Environmental Engineering Reference
In-Depth Information
PID controller with a higher intensity pulse amplitude modulation (PAM) actua-
tion that uses the same amount of electrical energy. The magnitude and duty cycle
of the PAM signal are updated at each time step based on the magnitude of the
continuous PID signal to match the consumed energy over one signal period.
8.10.1 Distributed Closed-Loop Control Experiments
For the results described in this section, each heater receives information from the
two adjacent optical ice sensors, and the heater stays on only if both of the adjacent
ice sensors detect ice. Ultimately, we can use information from all four sur-
rounding optical ice sensors adjacent to each resistor if necessary; however, this
makes the closed-loop controller slower by adding computations required for
signal processing. The initial control systems discussed here only use ice existence
information. Further experimental investigation is needed for a characterization of
the thermodynamic characteristics of different types of ice. Ice type and thickness
information provided by the OFDR ice sensing system can be used in future
versions of the control system design.
The control gains for each heater at the leading edge are calculated based on the
temperature information from the adjacent temperature sensors. For the initial closed-
loop test of de-icing, since there is only one temperature sensor for each ''column'' of
thermal resistors in the chord-wise direction of the blade, the same voltage was
applied for each column of resistors. Figure
8.24
shows different stages of ice for-
mation on the blade before the active de-icing control system was switched on in the
icing chamber. Figure
8.24
a shows rime and glaze ice formed on the blade in the icing
chamber. Glaze ice was created in the chamber by directly pumping humidity to the
blade surface under the condition of no wind. Uniform rime ice is created on the blade
when there is nonzero wind velocity in the icing chamber (from a fan).
A PID control has been implemented to each column of resistors in the network.
The applied voltage to resistor j is:
2
4
3
5
K
p
e
j
ð
t
Þþ
K
i
Z
t
d
dt
e
j
ð
t
Þ
v
j
¼
K
g
e
j
ð
s
Þ
ds
þ
K
d
ð
8
:
6
Þ
0
where K
g
is the gain of the op-amp circuit, e
j
(t) is the error between the desired
temperature (T
d
) and current temperature (T
a
j
ð
t
Þ
) on the blade for channel j at time
t, K
p
is the proportional gain, K
i
is the integral gain, and K
d
is the derivative gain.
Figure
8.25
shows the time history of the voltage and temperature for channel 1
(see Fig.
8.11
) from the beginning to the end of the activated PID control for de-
icing using distributed resistors, optical ice sensors, and temperature sensors,
where the maximum voltage indicates that approximately half of the maximum
power capability of the individual kapton resistor has been used. Similar behavior
has been captured with the remaining resistors. It takes approximately 3 min until
