You could simulate the spectrum with, e.g. the SMARTS code (http://www.solarconsultingservices.com/smarts.php), but you would still need accurate inputs, which usually implies at least a sunphotometer--unless you are lucky and close to an existing station with such equipment. If you're only interested in the _global_ spectrum (as opposed to the direct spectrum--much more difficult to model), you could potentially replace the sunphotometer by a pyrheliometer (to derive the aerosol optical depth) and a GPS-met station (to measure precipitable water in near real-time).

Predicting the solar spectrum under cloudy and/or rapidly variable conditions will always be difficult, unless you have a ton of additional observations. For those cases, a spectroradiometer is definitely the best solution.

You could simulate the spectrum with, e.g. the SMARTS code (http://www.solarconsultingservices.com/smarts.php), but you would still need accurate inputs, which usually implies at least a sunphotometer--unless you are lucky and close to an existing station with such equipment. If you're only interested in the _global_ spectrum (as opposed to the direct spectrum--much more difficult to model), you could potentially replace the sunphotometer by a pyrheliometer (to derive the aerosol optical depth) and a GPS-met station (to measure precipitable water in near real-time).

Predicting the solar spectrum under cloudy and/or rapidly variable conditions will always be difficult, unless you have a ton of additional observations. For those cases, a spectroradiometer is definitely the best solution.

You could simulate the spectrum with, e.g. the SMARTS code (www.solarconsultingservices.com/smarts.php), but you would still need accurate inputs, which usually implies at least a sunphotometer--unless you are lucky and close to an existing station with such equipment. If you're only interested in the _global_ spectrum (as opposed to the direct spectrum--much more difficult to model), you could potentially replace the sunphotometer by a pyrheliometer (to derive the aerosol optical depth) and a GPS-met station (to measure precipitable water in near real-time).

Predicting the solar spectrum under cloudy and/or rapidly variable conditions will always be difficult, unless you have a ton of additional observations. For those cases, a spectroradiometer is definitely the best solution.

You could simulate the spectrum with, e.g. the SMARTS code (http://www.solarconsultingservices.com/smarts.php), but you would still need accurate inputs, which usually implies at least a sunphotometer--unless you are lucky and close to an existing station with such equipment. If you're only interested in the _global_ spectrum (as opposed to the direct spectrum--much more difficult to model), you could potentially replace the sunphotometer by a pyrheliometer (to derive the aerosol optical depth) and a GPS-met station (to measure precipitable water in near real-time).

Predicting the solar spectrum under cloudy and/or rapidly variable conditions will always be difficult, unless you have a ton of additional observations. For those cases, a spectroradiometer is definitely the best solution.

Hi Ahish,

I think it is better to use MODTRAN or PC-MODTRAN.

Gives very accurate results. Look at:

http://modtran5.com/purchase/index.html

Cheers,

Frank

Thank you all ,i already made a science model to capture the radiation pattern into spectrum & do analysis into online image processing software.

I believe you want to measure the solar spectral IRRADIANCE (W/m2/nm) not the response, correct ?

However to test accurately your PV system i'll recommend to use a Solar simulator which is more stable that the direct solar exposure. Unfortunately it has also a cost.