Closed-Loop Tuning of Controllers ◾  81
without further intervention. In other words, the liquid level
response is integrating and is not self-regulating.
9.1 Trial-and-Error Tuning of Control Loops
Many process control loops are tuned by trial and error while
the controller is in automatic output mode. Everyone who has
ever tuned a process control loop has used trial and error at
one time or another. Some people have learned the art of tuning process control loops by trial and error and have become
highly proficient by learning from their own experience or
from the experience of a mentor or a network of practitioners.
Trial-and-error tuning consists of changing one or more of
the PID (proportional, integral, and derivative) actions in the
controller by some amount and observing the response of the
control loop to see if it changes for the better or worse. The
proportional gain may be changed by 10% or 20% or may be
doubled or cut in half. The integral and derivative may also be
changed by the same amount or turned off.
One good point about trial and error is that the controller is in automatic, and the controller settings can be returned
back to the initial values if they are recorded somewhere or if
they can be remembered. Another good point about trial-anderror tuning is that the usual equipment existing in a control
room for recording a process variable response with time can
be used without any additional tools or computer programs.
One concern about trial-and-error tuning is that a beginner
may not recognize whether the change in process variable
response is better or worse than the response was with the
initial values. Another concern is that trial-and-error experimentation may involve a long learning curve. Learning how to
tune a control loop may involve many setpoint changes and
process upsets. Also, trial-and-error tuning can sometimes lead
to mindless tweaking of the controller settings depending on
the individual and the plant circumstances.