A User's Manual for DELSOL3 - prod.sandia.gov - Sandia National ...

system variables, an optimal field is built up. At each design point power level at-
tained in the field build-up, the levelized energy cost is tested to see if it is a new
minimum. If it is a new minimum, the system design, performance, and costs are
saved. When the iterations are complete, the lowest energy cost system from the
allowed system variations is known at each power level.
Even if the user asks DELSOL to examine every system in the range of vari-
ables (IALL=l), not every system will be examined. DELSOL contains tests to
avoid searching over systems that cannot meet the minimum power of interest. If
a given tower height is too small to meet the minimum design power level for any
receiver, DELSOL automatically skips to the next tower height without search-
ing (in vain) for the optimum receiver size. Also, DELSOL checks to make sure
that cavity receiver geometries are consistent. Requests to allow aperture widths
which are wider than twice the depth of the cavity are ignored. This is due to the
DELSOL assumption that cavity receivers are configured internally as right circu-
lar cylinders centered on the aperture., so that the interior of the cavity will never
be greater than twice the depth of the cavity.
Other than for the above exceptions, DELSOL will examine every system re-
quested in the range of variables when IALL=l. However, considering every pos-
sible combination of system variables is an inefficient way to find the optimum
system(s). DELSOL, therefore, has an option to do a “smart” search over the
receiver variables. This option (IALL=O) is the default option. Smart search-
ing tries to consider the minimum set of system variables necessary to find the
optimum system(s). The search algorithm attempts to start iterating a receiver
variable at a value greater than its minimum allowed value and to stop iterating a
variable before it reaches its maximum allowed value. The search strategy makes
the following assumptions: (1) there is a single, well defined minima in each vari-
able, which is not significantly dependent on the choice of the other receiver vari-
able or the tower variables; and (2) the optimum tower height and receiver size
increase when the power level increases.
The first assumption of smart optimization, that a single well defined min-
ima in each receiver variable exists, allows DELSOL to find that minima for the
second receiver variable using the first choice of the first receiver variable which
results in the requested design point power level, and then to continue searching
for the minima in the first receiver variable for values of the second receiver vari-
able close to its minima. If increasing a receiver variable increases the energy cost
of each system whose power level can be reached by changing that variable, DEL-
SOL will stop the iteration and assume that a minima has been found for that
variable. However, in actuality the minima values will move slightly as other re-
ceiver and tower dimensions are changed. Therefore, after DELSOL has found a
minima for an inner variable and increases an outer variable to the next value,
it starts the iteration on the inner variable at the discrete value which is one less
than the minima found previously. For example, suppose the user has allowed re-
ceiver sizes of 10, 11, 12 ... 20. At THT = 150 the minimum receiver size that is
116