Hi Mike,
Thank you for the question. The trough modeling team has been working on it and here's what we found:
1. There is bug in the power tower field optimization code (it uses a program called PTGen) in SAM 2011.12.2 that is fixed in the attached file. The bug has to do with the way the code handles decimal values, and only affects the power tower model with large numbers of small heliostats.
The bug does not affect power tower systems with typical large heliostats, so you do not need to implement the bug fix unless you are modeling fields with small heliostats.
To implement the fix, please download the attached file. Be sure SAM is closed and go to your SAM installation folder (C:\SAM\2011.12.2 by default in Windows) and find the exelib\tools folder. Rename the PTGen.exe file to something like PTGen-old.exe. Then copy the attached version of the file to the folder (it must be named PTGen.exe).
We have not fully tested this version of PTGen, so if you encounter errors with the new version of PTGen, please revert to the old version by changing the file names.
2. Now for some comments regarding the power tower field optimization algorithm and small heliostats. SAM uses an model called DELSOL for the optimization (PTGen is part of DELSOL).
DELSOL uses an empirical curve to determine the "ideal" spacing between heliostats as the field is laid out. This curve was developed at Sandia National Laboratories quite a while ago, and the prevailing assumption at that time was that heliostats should be relatively large and should be spaced to avoid significant shading or blocking. With those assumptions, the solar field performance can be approximated by breaking it up into a set of large zones like the ones on SAM's power tower Heliostat Field page.
Now smaller heliostats are becoming more common. The empirical spacing curve in DELSOL may not be accurate when the number of individual heliostats in a zone is in the thousands rather than the tens. We therefore urge strong caution in using DELSOL in this type of situation because we have not done any analysis to show whether the results are realistic.
Best regards,
Paul.