What Do PVWatts' Numbers Really Mean?

a0128958
a0128958 Solar Expert Posts: 316 ✭✭✭
If I use the simple Version 1 of PVWatts, select the closet city to me (Fort Worth, TX), put in my system's DC Rating (36 panels * 225 W each = 8.1 kW), leave unchanged the DC to AC Derate Factor (at 0.77), and accept everything else as default (Array Type = Fixed, Array Tilt = Ft. Worth's latitude, and Array Azimuth = 180°), I get the following for Solar Radiation (kWh/m²/day):

Jan. - 4.32
Feb. - 4.77
Mar. - 5.50
Apr. - 5.98
May - 6.02
Jun. - 6.25
Jul. - 6.39
Aug. - 6.31
Sep. - 5.83
Oct. - 5.56
Nov. - 4.43
Dec. - 4.10

Year - 5.46

What I think these numbers represent is the actually measured insolation, averaged over many years of time, for each month, for wherever the measuring instrument was set up in Ft. Worth, and that there's no influence to these numbers by panel type, or panel orientation to the sun, or efficiency of the system. I.e. these are actual insolation numbers representative of monthly averages. Further i.e., the numbers take into account all atmospheric effects (cloudiness, humidity, dust, etc.), but not temperature, specific to Ft. Worth.

If this is correct, then is it correct to compute 'nameplate' performance of my system in this manner?:

Jan. - 4.32 kWh/m²/day * 31 days/month * 225 W/panel * 36 panels = 811 kWh
Feb. - 4.77 * 28 * 8.1 kW = 807 kWh
Mar. - 5.50 * 31 * 8.1 = 1001
Apr. - 5.98 * 30 * 8.1 = 1035
May - 6.02 = 1058
Jun. - 6.25 = 1032
Jul. - 6.39 = 1069
Aug. - 6.31 = 1055
Sep. - 5.83 = 955
Oct. - 5.56 = 981
Nov. - 4.43 = 794
Dec. - 4.10 = 769

Year = 16,128 kWh

Adding up these 'nameplate' energy amounts to accumulate for the year, it's 16,128 kWh. I think this is the annual theorectical output, specific to the Ft. Worth location, taking into account atmospheric effects, without regard to any derating factors, *and*, at a standard test condition of 25 °C.

I don't really care, though, about 'nameplate' performance at a test temperature. To get to the theoretical output of this system taking into account atmospheric *and* temperature conditions (i.e. 'weather'), it looks like I have to travel a different path by instead using PVWatts' AC Energy estimates.

Starting with PVWatts estimates at the default 0.77 Derate Factor:

Jan. - 811 kWh
Feb. - 807
Mar. - 1001
Apr. - 1035
May - 1058
Jun. - 1032
Jul. - 1069
Aug. - 1055
Sep. - 955
Oct. - 981
Nov. - 794
Dec. - 769
Year - 11,367

Then, to get to theoretical system output, taking into account all weather (atmospheric *and* temperature) conditions, but before any considerations for wiring, string design, shading, etc., these numbers need to be divided by the default 0.77 Derate Factor, i.e.:

Jan. - 811 / 0.77 = 1055 kWh
Feb. - 807 etc.
Mar. - 1001
Apr. - 1035
May - 1058
Jun. - 1032
Jul. - 1069
Aug. - 1055
Sep. - 955
Oct. - 981
Nov. - 794
Dec. - 769
Year - 11,367 / 0.77 = 14,762 kWh

I.e., for this system, taking into account all weather conditions for Ft. Worth, the theoretical best it can do is estimated at 14,762 kWh annually.

In my case, for my 8.1 kW system, for the last 12 months, I harvested 10,302 kWh. This is 70% of the 14,762 number. I think this means that when you account for my multiple orientations per string, only having one inverter, shading, not having direct 180° south orientation, wiring losses, inverter inefficiencies, and everything else, on an annual basis, it's costing me 30% of what the system could theoretically deliver. My fundamental question is, is this a reasonable conclusion?

Many thanks!

Best regards,

Bill

Comments

  • ggunn
    ggunn Solar Expert Posts: 1,973 ✭✭✭
    Re: What Do PVWatts' Numbers Really Mean?
    a0128958 wrote: »
    If I use the simple Version 1 of PVWatts, select the closet city to me (Fort Worth, TX), put in my system's DC Rating (36 panels * 225 W each = 8.1 kW), leave unchanged the DC to AC Derate Factor (at 0.77), and accept everything else as default (Array Type = Fixed, Array Tilt = Ft. Worth's latitude, and Array Azimuth = 180°), I get the following for Solar Radiation (kWh/m²/day):

    Jan. - 4.32
    Feb. - 4.77
    Mar. - 5.50
    Apr. - 5.98
    May - 6.02
    Jun. - 6.25
    Jul. - 6.39
    Aug. - 6.31
    Sep. - 5.83
    Oct. - 5.56
    Nov. - 4.43
    Dec. - 4.10

    Year - 5.46

    What I think these numbers represent is the actually measured insolation, averaged over many years of time, for each month, for wherever the measuring instrument was set up in Ft. Worth, and that there's no influence to these numbers by panel type, or panel orientation to the sun, or efficiency of the system. I.e. these are actual insolation numbers representative of monthly averages. Further i.e., the numbers take into account all atmospheric effects (cloudiness, humidity, dust, etc.), but not temperature, specific to Ft. Worth.

    If this is correct, then is it correct to compute 'nameplate' performance of my system in this manner?:

    Jan. - 4.32 kWh/m²/day * 31 days/month * 225 W/panel * 36 panels = 811 kWh
    Feb. - 4.77 * 28 * 8.1 kW = 807 kWh
    Mar. - 5.50 * 31 * 8.1 = 1001
    Apr. - 5.98 * 30 * 8.1 = 1035
    May - 6.02 = 1058
    Jun. - 6.25 = 1032
    Jul. - 6.39 = 1069
    Aug. - 6.31 = 1055
    Sep. - 5.83 = 955
    Oct. - 5.56 = 981
    Nov. - 4.43 = 794
    Dec. - 4.10 = 769

    Year = 16,128 kWh

    Adding up these 'nameplate' energy amounts to accumulate for the year, it's 16,128 kWh. I think this is the annual theorectical output, specific to the Ft. Worth location, taking into account atmospheric effects, without regard to any derating factors, *and*, at a standard test condition of 25 °C.

    I don't really care, though, about 'nameplate' performance at a test temperature. To get to the theoretical output of this system taking into account atmospheric *and* temperature conditions (i.e. 'weather'), it looks like I have to travel a different path by instead using PVWatts' AC Energy estimates.

    Starting with PVWatts estimates at the default 0.77 Derate Factor:

    Jan. - 811 kWh
    Feb. - 807
    Mar. - 1001
    Apr. - 1035
    May - 1058
    Jun. - 1032
    Jul. - 1069
    Aug. - 1055
    Sep. - 955
    Oct. - 981
    Nov. - 794
    Dec. - 769
    Year - 11,367

    Then, to get to theoretical system output, taking into account all weather (atmospheric *and* temperature) conditions, but before any considerations for wiring, string design, shading, etc., these numbers need to be divided by the default 0.77 Derate Factor, i.e.:

    Jan. - 811 / 0.77 = 1055 kWh
    Feb. - 807 etc.
    Mar. - 1001
    Apr. - 1035
    May - 1058
    Jun. - 1032
    Jul. - 1069
    Aug. - 1055
    Sep. - 955
    Oct. - 981
    Nov. - 794
    Dec. - 769
    Year - 11,367 / 0.77 = 14,762 kWh

    I.e., for this system, taking into account all weather conditions for Ft. Worth, the theoretical best it can do is estimated at 14,762 kWh annually.

    In my case, for my 8.1 kW system, for the last 12 months, I harvested 10,302 kWh. This is 70% of the 14,762 number. I think this means that when you account for my multiple orientations per string, only having one inverter, shading, not having direct 180° south orientation, wiring losses, inverter inefficiencies, and everything else, on an annual basis, it's costing me 30% of what the system could theoretically deliver. My fundamental question is, is this a reasonable conclusion?

    Many thanks!

    Best regards,

    Bill
    No system delivers 100%. It just means that for your particular system, the derating is about .70 instead of the .77 default. The best I have seen in the field has been about .80 or a little better.
  • a0128958
    a0128958 Solar Expert Posts: 316 ✭✭✭
    Re: What Do PVWatts' Numbers Really Mean?
    ggunn wrote: »
    No system delivers 100%. It just means that for your particular system, the derating is about .70 instead of the .77 default. The best I have seen in the field has been about .80 or a little better.

    Thanks. I didn't know how to figure the derating for my system without first computing theoretical output at 100% that takes into account weather conditions.

    Noting your comment, instead of all of this computing, I could have tried different derating numbers until I matched up PVWatts' annual estimate to my actual annual amount. Sound like the derating number would be 0.70. I'll go try this.

    Best regards,

    Bill
  • a0128958
    a0128958 Solar Expert Posts: 316 ✭✭✭
    Re: What Do PVWatts' Numbers Really Mean?
    ggunn wrote: »
    No system delivers 100%. It just means that for your particular system, the derating is about .70 instead of the .77 default. The best I have seen in the field has been about .80 or a little better.

    I had a moment today to put in to PVWatts (via trial and error with the Derating Factor) my annual harvested amount for the past 12 months (10,302 kWh), and indeed PVWatts responds with a 0.70 Factor. So, now I know that for an actual annual harvest amount, the actual Derating Factor can be computed (one number representative of the whole year) using PVWatts.

    Doing the same thing on a monthly basis, I see that my Derating Factor ranges from 0.41 to 0.88. It looks like my tilt, azimuth, and shading factors are particularly bad (much worse) in the Winter than in the Summer:

    SolarPVefficiency.png

    Best regards,

    Bill
  • sudeep
    sudeep Registered Users Posts: 2
    Re: What Do PVWatts' Numbers Really Mean?
    a0128958 wrote: »
    If I use the simple Version 1 of PVWatts, select the closet city to me (Fort Worth, TX), put in my system's DC Rating (36 panels * 225 W each = 8.1 kW), leave unchanged the DC to AC Derate Factor (at 0.77), and accept everything else as default (Array Type = Fixed, Array Tilt = Ft. Worth's latitude, and Array Azimuth = 180°), I get the following for Solar Radiation (kWh/m²/day):

    Jan. - 4.32
    Feb. - 4.77
    Mar. - 5.50
    Apr. - 5.98
    May - 6.02
    Jun. - 6.25
    Jul. - 6.39
    Aug. - 6.31
    Sep. - 5.83
    Oct. - 5.56
    Nov. - 4.43
    Dec. - 4.10

    Year - 5.46

    What I think these numbers represent is the actually measured insolation, averaged over many years of time, for each month, for wherever the measuring instrument was set up in Ft. Worth, and that there's no influence to these numbers by panel type, or panel orientation to the sun, or efficiency of the system. I.e. these are actual insolation numbers representative of monthly averages. Further i.e., the numbers take into account all atmospheric effects (cloudiness, humidity, dust, etc.), but not temperature, specific to Ft. Worth.

    If this is correct, then is it correct to compute 'nameplate' performance of my system in this manner?:

    Jan. - 4.32 kWh/m²/day * 31 days/month * 225 W/panel * 36 panels = 811 kWh
    Feb. - 4.77 * 28 * 8.1 kW = 807 kWh
    Mar. - 5.50 * 31 * 8.1 = 1001
    Apr. - 5.98 * 30 * 8.1 = 1035
    May - 6.02 = 1058
    Jun. - 6.25 = 1032
    Jul. - 6.39 = 1069
    Aug. - 6.31 = 1055
    Sep. - 5.83 = 955
    Oct. - 5.56 = 981
    Nov. - 4.43 = 794
    Dec. - 4.10 = 769

    Year = 16,128 kWh

    Adding up these 'nameplate' energy amounts to accumulate for the year, it's 16,128 kWh. I think this is the annual theorectical output, specific to the Ft. Worth location, taking into account atmospheric effects, without regard to any derating factors, *and*, at a standard test condition of 25 °C.

    I don't really care, though, about 'nameplate' performance at a test temperature. To get to the theoretical output of this system taking into account atmospheric *and* temperature conditions (i.e. 'weather'), it looks like I have to travel a different path by instead using PVWatts' AC Energy estimates.

    Starting with PVWatts estimates at the default 0.77 Derate Factor:

    Jan. - 811 kWh
    Feb. - 807
    Mar. - 1001
    Apr. - 1035
    May - 1058
    Jun. - 1032
    Jul. - 1069
    Aug. - 1055
    Sep. - 955
    Oct. - 981
    Nov. - 794
    Dec. - 769
    Year - 11,367

    Then, to get to theoretical system output, taking into account all weather (atmospheric *and* temperature) conditions, but before any considerations for wiring, string design, shading, etc., these numbers need to be divided by the default 0.77 Derate Factor, i.e.:

    Jan. - 811 / 0.77 = 1055 kWh
    Feb. - 807 etc.
    Mar. - 1001
    Apr. - 1035
    May - 1058
    Jun. - 1032
    Jul. - 1069
    Aug. - 1055
    Sep. - 955
    Oct. - 981
    Nov. - 794
    Dec. - 769
    Year - 11,367 / 0.77 = 14,762 kWh

    I.e., for this system, taking into account all weather conditions for Ft. Worth, the theoretical best it can do is estimated at 14,762 kWh annually.

    In my case, for my 8.1 kW system, for the last 12 months, I harvested 10,302 kWh. This is 70% of the 14,762 number. I think this means that when you account for my multiple orientations per string, only having one inverter, shading, not having direct 180° south orientation, wiring losses, inverter inefficiencies, and everything else, on an annual basis, it's costing me 30% of what the system could theoretically deliver. My fundamental question is, is this a reasonable conclusion?

    Many thanks!

    Best regards,

    Bill

    Hey Bill and other guys,
    I am very noob in all this. hmm in short, I am an Architecture student from India trying to do a Retrofitting of a Building with the help of Photovoltaics. I was going to use BIPV but I am little confused about these. Because while browsing, I saw many types of photovoltaics. Although I know that we can use PV's either on facade or use them as facade, I am not sure if we can use same panels instead of walls.
    So, anyways I sm trying to use PVWatts to find out how much electricity I can generate with it.
    So I had this Sketchup Model, I used Skelion Trial Plugin to just to try arranging some panels and get some results. Attachment not found.
    I got the total Power output as 65.39kWp (130Wp*503panels). So, after Entering this value in PVWatts tool, I got the following Result: -

    City: New Delhi
    Country/Province: IND
    Latitude: 28.58° N
    Longitude: 77.20° E
    Elevation: 216 m
    Weather Data: IWEC
    PV System Specifications
    DC Rating: 65.4 kW
    DC to AC Derate Factor: 0.770
    AC Rating: 50.4 kW
    Array Type: Fixed Tilt
    Array Tilt: 28.6°
    Array Azimuth: 180.0°
    Energy Specifications
    Energy Cost: 1.8400 rupee/kWh

    "Month", "Solar Radiation (kWh/m^2/day)", "AC Energy (kWh)", And Rupeee Column
    1, 5.56, 8098,
    2, 5.90, 7556,
    3, 6.65, 9215,
    4, 7.01, 9148,
    5, 6.91, 9056,
    6, 6.34, 8029,
    7, 5.28, 6973,
    8, 5.25, 7030,
    9, 5.89, 7641,
    10, 6.17, 8369,
    11, 5.74, 7668,
    12, 5.56, 7916,
    "Year", 6.02, 96700,

    So My doubt is I want to understand that How we got values in Each of The Column. OK so Solar Radiation is from Weather Files, I don't understand How did We got AC energy Values. And Even When Bill showed his calculation of following numbers, these dont sum up to the result.
    Jan. - 4.32 kWh/m²/day * 31 days/month * 225 W/panel * 36 panels = 811 kWh (1084.75Kwh according to calculator)
    Feb. - 4.77 * 28 * 8.1 kW = 807 kWh (4.77*28*8.1=1081.83)
    Mar. - 5.50 * 31 * 8.1 = 1001
    Apr. - 5.98 * 30 * 8.1 = 1035

    so obviously I am missing very silly thing, which on all should be laughing on, but I don't know why I am not getting it. So please tell me how these values came. Also I referred to this PDF which I have attached. In that PDF, on the last page they have showed how to calculate the output. I could not even use that method as I am confused and not sure about the values for some quantities like i am not bale to find Irradiance of New Delhi, India and also I don't know the Inverter Model using. Here is the link as having problems in attachment: - http://www.mbipv.net.my/dload/MBIPV%20Reports/C2/Putrajaya%20Perdana%20Bhd.pdf

    So overall I am stuck in between above said things.

    Really sorry for such a long Post, But it would be of really great help if you guys can help me on this. And also if you can clear some of my queries: -
    1) What if I want to put PV's on south facade of the building, Do I have to tilt them or I cant keep them vertical. Same goes for East and West Facade too.
    2) And after writing such a long thing i cant remember what i wanted to ask.. May be later...


    Thanking you,

    Sudeep
  • BB.
    BB. Super Moderators, Administrators Posts: 33,431 admin
    Re: What Do PVWatts' Numbers Really Mean?

    Welcome to the Forum Sudeep!

    So, far, as I understand the number.

    The Hours of Sun column was a fixed instrument that gathered the "heat" (Energy) from the sun over a ~20 year period (during the 1960's through the 1970's--at least for San Francisco, California, USA).

    The units were probably something like Watt*Hours or kWH per sq unit... To get "hours of sun" per day--The assumed average strength of the sun is around 1,000 WH/m2 near local noon (12pm).

    So, you may see units from other sources in kWH/m2 or Hours of sun, or a couple of others I have seen... But using 1,000 WH/m2 seems to be a good conversion between the various other units to hours of sun (or whatever units you wish to use).

    So, you may have 12 hours of "sunlight", but the equivalent hitting a flat plat is closer to 4-6 hours of "full noon time sun" per day.

    So, you have the base measurements which also include "average" weather conditions (for 40-50 years ago--how has pollution and changing weather affected your region?). Next, there is the "adjustment" based on the position of the array (tilt, azimuth, 1 and 2 axis tracking for PV Watts).

    The PV Watts program takes the basic measurements and "adjust them" using "math" to integrate the observed sun into numbers based on your installation. So you will see hours of sun change based on your specific panel setup.

    The kWH per day/month/year are (from what I have seen) is a pretty simple equation:
    • Array Wattage * Derating * Hours of Sun (month average) * days in month (Feb=28 days, etc.).

    So, Array Wattage is "DC Ratings" And DC Ratings * "Derating" is AC Ratings. DC Rating is "input power" based on "marketing numbers" and AC rating is actual usable power after all of the other deratings/assumptions about power losses/efficiency have been taken into account.

    The deratings are typically around 0.77 (~81% panel derating * 5% controller derating) for a Grid Tied System. For a typical Battery Backed system they tend to be around 0.52 to 0.61 or so:
    • 0.77 panel+controller derating * 0.85 AC inverter eff * 0.80 Flooded Cell bank Eff = ~0.52 "end to end" derating

    I have to go right now--But does this help (and make sense)?

    -Bill
    Near San Francisco California: 3.5kWatt Grid Tied Solar power system+small backup genset
  • sudeep
    sudeep Registered Users Posts: 2
    Re: What Do PVWatts' Numbers Really Mean?
    BB. wrote: »
    Welcome to the Forum Sundeep!

    So, far, as I understand the number.

    The Hours of Sun column was a fixed instrument that gathered the "heat" (Energy) from the sun over a ~20 year period (during the 1960's through the 1970's--at least for San Francisco, California, USA).

    The units were probably something like Watt*Hours or kWH per sq unit... To get "hours of sun" per day--The assumed average strength of the sun is around 1,000 WH/m2 near local noon (12pm).

    So, you may see units from other sources in kWH/m2 or Hours of sun, or a couple of others I have seen... But using 1,000 WH/m2 seems to be a good conversion between the various other units to hours of sun (or whatever units you wish to use).

    So, you may have 12 hours of "sunlight", but the equivalent hitting a flat plat is closer to 4-6 hours of "full noon time sun" per day.

    So, you have the base measurements which also include "average" weather conditions (for 40-50 years ago--how has pollution and changing weather affected your region?). Next, there is the "adjustment" based on the position of the array (tilt, azimuth, 1 and 2 axis tracking for PV Watts).

    The PV Watts program takes the basic measurements and "adjust them" using "math" to integrate the observed sun into numbers based on your installation. So you will see hours of sun change based on your specific panel setup.

    The kWH per day/month/year are (from what I have seen) is a pretty simple equation:
    • Array Wattage * Derating * Hours of Sun (month average) * days in month (Feb=28 days, etc.).

    So, Array Wattage is "DC Ratings" And DC Ratings * "Derating" is AC Ratings. DC Rating is "input power" based on "marketing numbers" and AC rating is actual usable power after all of the other deratings/assumptions about power losses/efficiency have been taken into account.

    The deratings are typically around 0.77 (~81% panel derating * 5% controller derating) for a Grid Tied System. For a typical Battery Backed system they tend to be around 0.52 to 0.61 or so:
    • 0.77 panel+controller derating * 0.85 AC inverter eff * 0.80 Flooded Cell bank Eff = ~0.52 "end to end" derating

    I have to go right now--But does this help (and make sense)?

    -Bill

    Hey Bill,

    Thank you so much for the reply. I understood what u said. And I also got that we will not get the expected output because of these deratings, but what I wanted to know is that how PV Watts tool Brings the value in AC Energy Column. I also understood the formula which u wrote, but the main problem is that I will have to convince my faculties in my college about it. All they would be seeing and asking is How come Jan. - 4.32 kWh/m²/day * 31 days/month * 225 W/panel * 36 panels = 811 kWh whereas calculator shows Something else. If you wil check the pdf which I have attached, you will see that in the end they have shown many factors which affects the output. Also Inverter's efficiency and other things. If I could show how these numbers actually came, then it would solve my problem, even if Output in Real will not be same as my calculation but it will be in some proportion.
    So currently the main thing is how did u get Jan. - 4.32 kWh/m²/day * 31 days/month * 225 W/panel * 36 panels = 811 kWh and not the real value, and which precise factor or value caused this difference.

    Thank you so much for giving me your valuable time.

    Thanks,

    SUDEEP
  • westbranch
    westbranch Solar Expert Posts: 5,183 ✭✭✭✭
    Re: What Do PVWatts' Numbers Really Mean?

    you did not use the derating value in this formula

    The kWH per day/month/year are (from what I have seen) is a pretty simple equation:

    Array Wattage * Derating * Hours of Sun (month average) * days in month (Feb=28 days, etc.).
     
    KID #51B  4s 140W to 24V 900Ah C&D AGM
    CL#29032 FW 2126/ 2073/ 2133 175A E-Panel WBjr, 3 x 4s 140W to 24V 900Ah C&D AGM 
    Cotek ST1500W 24V Inverter,OmniCharge 3024,
    2 x Cisco WRT54GL i/c DD-WRT Rtr & Bridge,
    Eu3/2/1000i Gens, 1680W & E-Panel/WBjr to come, CL #647 asleep
    West Chilcotin, BC, Canada
  • BB.
    BB. Super Moderators, Administrators Posts: 33,431 admin
    Re: What Do PVWatts' Numbers Really Mean?

    Sudeep,

    I am not sure I understand the question exactly... First, the equation you typed:
    4.32 kWh/m²/day * 31 days/month * 225 W/panel * 36 panels = 811 kWh

    Technically, you have a units problem... You have "kWh/m²/day" but no m2 or conversion factor (crystalline solar panels are around 13-17% efficient). Which was how they "sort of" documented your PDF.

    So, if you assume 4.32 kWH/m²/day divided by 1kW/m²/day as ~high noon solar energy (of course, this varies with altitude, weather, pollution, etc.)--Then you would have 4.32 "Hours of Sun" per day:

    36 panels * 225 Watt per panel * 4.32 hours of January sun * 0.77 panel+GT Inverter deratings * 31 days per January = 835,259.04 WH per January = 835 kWH per January

    To reproduce the PDF documentation using another method--It should be done like this:
    1248×803×46 mm per panel (135 Watt panel) * 88 units = ~88 sq meters (from document)
    Htilt = 1,368 kWh/m2 (irradiance at Bandar Baru Bangi, tilted at 30 degrees) appears to be per year from next equation -BB

    So, we could probably look up the solar panel efficiency, but since we have sqmeter and rated power, we can work backwards:
    • 135 watts / (1.248m * 0.803) = 134.7 Watts per sqmeter
    • 134.7 W/sqmtr * 1/1,000 W/sqmtr "full sun" = 0.1347 = 13.47% conversion efficiency
    • 88 sqmtr * 0.1347 panel conv eff * 1.368 kWH/(sqmeter*year) solar energy * 0.77 panel+controller deratings = 12.49 kWH per year

    Note there is another issue with the document... It appears that the 88 panels point in four different directions (assume North/East/South/West)... The Htilt number would, most likely not be identical for all four planes. There should be a different vault for each set (the installation is near the equator, so minor differences in orientation/tilt may not make a huge difference--But it is not a "no brainer" to assume that all are the same--Or at least they need to document that the Htilt was a combined value for all four directions).

    Farther north (for example) from the equator, the north facing panels will have substantially lower energy production. The East/West/South production may not be obvious as to which is best... If you have cool/clear mornings, the east facing panels may perform better than the west facing panels (hot afternoons+thunder clouds).

    The PV Watts data will account for weather, tilt, and azimuth--But it does not really account for temperature differences (the 0.4% reduction in degree C and 25C rise over ambient is documented in the PDF--And is a significant factor for deratings (I use 0.81 for panels to assume the worst).

    The mounting of the panels is "interesting". It appears that the panels are over cutouts in the roof. Sealing the solar panels against rain has got to be a big issue (I am not sure I would attempt an installation such as documented).

    Another interesting concept is the bottom of the panels may be in a condition space (cooling from air conditioning cools panels, increases output). Or it may not be (heat rises, little air circulation, raises solar cell operating temperature, reducing output). And there may be issues with "insulation" (hot panels with less than ideal insulative properties bringing heat into conditioned space).

    Regarding the hand waving on other losses, PV Watts does have a page documenting the range of values.

    I would not get too wrapped in the single digit accuracy for overall efficiency value of the system. It is probably difficult to measure system performance vs predicted performance to much more accuracy than 10% without some very good quality power meters and reference cells. Also, the year over year variation of solar ratio is probably on the order of 10% variation (and probably closer to 20% if you have significant variations in weather patterns (ocean temperatures, marine layer, etc.).

    Using PV Watts for New Delhi:
    "Station Identification"
    "City:","New Delhi"
    "State:","IND"
    "Lat (deg N):", 28.58
    "Long (deg W):", 77.20
    "Elev (m): ", 216
    "Weather Data:","IWEC"

    "PV System Specifications"
    "DC Rating:"," 8.1 kW"
    "DC to AC Derate Factor:"," 0.770"
    "AC Rating:"," 6.2 kW"
    "Array Type: Fixed Tilt"
    "Array Tilt:"," 30.0"
    "Array Azimuth:","180.0"

    "Energy Specifications"
    "Cost of Electricity:"," 1.8 rupee/kWh"

    "Results"
    "Month", "Solar Radiation (kWh/m^2/day)", "AC Energy (kWh)", "Energy Value (rupee)"
    1, 5.61, 1013, 18.64
    2, 5.94, 941, 17.31
    3, 6.66, 1143, 21.03
    4, 6.98, 1129, 20.77
    5, 6.86, 1112, 20.46
    6, 6.28, 984, 18.11
    7, 5.24, 856, 15.75
    8, 5.22, 866, 15.93
    9, 5.88, 946, 17.41
    10, 6.20, 1041, 19.15
    11, 5.79, 958, 17.63
    12, 5.62, 991, 18.23
    "Year", 6.02, 11980, 220.43

    So--I can see how you would be confused from the PDF document... It makes lots of design assumptions that may not be true. And has mixed the units so much, that I am not sure I would trust much of what was in the report.

    -Bill
    Near San Francisco California: 3.5kWatt Grid Tied Solar power system+small backup genset
  • ggunn
    ggunn Solar Expert Posts: 1,973 ✭✭✭
    Re: What Do PVWatts' Numbers Really Mean?

    To confound things even further, PVWatts may not be consistent within itself. I ran a pair of simulations with PVWatts v1 and v2 for the same location, and the results varied by nearly 10%, even though the location I chose was right on top of one of the NREL data gathering points.