Finally, a WiFi-enabled smart toaster. Where was this gadget back when I could eat a full loaf of cinnamon-raisin bread toast without gaining a pound? My old-school toaster dates from the 1980s — older than most of the instruments where I work — and does indeed make toast, and perfectly good toast if you're willing to babysit and pop it a couple times to check its progress. Since it still functions on this rudimentary level, I risk marital acrimony if I dare to replace it (or the microwave from decades ago that dims the lights when it runs) because my spouse is decidedly anti-tech, and prefers beach vacations to slick new appliances.
Indeed, it’s a struggle to justify a new one if the old clunker isn’t a fire hazard or completely FUBAR; and even Oprah isn’t pounding the toast like she used to. So, smart thermostats and washers and irrigation and refrigerators will remain at Lowe’s or on Amazon for some other (probably single) nerd to enjoy.
But what would truly be a huge benefit if it had the intelligence and connected-ness to butter our bread, or emblazon toast with today’s forecast?
In the process plant, there’s more at stake than burnt toast. Safety and environmental mishaps are egregious, and unplanned interruptions to production have far-reaching consequences. Intuitively, we know our process equipment and the process itself has “intelligence” to share, if only we knew where to listen, and what it was saying.
Our experienced operators are always listening to process equipment — pumps, motors, compressors, reactors and so on — for clues to the overall health of the equipment and the process. Even through normal hearing protection, subtle changes in characteristic sounds can spur some additional investigation. But like a raw acoustic sensor, a repeatable correlation with an abnormal condition requires analysis and validation by a human subject matter expert. And the action we take — if we choose to take any — hinges on the needs of the business and the commercial environment.
Some of our dreamier engineers imagine a microphone over the whole complex that could magically detect and interpret changes in the plant’s acoustic signature. Today, you can get your plant a gadget to listen for leaks. For instance, the Rosemount GDU-Incus can hover UFO-like over an area and alert you to changes in the amplitude of ultrasonic frequencies it monitors, which can be an early warning of an otherwise undetected leak. It has an effective range of about 20 meters, so you might need a zip line like we see for cameras on Monday Night Football. If you have a process running with toxic, flammable or asphyxiating gases at high pressure, this could be a piece of intelligence worth the investment.
But leaks to atmosphere, while concerning, are only one malfunction we wish we had ears to hear. If an Incus can hear leaks from afar, why don’t we integrate internal leak detection (or make it a specifiable option) for on-off valves and relief valves? Acoustic sensors for leak detection have been around for years, but the makers of tight-shutoff valves and safety-relief valves haven’t been quick to integrate them.
With a fieldbus valve positioner, or a HART-capable positioner connected to a multiplexor or a HART-enabled channel, you can create a DIY smart diagnostic, if you have a corresponding flow. For a given pressure drop, the valve position (or VFD speed) should correspond, within some noise band, to a flow. You can derive this relationship in an Excel spreadsheet.
It’s been many years since Valtek (now a division of Flowserve) introduced what might be the granddaddy of smart positioners, the Starpac, a unique assembly that integrated process taps and measurements for the connected valve’s upstream and downstream pressure, process temperature, and a calculated flow. In a modern plant, all the same data might even be on the same fieldbus. The “intelligence” may already be there to alert you that a flow isn’t going where it’s supposed to go, or that something in the process has changed.
While mashing a great amalgamation of data may yield some interesting patterns and detect anomalies, instead, consider what data you were lacking, or failed to recognize, the last time productivity got toasted. Once the toast is burnt, there’s no amount of jam or jelly that will render it edible.
Homepage image courtesy of Idea go at FreeDigitalPhotos.net