Buie SunShape Model

[giogio8427]

Hi,
I am interested in understanding the Buie's Sun model implemented in
Tonatiuh. I read the source code, in particular "SunShapeBuie.cpp" and
I have some questions about it:
1- If possible, I would like to have some bibliography about this
model. I downloaded Buie's thesis but I am not sure that you used this
reference to develop the sunshape function.
2- Reading the source code, if I understand well, this sunshape model
is equal to two different pillbox model, one for the disk region and
the other for the circumsolar region. Am I right?
3 - In line 85 of "SunShapeBuie.cpp" I don't understand the value used
in the function pdf theta (0.0038915695846209047). What is it?

Thank you very much

Kind regards

Andrea Giostri

[Manuel J. Blanco]

Hello Andrea,

Below, please, find some answers to your questions:

1. We developed Buie's sunshape plugin based on the following journal
articles:

- D. Buie*, C.J. Dey, S. Bosi. "The effective size of the solar cone
for solar concentrating systems". Solar Energy 74 (2003) 417–427.

- D. Buie*, A.G. Monger, C.J. Dey. "S unshape distributions for
terrestrial solar simulations". Solar Energy 74 (2003) 113–122.

However, the model defined in Buie's et alter articles is not defined
in terms of radiance (power per unit solid angle and unit surface
perpendicular to the direction of propagation), that is how we like to
define the sunshape. Thus, we needed to make some assumptions to
develop the corresponding Tonatiuh's sunshape plugin. In addition, we
introduced a correction to ensure that when the user select a sunshape
with a given percentage of circumsolar radiation, that percentage is
really achieved.

2. The way Buie's model is implemented in Tonatiuh's sunshape plugin
is NOT two pill-boxes. A pill-box sunshape, because of its simplicity,
allows using the inverse method to generate the stochastic samples.
Thus, it makes it possible to generate the stochastic samples from
explicit formulae. This is very efficient computationally.
Unfortunately, Buie's model cannot be simulated using the inverse
method, and one has to use the acceptance - rejection method, which is
computationally expensive. To improve the efficiency of the method,
what we do is to divide the acceptance - rejection method into several
rectangles, in order to take advantage of what we know about each one
of those rectangles. This is what may have given you the impression
that we are using two pill-boxes, but we are not.

3. The value 0.0038915695846209047 radians is the value for which the
probability density function is a maximum. The value of the maximum
varies with the circumsolar ratio, but the value of theta where the
maximum is obtained remains fixed.

Best regards,