A fluid catalytic cracker (FCC) utilizes an unforgiving piece of hardware between the reactor—where catalyst mixes with gas oil and “cracks” it to lighter fuels—and the regenerator, where “coke” (the heavy remains of said reaction) is burned off the same catalyst. FCC catalyst is regenerated in a continuous cycle, and the spent catalyst must always flow reactor-to-regenerator, lest a dreaded “reversal” take place.
The crucial equipment preventing this upheaval is a relatively large, customized valve called a slide valve, which has doors like a patio that slide open or closed to modulate the spent catalyst and retain enough to “seal” against backward flow. An operator seeing a waning differential pressure between the upstream and downstream sides might use his control system to pinch off on the flow of catalyst.
That we humans have innate biases is a focus—some might argue an obsession—of the current zeitgeist. But one’s tendency to seek a world unchallenged by strangers or uncertainty is even more pervasive, so argues psychology professor Daniel Simons and co-author Christopher Chabris in their new book, “Nobody’s Fool: Why We Get Taken In and What We Can Do About It.” While they focus on how scammers, con-men and other nefarious individuals outwit, deceive and defraud even the savviest among us, this vulnerability vexes our profession as well.
“We don’t intuitively appreciate how much variability should exist in numbers that describe human experiences, decisions and actions. We’re seduced by the simplicity of smoothness,” the authors contend. When measurement systems, controls and automation function well, the humans in charge acquire a similar expectation. “The slide valve should go to the position I want,” thinks the less-seasoned FCC operator, “it has the last few times when I was pushed up to work the board”.
Like the commercial airline or cruise ship pilot, the desire for a voyage unchallenged by perplexing happenings or upsets is likewise a goal for process plant operators on their shift. One might be predisposed to dismiss an uncommon alarm as an instrument malfunction or “noise.” If the differential pressure across the slide valve doesn’t react as expected when the valve is moved, one could surmise it’s an instrument malfunction, or a loss of purge on process connections prone to plugging up with solids. Even in a highly managed system, a given operator might “shelve” or suppress an alarm they deem to be irrelevant.
In recent decades, we’ve been compelled by changing workforce demographics and international standards to ensure all alarms are meaningful. Graphics should be grayscale and all color should be muted, only being used to indicate gross deviations from “normal.” A process plant can pay consultants, who will gather their P&IDs (piping and instrumentation diagrams), go away, and return with “optimized” alarm settings and priorities. Depending on the complexity of your plant, one might find this approach a little concerning. Even a team of in-plant process, safety and instrument specialists can argue about the consequences and urgency of a given abnormal condition. Alarms deemed to be noise—requiring no operator action—are justifiably removed from the configuration or relegated to “alert” or log-only status.
Noise in the realm of process measurement and control is what’s judged to be variation that has no information, meaning or value. Who is making that judgement? Usually, it’s the controls professional, who knows where to click the mouse to tune a low-pass filter. But the instrument specialist also knows there can be information in noise, as some contemporary pressure transmitters can monitor noise to alert them to plugged or plugging impulse lines (the tubing that connects an instrument to the process.) Experienced operators derive insight from noise: a pump or its motor is making an unfamiliar whine, or the routine thumpity-thump of a reciprocating compressor has acquired an unfamiliar rhythm.
Should we be more concerned that we’re removing noise that might convey an insight, something that could avert a calamity? Will console operators be so entranced by insipid grayscale graphics that potentially meaningful variation will slip by them unnoticed? Or will conditional logic suppress an alarm that could have averted product contamination? The FCC operator should treat the absence of a response with some concern, according to Simons and Chabris. “Our habit should be to treat the absence of noise as a warning to dig deeper.”