(506e) Chemical Engineering Analysis through Systematic Optimization

Xie, W. - Presenter, University of Minnesota - Duluth
Davis, R., University of Minnesota Duluth
With the rapid development of computing technology, computer-aided design skills are in high demand. The course, CHE5031 Chemical Engineering Analysis, has been offered at the University of Minnesota Duluth since 2015. This course aims to help students build a fundamental understanding of chemical engineering systems by developing mathematical and statistical models and simulations using digital computers. It also aims to develop a systematic understanding and a critical awareness of process optimization and analysis of results.

In 2021, MATLAB was selected as the computational language in the course. As a multi-paradigm numerical computing environment and proprietary programming language developed by MathWorks, MATLAB is a powerful tool for computer-aided design and Chemical Engineering analysis.

As shown in Figure 1, the course is organized into three major sections with 10 chapters:

1. Chapters 1 and 2 introduce modeling in chemical engineering processes and MATLAB skill preparation.

2. Chapters 3 through 7 introduce some advanced algorithms and computational methods with case studies in Chemical Engineering.

3. Chapter 8 through 10 integrate algorithms and methods for tackling complicated case studies in Chemical Engineering. There are many Chemical Engineering-related examples in each chapter at both the undergraduate and graduate levels.

This course’s most distinguishing feature is to apply systematic optimization for chemical engineering analysis to help students build their confidence for optimized data-driven decision-making. Students do not just stop learning on the existing optimization toolbox and global optimization toolbox in MATLAB - they gain added knowledge of the limitations of MATLAB’s optimization solvers. Additional topics include enhanced optimization methods, such as Lagrange’s method of converting constrained optimization to non-constrained optimization, conversion of inequality constraints to equality constraints, the penalty function method, and an enhanced constraints method to avoid the unwanted “zero” local optimal point.

The optimization techniques are demonstrated by case studies in model fitting, solving complex equations, and process optimization. Students also learn optimal experimental design to research influencing factors, their levels, the significance of each factor, and their combined effects. Besides the in-class exercises, students have four main assignments:

1) Model fitting of a particle size distribution via different ways by MATLAB apps and many optimization methods.

2) Optimization of the placement of a petrol station located on the road among 5 towns and further questions with weighted optimization, and with both equality and inequality constraints.

3) Optimization of a nodal network in a Cartesian grid to simulate steady-state 2D conduction heat transfer without internal heat resources.

4) Application of orthogonal experimental design for finding optimal parameters that minimize the operation costs in solvent extraction flowsheets using a ternary diagram.

Our Chemical Engineering Analysis course bridges the skills gap for our graduate students and some undergraduate students who selected this course, especially enabling them to use the “data science” techniques, skills, and modern engineering tools necessary for engineering practice.