Feedback on what to teach in process control

Staff -- Control Engineering, 7/1/2006

I read with great interest the Academic Viewpoint column "Process controls: what to teach" (CE, May 2006, p. 26). While I am not a chemical engineer, I have worked in the process control field (instrumentation R&D and design) for over 30 years. My academic education included two courses in control theory that covered many of the topics discussed in the column. Fundamental to these courses was a firm grounding in Laplace transforms and frequency-domain analysis. My ongoing education has built on the solid foundation I received from these courses.

I found the rankings given [on course topics] a bit incongruous. In particular, it is hard to see how someone can be expected to optimize a control system (#1) without having a basic understanding of the dynamics involved (#9) and how to properly tune a control loop (#7). The identification of bimodality of the answers when based on industry is interesting. It raised, in my mind, the question of whether or not there is a bimodal aspect related to whether or not the respondents were practicing engineers or were higher up the management chain.

David E. Wiklund, senior principal engineer, Rosemount Inc., Emerson Process Management

Expectations of process control engineers, at least with a chemical engineering background, have grown over the years to include issues like optimization and statistical modeling. I work with several people at the Austin Emerson location on those particular features.

Another reason loop tuning and frequency response ranked fairly low in the survey I reported is that about half of the respondents came from the batch process control perspective, such as pharma, which is not so focused on PID control as the continuous process industries.

If you would like to read more about these topics, I have co-authored a textbook with Dale Seborg entitled "Process Dynamics and Control" (Wiley, 2004, 2nd edition) that covers the current viewpoint on loop tuning (using more up-to-date correlations based on transfer function models in Chapter 12) and the use of steady state optimization for profit maximization (Ch.19) and dynamic optimization for constrained multivariable control (Ch.20).

Thank you for your interest.

Thomas F. Edgar, University of Texas at Austin