Diffuse and direct irradiance from global

RosarioRA

Hi everyone. I'm using data from an solar collector experimental instalation. I have information on the global irradiance on an horizontal surface and I'm trying to calcule the irradiance on the tilted surface. I've found a lot of methods (Liu and Jordan, for example), but I'm still confuse. In the paper for Marko Gulin et. al (Estimation of the global solar irradiance on titled surfaces), described that:


G_T=B_h r_b + D_h R_d + ρG_h R_h

Where
G_T= global solar irradiance on tilted surfaces
B_h= direct (horizontal) solar irradiance
r_b= direct irradiance conversion factor
D_h= Diffuse horizontal irradiance
ρ= foregrounds albedo
G_h= global (horizontal) irradiance
R_d= diffuse irradiance transposition factor
R_h= transposition factor for ground reflection

I supose the information I already have is G_h (factors are easily calculated). But, how can I calculate diffuse (D_h) and direct (B_h) irradiance if I have no other data from pyranometers.

Thanks a lot!!

BB.

Welcome to the forum Rosario!

Out of curiosity--Why?

Much of this ends up being statistical that (probably) approaches noise in terms of predictive ability... Weather is highly variable and, obviously, pretty unpredictable more than a week or so out for many places.

Also, to a degree, the equation leaves out the recipient of the radiation and its specific ranges of absorption such as thermal vs solar electric (i.e., snow reflects very well, dirt will reflect a different spectrum. UV is affected differently vs visible/IR light,snow in winter, grass in spring, dirt in summer, Tracking vs fixed plate, a few clouds in the right area can increase output and those same clouds 5 minutes later can decrease output, etc.).

I can see building a physical model--Say a pair of tracking solar cells. One cell with side shields to "track the sun directly". And a second one mount on the same frame with an "umbrella" that blocks direct sun but allows diffuse/reflected light onto the cell (thermal, solar electric, etc.

If the typical indirect flux is ~1/10th of the direct flux (my guess--And obviously changes a lot on weather, angles, etc.)--How "accurate" does the indirect illumination model have to be (if 10% error for indirect prediction/measurements--Then ~1% change of output).

-Bill

zoneblue

Check out https://pvpmc.sandia.gov/

RosarioRA

I just found the answer of my own question. In the book "Solar Engineering of Thermal Processes" by Duffie, in Chapter 2 there is an example of such a calculation (example 2.15.1). The only restriction is that this calculation is done for hourly solar radiation, i.e. not for irradiance. It is important to read carefully, because there are a lot of restrictions depending on your particular conditions. Also, the book "Solar Technologies for Buildings" by Ursula Eicker is very useful for understanding some concepts. With this information it is not so difficult to do the maths in excel or other software.

Thank you all!

froggersix

how does this help? the sun is the sun and every day is different anyway and you can't keep moving panels around so this is just an ad for some dumb books.

mulia

froggersix wrote: »

how does this help? the sun is the sun and every day is different anyway and you can't keep moving panels around so this is just an ad for some dumb books.

Well, I saw some mounting system with "sun-tracker" system. I think they're programmed to move every second to follow sun-path.

Just can't imagine keep moving those heavy panels.... So much power used by those motors...

BB.

I am not sure that the motors use that much power... But the cost and maintenance of tracking hardware certainly do not help.

One poster here really likes trackers because it gets you more hours of useful sun per day--Lead Acid Batteries need all of the hours of sun per day they can get charging (LA batteries last longer with 10% rate of charge 8+ hours per day vs 20% rate of charge at 5 hours per day).

From a cost point of view--It looks like mounting more panels facing SE and SW for a "virtual tracker" may be cost effective these days with the lower cost of solar panels.

Using PV Watts for San Francisco:

"Station Identification"
"City:","San_Francisco"
"DC Rating:"," 4.0 kW"
"DC to AC Derate Factor:"," 0.770"
"AC Rating:"," 3.1 kW"
"Array Type: Fixed Tilt"
"Array Tilt:"," 37.6"
"Array Azimuth:","225.0"

"Energy Specifications"
"Cost of Electricity:","12.5 cents/kWh"

"Results"
"Month", "Solar Radiation (kWh/m^2/day)", "AC Energy (kWh)", "Energy Value ($)"
1, 3.17, 284, 35.50
2, 4.15, 340, 42.50
3, 4.90, 450, 56.25
4, 5.90, 519, 64.88
5, 6.40, 584, 73.00
6, 6.74, 592, 74.00
7, 7.22, 651, 81.38
8, 6.75, 606, 75.75
9, 6.33, 547, 68.38
10, 4.95, 442, 55.25
11, 3.38, 292, 36.50
12, 2.92, 263, 32.88
"Year", 5.24, 5569, 696.12

"Array Azimuth:","135.0"

"Energy Specifications"
"Cost of Electricity:","12.5 cents/kWh"

"Results"
"Month", "Solar Radiation (kWh/m^2/day)", "AC Energy (kWh)", "Energy Value ($)"
1, 3.09, 277, 34.62
2, 3.99, 326, 40.75
3, 4.68, 425, 53.12
4, 5.81, 506, 63.25
5, 6.20, 558, 69.75
6, 6.24, 537, 67.12
7, 6.76, 597, 74.62
8, 6.21, 548, 68.50
9, 5.97, 507, 63.38
10, 4.72, 418, 52.25
11, 3.35, 290, 36.25
12, 2.82, 252, 31.50
"Year", 4.99, 5240, 655.00

"Array Type: 2-Axis Tracking"

"Energy Specifications"
"Cost of Electricity:","12.5 cents/kWh"

"Results"
"Month", "Solar Radiation (kWh/m^2/day)", "AC Energy (kWh)", "Energy Value ($)"
1, 4.39, 398, 49.75
2, 5.61, 466, 58.25
3, 6.60, 614, 76.75
4, 8.11, 725, 90.62
5, 9.06, 842, 105.25
6, 9.37, 833, 104.12
7, 10.20, 930, 116.25
8, 9.19, 836, 104.50
9, 8.51, 741, 92.62
10, 6.72, 606, 75.75
11, 4.70, 414, 51.75
12, 4.13, 381, 47.62
"Year", 7.23, 7786, 973.25

If you can install ~1.5x more solar panels for the same price (or less) than 1x 2 axis tracking--You would be better off.

-Bill