DELTAE is versatile in the way it uses different model parameters as guesses to meet its targets: length or volume (to achieve resonance at fixed frequency), stack length and position (to meet an efficiency and amplitude), or stack diameter (to get adequate power), for example. When such geometric variables are released to the solver for manipulation (or when they are made to change in a plot loop), there are often certain geometric relationships to other parameters that we would like to see maintained. For example, if the area of a duct increases, we must increase the associated perimeter as well. Another common wish is to lengthen a segment while simultaneously shortening another segment to keep overall length constant. Also, in a stack made of a constant thickness material in a duct of fixed diameter, we cannot blindly vary the pore size and expect the porosity to remain the same-this could lead to a misleading optimization if we are faced with these constraints. If we go to the trouble to calculate a porosity for our initial segment, we want DELTAE to respect it for the values it chooses as we run the model. `Special modes' were introduced to link parameters for just these purposes.
A special modes dialog appears automatically whenever a parameter linking capability is recognized for a variable that is chosen as a guess vector member:
MAIN: (rpwPncTCgudvomfst e?)> u
Guess/Target Address=? ( 0a) 4d
Selection: STKCIR:r0
Add to the guess vector (y/n)? y
Special modes can be enabled as this parameter is varied
(Only one mode per segment possible):
Mode Description
0 Normal mode (no inter-related parameters)
-1 Adjust porosity while y0(r0) varies (const. Area, L0)
Mode=? ( 0) -1By selecting -1 for the special mode, we have asked DELTAE to remember the following constant before it begins iterating:
where r0, poros are the pore size and porosity of our initial stack. We assume that the effective plate material half-thickness is L0=r0*r0(1/poros)-1. During the iteration, as r0 is changed, DELTAE assumes porosity changes as an ideal porosity would and calculates it from the following:
and the effective plate thickness is maintained.
If we create a plot varying the area of our first INSDUct (parameter 2a, in most of the examples of the previous chapters), the dialog looks like
MAIN: (rpwPncTCgudvomfst e?)> p
define plot variables. One or two inputs (a-j)
and up to 10 outputs (A-J) can be plotted)
Plot Parameter Address=? ( 0A)2a
use for outer or inner (2d) plot loop (o/i)? o
Outer (or 1-D) Plot Loop:
Independent variable is ISODUC:Area
Plot begins at ISODUC:Area = 1.2920E-02 m^2
New value (<CR> to keep)=?
Plot ends at ISODUC:Area = 1.2920E-02 m^2
New value (<CR> to keep)=? 2e-2
with a step of: 1.00
New value (<CR> to keep)=? 1e-3
Special modes can be enabled as this parameter is varied
(Only one mode per segment possible):
Mode Description
0 Normal mode (no inter-related parameters)
-2 Maintain consistent Perimeter as initial Area varies
Mode(n)=? ( 0)-2
By selecting -2 for the special mode, we have asked DELTAE to remember the constant:
and, later, to calculate the perimeter from
This relationship keeps circular ducts circular and maintains the aspect ratio of rectangular ducts.
A very complicated example, even if somewhat confusing, can give some idea of the power of parameter linking. Interesting iterations can be done by using sameas parameters in combination with length parameter (always `c', for segments that have a length) linking. For example, if segments 2 and 7 are ISODUcts, and segments 4 and 5 are STKSLabs of equal length (but different material or porosity, perhaps), we can iterate using stack length, keep these lengths equal, and keep the overall length and stack center position constant by doing the following: