A couple decades ago, I recall an incentive at a company to streamline the controls' upgrade process. The company had a number of likely panel-to-DCS projects at units ranging from blenders and wastewater treatment to alkylation plants and reformers. The company-wide global standard for the DCS and field devices was fixed. There were few choices to be made apart from "how many of X do we need." OSHA 1910 had been released, so to ensure compliance, the company had shiny, new up-to-date piping and instrumentation diagrams (P&IDs). The mature processes that had been cranking out product since the 1960s or earlier could be considered "saturated." Any variable the engineers had ever imagined controlling already had a transmitter and/or a valve.
In the zeal to optimize costs, another concept came into play called "minimum adequate documentation" or MAD, if you will. Installers were always coming back with cleverer ways to route conduit. Why pay an EPC (engineering, procurement, construction) contractor for a conduit plan? Some jobs were tried with location plans only, and supplemented with computer-generated termination schedules. Maybe the field junction boxes were sited, but these were somewhat fluid. Could a creative and specifically qualified installer be "delegated" the task of field-conduit routing?
Even in those days of MAD-ness, one essential document was still the loop drawing. By the 1990s, these were being populated by the same database engines that created the termination schedules. Lacking, though, were the accoutrements of prior-day loops. The first DCS job I worked on in the early 1980s had essentially hand-drawn loop diagrams, and each one had a little P&ID excerpt on it! There was a wealth of information, aiming to document everything a tech or troubleshooter would need to track down an issue without creating any more. But adding this meant hours—and not cheap hours—of work. And, in time, the client's project managers were promised efficiencies owing to their EPC's cleverly integrated, computerized database. Loop drawings were fast becoming mere "wiring" diagrams.
Recently, the question of whether loop diagrams were needed for a fieldbus project came up on the ISA "CONTROLS" list-serve (www.isa-online.org/cgi-bin/wa.exe?INDEX). Walter Driedger of Colt Engineering set us all straight. "According to ISA 5.4," Driedger wrote, "a loop drawing shows an entire control loop from the transmitter to the controller to the valve to the pipe and back to the transmitter. This is basic to the understanding of a control loop." It's not a "current" loop we're aiming to depict. Fieldbus segment diagrams show how devices are connected to a segment, but not how they are integrated for control.
Should we go back to showing all loops on loop drawings, regardless of the technology? In the days of devices digitally integrated through Profibus, Foundation fieldbus and wireless, do we have a prayer of getting all the information on an 11 x 17 loop sheet, let alone keeping it up-to-date?
Some are finding it easier to make their segment diagrams on DCS graphics, and use the DCS database to self-document all the ways a given variable is interconnected. Chuck Carter, of the Fieldbus Center at Lee College (www.lee.edu), insists all his trainees learn how to navigate the "live" DCS configuration for this very purpose. It's simply too risky to rely on paper, no matter how professionally and meticulously it's composed and maintained.
We're lucky to keep P&IDs and loop sheets up-to-date, let alone add annotation for all the references where the variables might interact. There's a path to getting the needed information from our control systems, but our culture needs to accommodate it (Instrument techs need to touch the DCS! Egad!) And, our suppliers need to help us get to it with maximum ease and minimal risk to the plant. Loop diagrams: It may be time for them to depart the physical world and pass to the virtual one.
