% (using MATLAB, Python, TensorFlow and/or Keras)
Textbook: There is no textbook. Several papers from the literature will be made accessible during
the course.
1. Venkatasubramanian, V., Artificial Intelligence in Process Engineering: Experiences from a Graduate
Course. Chem. Eng. Educ., Fall Issue, 188–192, 1986.
2. Venkatasubramanian, V., The promise of artificial intelligence in chemical engineering: Is it here,
finally?. AIChE J., 65: 466-478, 2019.
