Start-up can be a tense time as licensors, process engineers and project managers anxiously await the verdict on their design and implementation. It's the worst time for the measurement and control infrastructure to be any kind of distraction or sideshow. We want to be perfect, and when we are, they don't even know we're there, rather like the background music in a drama.
Consider the term HMI, or human-machine interface. What exactly are these machines with which our operators interface? They interact with our "machine" for sure—these days almost certainly a using a Windows box with some real-time graphics capabilities and connectivity to the other machines that constitute our DCS, including field devices. But the real operator interaction is with the process.
Also Read: Improved Fieldbus Operations with Advanced Diagnostics
I suppose you could say a hydrocracker or reformer is a machine, and such processes have plenty of machines such as compressors, pumps and valves routing process fluids through vessels and pipelines. But our operator's ultimate interface is with a process, not a machine. If operators don't manipulate the process to produce saleable product—hopefully making something more useful and valuable out of something less useful and valuable—then their mission is a failure. The more all the machines just do their job, the more effective the operator can be at keeping the plant safe, reliable and productive. A bit paradoxically, the controls specialist strives to deliver measurements and automation transparently, and stay out of the limelight.
Our profession supplies the process' "nervous system," if you will, providing most of the operator's sense of its condition. They might hear a little fieldbus music playing when they wonder, "Did the control valve move when I asked it to?" Digitally integrated control valves report their position feedback in real time, and if you've taken pains to show this on your HMI (recommended), they can see nearly immediately; "I've clicked the up-arrow twice and nothing happened," which in turn validates why a flow hasn't changed.
More music might play for them when a flow exceeds its calibrated full scale, but keeps on indicating a valid flow, instead of saturating at 21.7 mA. They may not hear it at first, but fieldbus brings a whole distinctive layer of data validation. Every device plays a relentless refrain of "I'm here. I'm happy. I'm communicating, and my measured value Y (as a floating point value in engineering units) is valid at timestamp X ±1 millisecond." An array of conditions that could indicate a measurement is questionable or "bad" are monitored and updated with similar rigor. Where fieldbus devices provide an interface to the process, each instance is imbued with a layer of diagnostics and intelligence designed to provide a more truthful and reliable version of reality and to instantly inform us when there's a fault.
Operators may not seem to care much about the nuances of how our "machines" deliver information. But when they're spending a tense 12 hours guiding the process through the straits of some challenge like extreme weather or an unanticipated feed change, relying on "one version of the truth" can mean the difference between safe harbors and following a siren-song into the rocks. Operators want to know their instruments aren't lying to them. That's the tune they like.
