|
|
FAQ (Frequently Asked Questions)
Category:
CDS 110b Winter 2005
Identifiers: H10 H11 H14 H15 H9 L14 midterm review
-
Are there matlab m-files to convert controllers to work with the dfan simulink block?
Submitted by: macmardg
Submitted on: January 4, 2005
Identifier:
H9
Yes - download the file runsim.m off of the web site and type help runsim to see instructions; this uses the controllers for x and theta axes, assembles the full controller, calls simulink, and plots and returns the results. Let me know if you have a problem with it that isn't answered in the instructions.
[Back to Top]
-
How does the sign of the plant affect the control design?
Submitted by: macmardg
Submitted on: January 5, 2005
Identifier:
H9
The plant for the x-axis is negative (that is, instead of requiring negative feedback to stabilize it, it requires positive feedback). Note that PC=(-P)*(-C), so that if one designs a compensator for -P, as usual, then changing the sign of the compensator will stabilize P.
Also note that the need to change the sign of the compensator will affect plots such as the root locus. The rule about the locus existing to the left of an odd number of poles/zeroes applies for negative feedback. Changing the sign of the feedback will mean that the root locus will exist to the right of an odd number of poles and zeroes.
[Back to Top]
-
There's a typo in 1(c): "rad/sec" should be "rad"
Submitted by: macmardg
Submitted on: January 5, 2005
Identifier:
H9
You're right. Oops. Also, the text isn't clear that question 1 is to design control for theta, and question 2 is to design the (outer loop) control for x
[Back to Top]
-
I dont understand the actuator saturation
Submitted by: macmardg
Submitted on: January 5, 2005
Identifier:
H9
There is a maximum side force u that the physical ducted fan can produce (i.e. an actuator saturation). The forces are generated by a fan, and the thrust can be vectored through some angle by a set of flaps (controlled by a model airplane servo motor).
To stabilize the fan in hover, the thrust must be set at f_2 = m*gamma = 7N. (Note that the class notes may be different from the problem set. The control u = f1 is the side force produced by vectoring, as opposed to the axial force f2 required to counteract gravity.)
If we assume that the total thrust does not change as the thrust is vectored, then the maximum side force is equal to the nominal thrust times the sin of the vectoring angle of 40 degrees. (This is the behaviour one would get if the vertical control were open-loop, that is, simply set at a fixed thrust. If there were a control loop closed on the vertical position, then a better assumption would be that the vertical component of the thrust remained fixed at 7N. The actual saturation algorithm used in the Simulink is closer to the latter and somewhat more complex.)
Almost any controller you devise will result in an actuator saturation even for very small step inputs in the desired position. This is not necessarily a problem, but does mean that the actual system would not respond in the same way as your linear model does. You do not need to redesign the controller to obtain better performance in the presence of the saturation.
[Back to Top]
-
I'm still confused about the actuator saturation!
Submitted by: mjdunlop
Submitted on: January 8, 2005
Identifier:
H9
The 40 degree angle listed in 1c is the angle of the fan flaps. Note that this is NOT theta, which is the angle the entire fan is tilted. If the maximum thrust generated coming out the back of the fan is 7N you can figure out what component of that is a purely sideways force (u) when the flaps are at a 40 degree angle. This number is the maximum value of u you can have, so compare that to what your plot of u vs. time looks like.
[Back to Top]
|
