By John Rezabek, Contributing Editor
The other day, the long-discussed and oft-debated topic of "PLC or DCS" came up again on the ISA email listserv "control." The subject has been debated on the list numerous times—maybe someone out there keeps assigning it as a term paper or senior project, and his or her students end up posing the question to our group. The idea of convergence, i.e. PLC ≈ DCS, has its advocates, and the case has grown stronger in recent years. There's even a camp proclaiming PC (Windows box) ≈ PLC ≈ DCS. The arguments are not without merit, but the focus is on functionality: My "Y" can do everything your "X" does. But, good intentions, especially when executed by novices, can have unintended consequences. The hard part is making sure it doesn't do what you don't want.
The hydrogen plant near me, I'm told, is a clone of an earlier plant. The former was built to serve, we'll say, a pancake syrup plant. In its earlier incarnation, it was controlled by PLCs. Now, it has a PLC for pressure swing adsorption (PSA), but uses a scaled-down DCS for the rest. Why the change? The second plant, while nearly identical in capacity and footprint, serves a plant from a "large process industry" culture that was willing to pay for the more costly DCS. The former plant was unmanned; the latter has at least one operator on duty 24/7. A syrup factory outage may be undesirable, but that client's culture was OK with the risk of an occasional outage, which the latter client found untenable. Margins on organic chemicals and the cost of downtime relative to pancake syrup probably is a big part of the equation.
In a large process industry culture that reveres bulletproof reliability, few have much patience for beta testing or "science projects." As a friend and former colleague loved to say, "We race to be second." Let the other guy take his lumps with Serial No. 1. If it pans out, then we'll be close followers. What students and novices sometimes fail to grasp, is that control systems, no matter how well-conceived, are often fraught with unintended consequences and features in early iterations, and may require years of interaction with sophisticated users before all the bugs are out. This important, but less tangible aspect of a control system's heritage is absent from the "PLC ≈ DCS" debates. What user communities have shaped the control systems under consideration? Where can you find your "flock" where your concerns and priorities are understood?
A related question I hear a lot is "Should I choose HART, Profibus or Foundation fieldbus?" A little effort and research could lead to finding your flock. One good place to look is free seminars by trade groups. Profibus Nutzerorganisation (Profibus User Organization) has free seminars worldwide. Go to one, and if you recognize your peers and competitors attending and/or presenting papers, then you may have found your home. However, if all the talk is about soup and syrup and your business is gasoline, you may want to move on.
The Fieldbus Foundation routinely has free seminars too. The foundation is unique in maintaining an end-user council whose members (from companies like Shell, Mitsui, Saudi Aramco, Reliance, Suncor, Apache Corp. and others) aim to influence the development and prioritization of new features. The HART Communications Foundation regularly participates in industry shows, so you can gauge your kinship with fellow users there.
Finally, system-supplier/end-user meetings are an excellent way to find your kin. Again, you can go and listen to what's being promised for upcoming releases. Do they sound like what you've been wishing for your plant? Do you find the papers presented are about subjects you or your management find near and dear?
Of course, you can argue that running with the herd means your results will be—at best—undistinguished, but do you really want to be a pioneer? If you and your company agreed to such an adventure, I'd be interested in how it worked out. Write to me at firstname.lastname@example.org.