Professional Documents
Culture Documents
Lab 1 Final
Lab 1 Final
No. 1 G9
Laboratory No. 1 G9
Objectives To demonstrate the characteristic of PB (Proportional Band) + I (Integral Action) + D (Derivative Action) on a pressure process control loop. Results: Experiment 1 PB = 200 % I = 3 sec D=0 With 5 second change to 40 L/min
Figure
1:
Start
of
Experiment
1
Laboratory No. 1 G9
Laboratory No. 1 G9
Laboratory
No.
1
G9
Figure
7:
PB Changed to 10
Discussion
To
make
it
easier,
each
figure
will
be
discussed
in
order,
rather
than
in
one
bunch.
Experiment
1:
Figure
1:
Figure
1
shows
the
settings
at
which
the
experiment
started
which
were
PB
=
200%,
I
=
3
sec
and
D
=
0.
The
200%
PB
setting
is
rather
high,
causing
higher
controller
output
and
hence
a
larger
error
when
correcting
the
disturbance
of
changing
the
valve
to
40L/min.
There
would
be
an
even
larger
error
if
not
for
the
I
=
3
sec
setting
which
corrects
the
error
and
produces
the
sine
wave
like
shape
shown
in
the
graph.
Figure
2:
In
figure
2,
the
same
settings
are
kept
but
an
additional
disturbance
is
added
and
a
step
change
to
75%
is
added.
The
presence
of
the
disturbance
yields
the
same
results
as
that
of
figure
1
except
that
it
is
in
opposite
direction
(goes
upward
first
then
goes
downward).
This
is
due
to
coincidence
that
the
graph
was
oscillating
in
the
upward
direction
when
the
disturbance
was
applied.
When
the
step
change
was
applied,
the
system
was
slow
to
adapt
to
the
change
but
corrected
itself
quickly
in
two
major
oscillations.
15485634 NICHOLAS ANG
Laboratory No. 1 G9
Figure 3: In figure 3, the derivative time setting is changed from 0 to 5, the system is brought back to a set point of 75% and a disturbance is introduced as before. The system takes the same time to recover from the set point change but corrects the disturbance in a quicker time than in the two previous figures. This is likely due to the derivative effect correcting the lag caused by the integral time function. Figure 4: In figure 4, the set point is changed and the derivative time is changed to 20. The systems response to the step change is starting to get worse which suggests that the derivative time change is affecting the systems response to step changes in a negative way. Figure 5: In figure 5, a disturbance and step change occurs with a derivative setting of 20. As before similar results occur for the step change but the reaction time and magnitude correction for the disturbance is much quicker and smaller than before. This indicates that a higher derivative time setting allows quicker and better responses to disturbances. Experiment 2 Figure 6: In figure 6 a PB setting of 100 and a large I setting of 200 s is used. Although the PB setting is lower than before, there is still a large error and since the I value is exceedingly large, there is nothing to correct the PB error and keep it oscillating so it just decreases linearly at a slow rate. Figure 7: In this figure, PB is changed to 10, which allows the error to decrease, as there is lower controller output and higher gain. However, gain should not be too high as the response may become exceedingly oscillatory and unstable. Again the system is correcting itself closer and closer to the set point rather than using a sine wave approach as the I value is very high. Figure 8: In this figure, PB is changed to 6 and 5.5 which results in the system getting closer to the setpoint. However, as mentioned before, the system now has higher gain and the curve starts to become more oscillatory and unstable. Hence, the system should have a good compromise between proportional band and gain such that the value chosen is of low error and stable, with as little oscillation as possible. Conclusion Based on the results, a larger proportional band will result in a larger error from the set point due to high controller output. This causes the system to be further from the setpoint and responds slower to disturbances such as set point and flow rate changes. The PB value should also not be too small as this results in high gain, causing unstable oscillatory behaviour. The integral time is also critical to increase the response time of the system and to correct the error using a sine curve and sum of error approach rather than letting the system slowly approach the set point as seen at high I values. Lastly, the derivative time has a small effect in allowing the system to respond to disturbances faster as it smooths out the lag caused by the integral time function.