Page 56 - Practical Control Engineering a Guide for Engineers, Managers, and Practitioners

Introduction to Developing Control Algorith11s 31

projects were not fun and did not use my skills properly. I proceeded
to avoid them like the plague.
Surprise Sub
After several years of avoiding projects like the one described (and
paying for it with an unimpressive rate of promotion), I got involved
with a junior engineer, whom I will call Bill, as his mentor. Bill was
formally trained as an electrical engineer and had morphed into a
software engineer primarily engaged in programming the minicomputer-
based data-acquisition and control systems. Without much prior
experience, Bill was charged with the instrumentation and control
responsibility for another new process. By this time our group had
started installing analog output cards and using the minicomputer
to close control loops and replace stand-alone analog controllers. I,
as a semisenior, semiexperienced associate, was supposed to be
available to Bill for questions and counsel.
Bill invited me over to the site once, well into the installation, and
showed me around. Since he was programming the computer and
since I had minimal knowledge of software systems, the tour was
nothing more than a long cup of coffee. I really gained no idea of how
he was doing, except that everything looked OK and Bill was cool. I
heard no more from him. The process started up without incident
until a few days into the operation when a major flaw in the process
design was discovered and the boys who owned the process went
back to the drawing boards for several months.
When the redesign was finished, three things were apparent. First,
nothing topologically had changed with the process so all of Bill's work
would supposedly still be applicable. Second, Bill's 5-year company
anniversary had arrived and he had planned to spend his extra 3 weeks
of vacation (5 weeks total) in Europe with a bunch of like-minded
youngsters. So, even though the vacation coincided with the restart of
the process he still was allowed to go because of all the trip commit-
ments he had made. Third, because Bill's stuff had apparently worked
during the first abortive bring-in, there was no need for his presence
and Uncle Dave (moi) could simply drop by on the start up day and be
available for questions, should there be any (unlikely).
This situation bothered me. I had no idea what Bill had done. So,
even though the first process bring-in had been uneventful from an
instrumentation point of view, I dropped by during the week before
the second bring-in and proceeded to dig into the software. I found
that there were several proportional-integral control loops, some con-
nected in a cascade configuration (things we will talk about later).
To check things out I did the obvious. I put test signals into the
computer and monitored the data-logging CRTs to see what the con-
trol loops did. To make another long story short, there was a plethora
of errors, some serious (the cascade loops) and some minor (derived
variables based on data-logging inputs). I could correct the FORTRAN