Ask the Automation Pros: How Is Process Control Different from Other Industrial Control Disciplines?

The following discussion is part of an occasional series, “Ask the Automation Pros,” authored by Greg McMillan, industry consultant, author of numerous process control books and 2010 ISA Life Achievement Award recipient. Program administrators will collect submitted questions and solicit responses from automation professionals. Past Q&A videos are available on the ISA YouTube channel; you can view the playlist here. You can read posts from this series here.

Russ Rhinehart’s Question

What are the unique attributes differentiating process control from other domains, such as mechatronics, aero, automotive, power, electronic, etc.?  And how do these differences shape essential techniques and necessary understanding?

Russ Rhinehart’s Thoughts

1. Time-constants in process control are long relative to other domain applications.  This means that continuum mathematics such as Laplace analysis are fully adequate, as opposed to needing z-transform or back-shift operator analysis. It also means that 0.1 to 10 Hz control frequency is fully adequate, which relieves demand for computational speed.
2. There is no controlled oscillation, we tune to avoid oscillation, and abhor the ultimate method of tuning because of safety and other issues, so frequency analysis has little utility.
3. Delays (deadtime) can be very long (analyzers, lab analysis, transport time) meaning that deadtime compensation or accommodation methods (such as cascade) are very important strategies. 
4. The environment is often dangerous due to fugitive emissions and flammable or toxic materials.  So electronic devices often need to be replaced by pneumatic devices or need to be enclosed in spark-proof enclosures. 
5. There are often safety overrides during which a secondary controller takes over, related to excessive heat, reaction runaway, condensing vacuums, …  This means that tuning the unselected controller takes some skill, and integral wind-up is a concern making external reset feedback very useful.
6. There are numerous constraints (safety, operability, cavitation, specification, degradation, …) which need to be accommodated in the control action. 
7. Process gains and dynamics continually change in time. Throughput is always changing due to market demand, and this changes delays and time-constants. And since the process responses are nonlinear, changes in production rate also affect steady state gains. Additionally factors that affect the process (ambient losses, fouling, catalyst activity, raw material composition) are also continually changing. And we continually change tank level setpoints for inventory management, which changes mixing time constants and blending of upsets. So, the model that might give insight as to controller coefficient values cannot be known. PID controllers need to be continually adjusted while online. Usually, they need a heuristic fine-tuning method, because formal model-based approaches give imperfect results, upset the process and take too much time and attention. This is very unlike mechanical, electrical or physical equipment in which the model does not continually change.
8. The education level of the folks who do the adjusting are not degreed engineers, or even PhD designers. They are operators and technicians, some with an associate degree, many with just a high school diploma. This requires simplicity of concepts, language and maintenance procedures.  For instance, although I don’t like how integral time and derivative time are usually explained, it seems appropriate to the mathematical skill of those operating the controllers. [For integral time: “If one has a standard form of the PI controller and the disturbance makes a step and hold, and the controller output is not connected to the process, then the integral time is the same as the time it takes the integral action to repeat (duplicate) the initial proportional bump. The integral action is termed reset action, because it resets the controller bias, and the time constant units are described as minutes-to-repeat.” And for derivative time: “If the standard form of the PD controller has a ramp disturbance, but no effect on the process, then the derivative time is the time the proportional action duplicates the step output from the derivative term.”]
9. Many processes are multiple-input and multiple-output, and interaction requires any of several decoupling strategies (ratio, de-tuning one loop, one-way static decoupler, dynamic decouplers, model-based control).
10.  In the process industry, Goodness of Control Metrics mostly relate to the regulatory period. Processes operate at extended steady periods between set point changes. Regulatory metrics are in contrast to QAD-like metrics for set point changes. Variance in regulatory periods relates to the setpoint deviation from specification and constraints, which is required to prevent or minimize violations and waste or seconds/recycled products. Reduction of variance in regulatory periods is the economic key. Rather than IE, IAE, ISE, ITAE, etc. for set point changes, variance, equivalent to the time-normalized ISE (nISE) during extended regulatory periods, is the key metric related to the set point deviation from specification or constraint. Tuning should be to minimize nISE, not metrics related to set point changes. 

Julie Smith’s Thoughts

1. In discrete industries, you can see the product as it is being made (think a bottling line or parts assembly). In chemical processes, you cannot! Everything is contained in pipes or vessels, by design. So the sensors are critical to judge whether the process is operating where it should be. They are also key to inferring quality, which is usually measured after the fact and offline.
2. Both chemical and discrete industries must manage risks regarding personnel safety. But chemical process also must consider risks to the community. Loss of life or limb from a machine incident is terrible, but it affects mainly those within the plant fence. A chemical release, however, can travel offsite and adversely impact those outside the fence line. Much additional effort (layers of protection such as overrides and SIFs) goes into preventing this from happening. 

Erik Cornelsen’s Thoughts

1. A core duty of automation and controls engineers involves developing and maintaining PLC and DCS control software. Yet the industry faces a persistent challenge: While modern software engineering has embraced continuous integration and continuous delivery (CI/CD) pipelines for automated building, testing and deployment, industrial automation largely relies on manual program updates. This outdated approach introduces unnecessary risks, consumes valuable engineering time and significantly limits the pace and reliability of control system improvements.
2. Industrial automation systems also employ fundamentally different communication protocols compared to traditional IT networks. Instead of following the standard 7-layer Open Systems Interconnection (OSI) model like HTTP, FTP or SMTP, industrial protocols such as PROFINET, EtherNet/IP and Modbus TCP are purpose-built for deterministic, real-time performance in distributed control systems. These specialized fieldbus and industrial Ethernet solutions prioritize reliable, time-sensitive data exchange over strict adherence to layered protocol architecture, making them essential for modern automation environments.

Greg McMillan’s Thoughts

I did a Control Talk column about 20 years ago with an expert in machine control for parts manufacturing. I cannot find it but from what I remember, an incredibly fast controller execution time was required because there was no process time constant, no backlash or stiction and nearly no dead time. Disturbances could be considered as being on the process output and were minimal. The main goal was a fast setpoint response (e.g., servo mechanism response), and possibly motion control. It would be great if we could get some input from experts working in this and other industries that are so different than our world of process control.

In process control, we need to deal with many variable gains, dead times and time constants in the measurements, final control elements, equipment and process. Often unrecognized is the effect of lost motion and resolution limits on a control valve’s 86% response time and the valve to system pressure drop ratio on a valve’s gain seen as slope in the installed flow characteristic. Even less recognized is the effect of inverter design, input card resolution, rate limits and operating to static pressure ratio on variable frequency drive (VFD) dynamic response and rangeability. Thermowell and electrode time constants can change by orders of magnitude depending on installation and process conditions or in the case of pH simply by age. The problems with fast integrating processes (e.g., fast furnace pressure) and especially runaway processes (e.g., highly exothermic reactors) can be severe to the point of requiring loops to stay in automatic with fast and aggressive response.

I have used first principle ordinary differential equations to detail how process gains, time constants and dead times change with operating conditions including the differences between self-regulating, near-integrating, true integrating and runaway processes as seen in ISA-5.9-2023 Annex F. I have been fortunate to have has close connections with key experts in pH measurement getting many details rarely known about pH electrodes. Plant applications and pilot plant tests of different types of electrodes revealed a lot of misconceptions in the literature. I have documented this knowledge in my ISA book Advanced pH Measurement and Control Fourth Edition. I accumulated articles by temperature measurement experts dating back 40 years or more ago on the how to get the best sensors and thermowell installations. I have personally experienced many problems of valves with supposedly high-performance valves that are on-off valves posing as throttling valves. I have documented the many valve response problems and solutions in ISA-75.25.02-2024 Annex A. While the literature often gives a view of VFDs as offering a so much better performance than control valves, I was able to track down in a key book and article detailing the severe limitations often experienced in VFD performance often in an attempt to reduce VFD cost. I summarized this knowledge in my response and follow-up in the ISA Interchange “Ask the Automation Pros” post “What Is the Best VFD Design and Installation Plan.”

Ed Farmer’s Response

Process Automation per the International Society of Automation

The dictionary defines automation as “the technique of making an apparatus, a process or a system operate automatically.”

We define automation as "the creation and application of technology to monitor and control the production and delivery of products and services.” To read the full ISA definition of automation, go here.

Operations Research

Definitions: Operations Research (McGraw-Hill Dictionary of Scientific and Technical Terms)

Math: The mathematical study of systems with input and output from the viewpoint of optimization subject to given constraints.

SciTech: The application of objective and quantitative criteria to decision making previously undertaken by empirical methods.

It can be applied to any situation in which what, when and how the steps of a process are selected, organized and executed. Goals may include safety, quality, speed, simplicity, economics, schedule and perhaps applications-special factors, all with the intent of optimizing the results.

Applying the technique involves understanding the process and the constraints on it. That usually begins with a functional diagram and examining the functionality, and often the cost, of each step. Often, carefully audited trial runs help elevate understanding. Once all is understood, the implementation of the effort is tested, implemented and periodically re-evaluated. Automation engineering guides the development and various process-related skills facilitate development, construction, safety, operation, maintenance and economics.

Many of the skills involved come from application-dependent issues and require people who know how each subject area works. The use of the involved automation technology requires understanding of the science, engineering, installation, maintenance and operation of the system. That often involves a spectrum of skills and people.

Process and Instrumentation Diagrams, P&ID

In process control work, a specific application is developed, discussed, evaluated and built from a “process and instrumentation diagram,” commonly referred to as “the P&ID.” There may be a few P&IDs, starting with an overview and more showing the details of various sections, process, sub-processes and sections within it. It’s important for all involved to understand that kind of information and recognize design requirements. In automation, projects usually begin with a P&ID. Additions, modifications and enhancements are worked in, discussed, evaluated and modified as necessary. It is important to understand them. It is essential that, in automation-related projects everyone involved understands the P&ID concept.

Process automation — a broad, deep, complex and incredibly interesting aspect of manufacturing!

About the Authors

Russell Rhinehart has experience in both the process industry (13 years) and academe (31 years). He is a fellow of ISA and AIChE, and a CONTROL Automation Hall of Fame inductee. He served as president of the American Automatic Control Council and editor-in-chief of ISA Transactions. Now “retired,” Russ is working to disseminate engineering techniques with his web site (www.r3eda.com), short courses, books and monthly articles. His 1968 B.S. in ChE and M.S. in NucE are both from the U. of Maryland. His 1985 Ph.D. in ChE is from North Carolina State U.

Julie F. Smith is the global automation and process control leader for DuPont. She has 35 years of experience in the process industry, having been part of numerous engineering and operations activities across the globe. She has written several papers and columns highlighting the value of modeling and simulation. Julie has a BS in chemical engineering from Rensselaer Polytechnic Institute and an MChE from the University of Delaware.

Erik Cornelsen is an automation and process control engineer at DPS Group, a leading system integrator based in Scotland. With over a decade of experience, Erik has worked and lived in six countries, contributing to diverse industrial sectors, including food and beverages, logistics and construction materials. He holds a Master’s Degree in mechanical engineering from INSA de Lyon (France) and is a Chartered Engineer, a member of the Institution of Mechanical Engineers (UK), and an active member of ISA.

Gregory K. McMillan retired as a Senior Fellow from Solutia Inc in 2002 and retired as a senior principal software engineer in Emerson Process Systems and Solutions simulation R&D in 2023. Greg is an ISA Fellow and the author of more than 200 articles and papers, 100 Q&A posts, 80 blogs, 200 columns and 20 books. He was one of the first inductees into the Control Global Process Automation Hall of Fame in 2001, and received the ISA Lifetime Achievement Award in 2010, ISA Mentor Award in 2020 and ISA Standards Achievement Award in 2023. His LinkedIn profile is: https://www.linkedin.com/in/greg-mcmillan-5b256514/

Ed Farmer completed a BSEE and a Physics Master degree at California State University - Chico. He retired in 2018 after 50 years of electrical and control systems engineering. Much of his work involved oil industry automation projects around the world and application of the LeakNet pipeline leak detection and location system he patented. His publications include three ISA books, short courses, numerous periodical articles and blogs. He is an ISA Fellow and Mentor.