Nonlinear Multiple Input Multiple Output Process as a Final Project: Bringing Motivation to the Control Classroom

This paper presents the results of applying a final project in an undergraduate chemical engineering control class. The final project is a nonlinear simulation of a process that has 3 or more inputs and outputs that must be regulated. The simulation includes additional disturbances such as input/output noise and load disturbances such as changes in feed concentration. The purpose is to bring the sometimes esoteric world of control which is often taught overwhelmingly in the linear Laplace domain back to the reality of actual process applications that are tricky to control. Not only does this provide motivation for the students to learn control but is introduced throughout the course in parts to accentuate this reality prior to each exam in the class. The first part (20%) has the students perturbing the system using step responses and calculating process gains. Each student is a member of a team working on the process with each student being responsible for one or two inputs. This is completed and graded prior to the second exam which is control of a nonlinear single input single output process. Parts 2 and 3 (60%) are done prior to the third exam where students control a nonlinear 2x2 system. As a group, students calculate the Relative Gain Array to determine best input-output pairings and their tuning strategies. Students then are responsible for one or two loops to develop Multivariable Single Input Single Output controllers using First Order Plus Dead Time models for each loop and Internal Model Control based – PID tuning. Part 4 (20%) takes place at the end of the course where the students close all the respective loops for the system, evaluate baseline controllers, make any tweaks in the controllers, then run various disturbances both load and/or noise in the inputs or outputs. Currently have three systems: a waste sludge treatment plant with inputs gas and mass flowrates and recycle fraction, a bioreactor with inputs of feed, base, oxygen and jacket flowrates, and a non-isothermal jacketed reactor with distillation column with recycle to the reactor with inputs feed and jacket flowrates for reactor, reflux ratio and reboiler for distillation column and recycle stream from distillation column to the reactor. All models are in Matlab and available for use. Overall the students have shown increased efficacy resulting higher grades for the second and third exams due to the final project being introduced. It also is nice way to introduce something that has similar complexity for regulating actual unit operations in industry.