This Control Talk column appeared in the January 2018 print edition of Control. To read more Control Talk columns click here or read the Control Talk blog here.
Greg: In my golden years of gradual retirement to being a part time developer of models for Emerson’s Process Simulation team, I have been fortunate to be able to focus more on writing, mentoring and teaching. I get to use most of what I have discovered in my roles in engineering technology and research and development to further advance what I initially learned from Shinskey, as detailed in “The greatest source of process control knowledge.” Particularly rewarding is seeing basic and advanced process control, in my specialties of pH, bioreactor, chemical reactor and compressor modeling and control, come to fruition via the virtual plant. Emerson’s Process Simulation team has five to 10 coops and interns each year to work along with the 30 or so young engineers who have rapidly gained proficiency in providing dynamic models for operator training and process improvement. The enthusiasm and computer skills, combined with the education in bioprocess, chemical, mechanical and systems engineering, and exploiting the knowledge of key technical resources, helps lead to a fast takeoff in modeling and control productivity. The culture and talent reflect the potential for new companies using the incredible power of today’s computers and software to fully exploit decades of knowledge to revitalize the automation profession. The knowledge extensively flourished in the 1970s and 1980s, when brilliant specialists at companies like DuPont, Monsanto and Foxboro developed and, most importantly, published their expertise. Most of the articles, papers and books that form my core knowledge, including equations for first-principle models and test results, are from this era.
Stan: Here, we have the opportunity to visit with two interns: Lydia Schreiber, a student in bioprocess engineering at University of Missouri (Mizzou), and Patrick Brennen, a student in chemical engineering at Missouri University of Science and Technology (MS&T), to understand more about the path from university to industry. We begin with the question, “What led you to choosing your major and your university, and what are some of the courses?”
[sidebar id =1]
Lydia: My dad has a master’s degree in civil engineering, and is working with the Environmental Protection Agency (EPA) focusing on regulations for carbon emissions in plants. I was intrigued by how a 3D printer was able to produce stem cells. I initially was thinking of the biomedical track. When I got to school, I realized there were many paths in bioprocessing, such as food, pharmaceutical and environmental. What turned me off in the pharmaceutical path were the long research, testing and development time and the extensive red tape in the commercialization and production of drugs. The environmental emphasis has a lot of overlap with civil engineering dealing with hydrology and soil. I am interested in sustaining natural resources. I think I could help people on a greater scale through sustainability rather than through biomedical.
Here are some courses I've taken and am about to take that are required for bioengineers, plus some technical electives. The program is pretty flexible beyond the core requirements, and developing an emphasis within the bioengineering major is fairly easy. Mizzou also has a ton of extracurricular paths to explore your interests, like research with the strong agriculture and biology departments or clubs like “Sustain Mizzou” and the “Biomedical Engineering Society”.
Core courses
Engineering Analysis of Bioprocesses: This class brought together conservation laws and problem-solving methods in the context of biological processes. We used the book Bioengineering Fundamentals by Ann Saterbak.
Biomaterials: This course relates material structures to their properties and discusses their use in the context of biological engineering. e.g., implant design.
Heat and Mass Transfer in Biological Systems: The name is pretty self-explanatory.
Applied Electronic Instrumentation: Intro to fundamentals of electronics, analog/digital circuits, signal conditioning, computer interfacing and measurement techniques in computer-based instrumentation systems.
Technical electives
Computational Neuroscience: We used differential equations and scientific laws to guide our understanding of neurons. We also built neuron simulations based on those laws using simulation software developed by Yale professors called NEURON.
Intro to Biomedical Imaging: We learned about the physics and mathematics behind data acquisition in various types of medical imaging machines (e.g., CAT scans, X-rays, MRI).
Other courses I'm interested in, schedule permitting: Materials Engineering, Biomolecular Engineering and Nano Biotechnology, Watershed Modeling, Biomass Refinery Operations, Feedback Control Systems, Food Process Engineering, Soil and Water Conservation Engineering
I really appreciate the variety of courses and faculty we have in our department. There's a lot of collaboration and overlap with other departments, too. I think that's why I chose Mizzou—the departments beyond engineering are strong and eager to work with us. It's opened up my mind to what engineering is and could be.
Patrick: My whole life, I’ve been exposed to many engineers in my family, including my dad and a few cousins. I am good in math and science and wanted to apply these skills in industry, so engineering was a good fit. I liked the MS&T campus and I’ve only heard good things from people who went there. I decided on chemical engineering because it was more about the process of productively transforming chemicals from point A to point B and the big picture, rather than the chemistry of developing a new molecule. I liked environmental, but there is more of a research than industrial focus in those courses.
Our curriculum is pretty full with standard classes like Thermodynamics, Heat Transfer, Chemistry, Mass Transfer and Fluid Flow. I was able to use my technical electives to explore the field of research. Then, when I decided that research wasn't for me, I went on to look at industry and found different positions to try.
Overall, I'd say MS&T has a bit of a different approach than Lydia's experience. There is not much room for making it your own (although I have recently seen that they are consolidating some classes to give more room for technical electives). Instead, the school has been very successful in getting the students the experience that they are looking for, in both research and industry. There is a huge variety of research opportunities with professors and if you don't see one you like, they'll give you money to research something on your own. There are also hundreds of companies that come down to campus willing to hire interns and coops in the industry. The ease of finding useful experience has been the most crucial part of my own education.
Greg: What type of computer modeling do you use at your universities?
Lydia: We mostly use MATLAB for the dynamic models, control system and PID tuning. The models are developed by professors and graduate students.
Patrick: The text book for our Process Control class emphasizes the frequency domain and model algorithmic control. But the person teaching our control theory course is from industry, so I’m expecting more industry focus in the future. Rolla already has a reputation for preparing students more for industry rather than academia. We have a new virtual plant lab. I just finished developing the model for a Texas A&M lab methanol and water distillation column to run with the current control system. Emerson retiree Doug White and advanced applications specialist James Beall, who are Texas A&M alumni, are helping to improve the lab and create a virtual plant as well. We are looking forward to moving beyond what we are presently doing in our virtual plant, which is configuration of a simple PID loop with a tieback simulation.
Stan: What type of job are you seeking?
Lydia: I need to do some research on the types of jobs out there. I would like to be in a development group to solve environmental problems and to improve existing systems through innovation. I am talking to classmates and searching the Internet. There is a job fair at the university where a hundred or more companies show up to talk to engineering students.
Patrick: I have worked at a chemical plant, where I spent a lot of time just to get the paperwork done to install a control valve. I like environmental technologies, but am concerned the paperwork will be even more overwhelming. I enjoyed the experience of using the virtual plant. I am attracted to pilot plants and greenfield projects that could offer more in terms of how dynamic simulation is used for innovative solutions, and more room for true engineering work. There is also a job fair at the university where 300 companies show up to talk to engineering students, so there are plenty of options out there.
Greg: From what I have seen in most university courses, there are few practical courses on modeling and control. While there are some exceptional books written by former professors such as Karl Astrom, Bill Luyben and Cecil Smith, nearly all of the university books on modeling and control focus on frequency response, Laplace transforms, Z-transforms and special algorithms developed by universities not realizing that they have never made much of an entry into industry and are not a better solution than PID for single-loop control and model predictive control (MPC) for multivariable control and optimization. The new algorithms usually ignore the most important need of minimizing the effect of disturbances at the input (not output) of the process, and interactions by tuning. The tests fail to use the extensive, well documented and tested capability of the incredible spectrum of readily configurable PID and MPC features and options. Often, the tests do not use modern methods of tuning the PID. A common mistake made in the literature by developers of new algorithms or tuning rules are tests with zero dead time to show the supposed greatness of their ideas. Dead time is never zero just by virtue of having an automation system. If the dead time was zero, prefect control would be possible because the controller could immediately see the disturbance and any correction made. If there was no noise, the controller gain could theoretically be set at infinity or, more realistically, the high gain limit allowed in configuration; similarly, the reset time could be set at zero or, more realistically, the low reset time limit allowed in the configuration. The controller is always stable despite changes in open-loop gain or open-loop time constant.
Instead of reinventing the wheel in MATLAB, to revitalize process control, universities need to use the virtual plant with all the power of modeling and control as practiced in industry built in. The virtual plant can provide realistic, flexible and fast exploring, discovering, prototyping, testing, justifying, deploying, testing, training, commissioning, maintaining, troubleshooting and auditing to achieve continuous improvement, showing the “before” and “after” benefits of solutions from online metrics. See the August, 2017 Control feature article “Virtual Plant Virtuosity” for insight into the future of modeling and control. We all need to spend more time exploiting the extensive untapped capability of PID, which is proven to provide near optimal load disturbance rejection in the research documented by Bohl, A. and McAvoy, T. in “Linear Feedback vs. Time Optimal Control II - The Regulator Problem” (Ind. Eng. Chem., Process Des. Dev., Vol. 15, No. 1, 1976).
Hunter Vegas and I are seeking to partner with a university, Control and ISA to provide a flexible practical online degree with self-learning virtual plant demos and exercises based on the upcoming Sixth Edition of our Process/Industrial Instruments and Controls Handbook. The courses would use a succession of lecturers who are ISA Mentor Program resources and protégés, ISA book coauthors and handbook contributors. My goal is to make the most out of using PID, MPC and process knowledge for innovations in control systems that improve process safety and performance.
“Top 10 expectations of engineering graduates”
10. Always working with the latest technological advances.
9. Everyone loving the new and exciting changes you are making to their processes.
8. Traveling to towns with more than 500 people.
7. Expense paid trips to see the world.
6. Rock at work with headphones on.
5. Play ping pong.
4. As much free coffee, tea and soda as you can drink.
3. Lunch time board games.
2. No more Laplace and Z Transforms.
1. No homework!
