Figure 6 – PITOPS process dynamics identification
As it can be seen from Figure 6, identified process model (shown
with blue curve ) in PITOPS shows accurate model behavior which
perfectly follows the raw reactor temperature data (shown in red )
containing slow dynamics with a dead time of 4.5 minutes, process gain
of 2.5 and a process time constant of 25.7 minutes. Considering the slow
process dynamics if this process had to be identified using old-fashion
methods, it would take much time of process engineer compare to PITOPS.
After process identification, the next step would be to find optimum PID
tuning parameters. PITOPS uses precise and accurate algorithm for
identifying process model and based on the process needs it tunes PID
tuning parameters, as described earlier.
In this example, PID values computed from old-fashion PID tuning methods
will be compared with PITOPS for a SP change and PV overshoot. Since
this is a temperature controller, it needs to have very tight and stable
control action without and oscillations and PV overshoot. To easily
compare the PID tuning parameters computed from both methods, trends of
both methods will be shown one after another. IMC and Lambda PID tuning
standard will produce similar behavior where calculated PID tuning
parameters gives very slow and sluggish control action without PV
overshoot, as shown in Figure 7. ZN and CC PID tuning standard will
produce somehow satisfactory control yet it will be very aggressive,
oscillatory and with PV overshoot, as shown in Figure 8. Last shown PID
tuning method is PITOPS. In house PID tuning standard will produce
tight, stable and fast control action behavior without PV overshoot, as
shown in Figure 9.