Technology and People Connections: Keys to the Future
Participating in this Control Engineering 50th anniversary interview were Jerry Gipson (firstname.lastname@example.org), director of Engineering Technology, and Eric C. Cosman (email@example.com), automation architect, from The Dow Chemical Co., Midland, MI. In your experience, what do you consider to be the top three advances made in control and automation over the past 50 years? Which one of these do you th...
Participating in this Control Engineering 50th anniversary interview were Jerry Gipson ( firstname.lastname@example.org ), director of Engineering Technology, and Eric C. Cosman ( email@example.com ), automation architect, from The Dow Chemical Co., Midland, MI.
In your experience, what do you consider to be the top three advances made in control and automation over the past 50 years? Which one of these do you think is the most important and why?
Gipson : Of course, I don't have the full perspective of 50 years. But one of the things that pops out is the emergence of distributed control systems and the associated architectures. This speaks to significant changes in how we view the design of control and automation systems and applications. As we continue to apply distributed architectures, the possibilities are endless. This has to be one of the top advances.
Another one—of course it's peripherally related, but very much integrated with what we're doing in control and automation—is the emergence of sophisticated IT networks and infrastructure, which allows us to connect and share data across the plant and the enterprise. This represents a real breakthrough that has come to the fore in the last 20 years.
A third area, and one that I believe Dow Chemical has a key perspective on, perhaps because we've taken a leadership role here, is the very tight integration of basic process control and safety systems technology. That has to be one of the top three.
If I had to choose one as most important, I would choose distributed control technologies because of the core focus on architecture and on allowing continuous improvement, as we go forward, on how we design and implement these systems in total.
Cosman : Compared to when I first studied process control almost 30 years ago, the most obvious change was the move from analog to digital. That fundamental shift was probably the biggest change, and I look at that as the beginning of a whole continuum of changes that have brought, almost year by year, a closer alignment of the disciplines of control and information technologies. Each step has brought these two groups closer and closer together until at this point we are almost 'joined at the hip.'
Another shift that I see—and this is probably somewhat 'Dow-centric'—is the tight connection between chemical process engineering and process control. In some industries, process control has been a discipline that has been sort of added on. The Dow philosophy for probably 30 years or more has been that control engineers are process engineers first. To understand how to control a process, you must know how to design it. To me, that's something that sets the process industries apart from other industries that use process control. As you bring these sophisticated information technologies into areas like modeling and advanced control, clearly you'd better be very intimately involved in the process and understand it thoroughly. We're not in the loop-based control business anymore. We haven't been for a long time.
What do you think will be among the next significant advances in the controls, automation, and instrumentation arena in the next 5 to 10 years, and––of course this is crystal-ball gazing––50 years in the future?
Cosman : The first thing that comes to my mind is the massive distribution of intelligent devices. Everything, down to the smallest discrete pieces, has some degree of intelligence built in. It could be passive intelligence like an RFID tag or it could be active intelligence. There is not now—but the trend toward this will continue, I believe—a common brain. That will lead to huge challenges to designing a true distributed control system of any kind. It is challenging today and will become even more challenging in the future. To me, that's one thing I see happening.
A sidebar to that is that everything is networked. Everything will have a network address of some kind. The most innocuous, benign things that you can possibly think of will not only be intelligent, they're going to be able to talk about it. We will be able to reach dynamics of processes that we have been unable to reach up until now. The volume of information—the din of the data coming in––is going to be overpowering.
Gipson : Number one on my list of advances is probably similar to Eric's. We're going to see the process control and automation systems and applications functionalities spread throughout the process. We're going to see them distributed even on networks in dramatically different ways than we do today. We can already see where this is going because some of these architecture advances and equipment capabilities are already starting to appear on the market today and are being adopted by some users.
I think we'll also continue to move toward dramatic improvements in reliability and robustness. Whereas we sometimes have to use redundancy in our automation systems, we'll see a higher device reliability and overall an ability to design integrated systems that are more intrinsically operable and robust.
Also important will be bringing more of our engineering knowledge to—and making it available within—the controls and automation framework. We talk about advanced approaches by which we collect and analyze data to develop a process. We do engineering and modeling to characterize the process. But my view is that within Dow and within industry today we use relatively little of that engineering knowledge in the design and operation of our automation systems.
Going forward, we will apply that engineering knowledge and that will result in two things: First will be a better overall design as well as the design of automation systems being incorporated into the design of the overall process. Second, we'll also see better operation of the process to higher profit levels and higher environmental health and safety standards, where we're always driving for perfection. I think that is absolutely critical: engineering knowledge will allow us to take our processes to the ultimate.
What will be the biggest challenges facing manufacturing engineers in the next decade with regard to plant infrastructure, technology, and the business of manufacturing in general?
Gipson : The term manufacturing engineer probably means different things as you go from company to company. At Dow, the manufacturing engineer is in charge of running the plant. He or she is not in charge of developing the technology or improving the technology. So using that definition, the biggest challenge for the manufacturing engineer is to accept from the people who develop the technology and design the processes something that is seamless in terms of a well-designed, robustly operating plant, knowledge transfer, and a definition of consistent processes by which the manufacturing engineer is going to do his/her job.
The challenge is that in areas where we're not doing a good job of giving manufacturing the guidance and knowledge necessary to do the best job of running the plant, there's going to be additional work on the part of the manufacturing engineers to overcome those gaps. I see it as dealing with systemic issues and gaps in performance.
Cosman : Historically, manufacturing engineers have been focused on responding to dynamics of the plant, the process, and the physical aspects of it. That has been their challenge. Things change because things break, because the process isn't operating the way it is supposed to. The fact is: we're getting better at that. As we get those processes more under control, and operating more reliably, we're designing plants to run uninterrupted for 5 to 10 years. What that means is that the manufacturing engineer is going to have to switch gears and change his thinking in the sense that the disruptions to which he's going to have to respond in the future will not be process-oriented; they will be business-oriented.
We are at the point where business changes, up to and including mergers, acquisitions, and divestitures, happen more frequently than physical process changes. You may design a plant to last five years and find it changes hands three times in that period. As we see the amount of business change in the industry, with consolidation and repositioning of product lines, it is realistic to think that you could have a plant running for a five- or ten-year period and have three different owners. Now, a manufacturing engineer is responding to a whole different set of dynamics than in the past.
Manufacturing engineers are much more plugged into what the business is doing, what the business conditions are, what business decisions are being made and why, and even participating in those decisions. It's not enough to just run a large, complex physical process anymore. Now they're running an even larger and more complex process—but a business process instead of a physical process. This change clearly demands people who are very flexible, agile, willing to learn constantly, and willing to change on a moment's notice.
Looking back over your career, have there been any unexpected surprises for you in terms of new technology, business developments, or industry direction?
Cosman : The spread and pervasiveness of network technology, particularly the interconnected, Internet kind of network technology over the past 10 years, happened faster than anyone realized or expected. People woke up one morning and found that they were much more connected than they were prepared for.
On the other hand, we've been talking for a long time about the pervasiveness, adoption, and in some cases intrusion of what's generally called commercial off-the-shelf technology into the control space. But in my opinion, this has not happened as fast as I would have expected. The reason, I think, is the inherent difficulty of replacing a very large installed base. If you look at some of the statistics from organizations like the ARC Advisory Group, many companies are not ready for the current state of technology. Nobody is going to go out and spend the money to rip out existing systems. So that tells you that the spread of COTS [commercial off-the-shelf] technologies is not going to be limited by the availability of the technology, but rather by the organization's inherent ability to absorb it.
Gipson : I'm not overly surprised at the developments that have occurred, but I'm pleasantly surprised at how quickly some of them have occurred. For example, in the early '80s, when we talked about networks and network technology and where it was going, and about the emergence of desktop applications, it sounded like these would take quite a long time to achieve, although we could definitely see them out there. Now, just 20 years later, it is amazing what has been accomplished.
One outcome of technology advances that has been somewhat unexpected is cybersecurity. I'm referring to the magnitude with which we are being challenged by new and very creative forms of attack, and also the level at which businesses have to invest and industry has to respond with new technologies, even totally new businesses, around this whole area of cybersecurity.
Another thing—and one I think many people miss when looking forward from years ago—is the level of desktop functionality we've achieved: what can be deployed on the desktop, the computing power that is available on the desktop. That clearly has been a good story, but one that was not foreseen years ago by many.
To advance their careers within manufacturing, what should engineers focus on in the next 10 years?
Gipson : The basic answer is fairly simple: keep your technical knowledge current. Most people think in terms of their own disciplines: chemical engineers think in terms of chemical engineering knowledge or tools associated with chemical engineering, but it's also understanding how our discipline, whatever it is, operates within industry and within a specific company. It's not just a raw technical skill, but applying that technical skill in the company and industry.
Partnership skills are also very important. We at Dow spend a lot of time talking about partnership skills, and partnership means understanding how we are going to relate to other people, how we're going to work together, what we are going accomplish together, and having a clear understanding of how those outcomes are going to be accomplished. This is critical for everyone, but sometimes a challenge for engineers because we think so much about our technology and how to accomplish technical outcomes.
The last point would be to focus on human relations skills—good communication, respect and responsibility, a whole host of things on the soft side of accomplishing our mission both as professionals and as human beings.
Cosman : My knee-jerk answer to this question is: as many things as possible! Engineers need to take a broad view. Get involved and be engaged in and learn about as many things as possible. My advice to a young engineer today would be—and you don't necessarily have change jobs to accomplish this––you have to be aware of, be conscious of, and be open to different ways of doing the things you are already doing and getting involved in different aspects of the same kinds of jobs.
Can you provide an example based on your experiences where paying attention to processes and culture helped the automation implementation (and company)? Are companies looking at processes and culture as part of major technology implementations? Should they be?
Cosman : This question is tailor-made for what we are doing at Dow right now. We are in the middle of this and have, over the past couple of years, been heavily engaged in bringing fundamentally new controls technology into the company. We recognized early in the game that doing this successfully involved people, organization, and cultural problems more than technology issues. The reason for this is because of the COTS technology we are installing simultaneously—a control system, an IT infrastructure, and an information system—all within the same product offering. Of these three, we, the controls organization, are only good at one of these, reasonably good at the second one, and not at all good at the third. What that meant was partnership. Partnership not just between us and our IT organization, but us, our IT organization, our automation vendors, and our IT service providers. We get into project meetings where there are representatives from four or five companies involved. If that doesn't speak to the need for culture and understanding different points of view, then I don't know what does.
Now, whether or not other companies are looking at these issues, I would say it's all over the map. I've worked with a lot of companies in the cybersecurity area. It is always dangerous to over-generalize, but the large, integrated, more sophisticated companies have seen this and, to one level or another, are responding to it. But there are those at the other extreme of the spectrum who have not even seen it yet. A large majority are somewhere in the middle: they know they need to do it but are struggling with the how. We don't have all the answers, but we are certainly ahead of the game.
Should companies be doing this? Absolutely!
Gipson : Eric gave, I believe, the best example in our current experience here with our new control system implementation. However, we can also look at an earlier example. About 20 years ago, we developed an internal technology—both the hardware and the applications for process control. We recognized there that we could not just work on the technology. We dealt not only with the processes by which we develop, apply, and standardize the technology, but also with how we use them in the practice of running the plant. That inevitably crosses over into culture. At Dow, we standardize approaches as much as possible. If you go to a Dow plant anywhere in the world, you will see and hear things that are common across the globe. This whole philosophy of combining technology with processes and culture has yielded significant benefits, and we feel strongly about carrying this concept forward.
'Should companies be looking at process and culture?'—that is really the important part. We have no choice. We have collective interests in industry that drive us to bring together all these elements because together they constitute total performance. If you talk about our common vested interest in achieving perfection, plus the desire to be more productive and competitive in the operation of our plants and our businesses, all of this drives us to bring together this total performance package.
Gipson is involved in process automation, process engineering, and process safety engineering. A Dow employee for the past 24 years, he has worked in R&D and manufacturing, and has had experience developing and scaling up processes and background in modeling as well as a variety of other engineering activities. He says of his role at Dow: 'We have key responsibility for running our plants safely, productively, and cost-effectively.'
Cosman's responsibilities include system architecture definition, technology management, and integration planning for manufacturing systems. He also represents Dow on various standards committees and industry focus groups. Having worked for Dow since 1976 in Canada and the United States, with assignments in process engineering, process systems software development, telecommunications, automation architecture, and consulting, Cosman is a chemical engineer and a member of ISA and CIDX (Chemical Industry Data Exchange). He is active in a variety of user groups and organizations, and has presented and published papers on various topics related to the management and development of information systems for process manufacturing.