# Solar flux calculation

Posted by mundt on August 2, 2008, 11:00 pm

I wish to calculate the solar flux falling on my panels (Ip) for
modeling purposes.

I have:
1. the insolation from a local meterology station (at local solar
noon)  =  Is
2. the solar declination (at solar noon)  = Td
3. my latitude = Tl
4. my panel tilt (azimuth) =Tp
5. my panels face due south

To correct the insolation for the declination of the sun, I believe
that I need to divide the measured insolation (Is) by the cosine of
the actual sun angle (Td-Tl). Thus the solar flux on a surface NORMAL
to the sun is Is/cos(Td-Tl).

To correct for the tilt angle of my panels, I believe I need to
multiply the NORMAL flux  by the cosine of the angle between the
normal surface and the actual panel surface, i.e. cos(Td-Tl-Tp).

The flux my panels then see is thus:    Ip = Is *cos(Td-Tl-Tp)/cos(Td-
Tl).

Is this a correct formulation?

Posted by Bill Kaszeta / Photovoltaic Re on August 4, 2008, 2:38 am

Not so simple.

There are two major measurements made by many observing stations,
global irradiation on a horizontal surface (using a pyranometer) and direct

Sunshine on a tilted surface has three components, direct normal (the
component that can be concentrated with a magnifying glass), diffuse
(the blue sky and/or reflections from clouds), and reflected from the
ground (albedo).

There is some good information at
http://www.nrel.gov/rredc/models_tools.html
and the related site.

Your calculations apply to the direct beam component on a clear day.
You need to make sure you understand the origin of the data you are using
before applying calculations.

There are computer programs that do these calculations, but
the old links I have no longer function.  There are solar radiation
handbooks with full details.

Bill Kaszeta
Photovoltaic Resources Int'l
Tempe  Arizona  USA
bill@pvri-removethis.biz

Posted by mundt on August 4, 2008, 4:47 pm
On Aug 3, 7:38pm, wkasz...@cox.net (Bill Kaszeta / Photovoltaic
Resources) wrote:

Thank you Bill,

My local insolation data appears to be from a horizontal pyranometer
(data at  http://gate1.baaqmd.gov/aqmet/Met.aspx?PID=SI ).

My objective is performance modeling of my solar installation, i.e.
tracking performance changes over time. I am modeling the output of my
installation as a function of independently measured insolation,
temperature, and wind speed. I am hoping for a model accuracy
(resolution?) of less than 1% so that single cell failures could
potentially be detected.

I am selecting data only from "clear" days (by inspection of output
power vs. time graph). It looks to me like the "blue sky" component
can be ignored (for my goal) - it should effect both the pyranometer
measurement and my panel output equally (maybe some very small
spectral response related differences?).

I mainly wanted to confirm that my correction to the measured
insolation (normal direct component anyway) accurately accounted for
the geometric mounting difference between the pyranometer measurement
and my panels.

Thank you again for you comments.

Randy