Announcement

Collapse
No announcement yet.

Comparing calculated hourly solar

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • Comparing calculated hourly solar

    I currently have 2 arrays at 322 azimuth. I hope to add another inverter and an array at 232 and another array at 52. I want to compare hourly predicted outputs of each array throughout the year. I know that I can generate hourly output for each array in PVWatts but the data tables are large and unwieldy. Is there a solar output prediction program that displays hourly output in a fashion that is easier to compare the hourly outputs from these three different azimuths?
    In an output limited system, I hope to avoid peak generation around the same time for all of my panels.

  • #2
    The short answer is no - at least not that I know of.

    FWIW, I've not found an ~ 9,000 row spreadsheet much more different or difficult than a 13 row spreadsheet by more than a bit. 3 runs, 1 per orientation, combine or compare the hourly outputs as you wish. SAM works similar, but from how you write, it sounds like the input required for that model will be more than your willing to spend time on.

    Other models I know of are a bit more tedious than PVWatts. If you want hourly output, you're pretty much stuck with a several thousand row output and then either paring it down, or have a model that will perform that task. I'd guess that such models are not cheap. Maybe I'm wrong.

    Comment


    • #3
      Aurorasolar can easily model this but it is not free.
      OutBack FP1 w/ CS6P-250P http://bit.ly/1Sg5VNH

      Comment


      • #4
        When I first wrote this question, I was trying to design a system that I have since learned will not be permitted. As background, I have SE6000A with 24 panels and power optimizers which total 7.54 Kw. I rarely clip and don't export as much power as I would like (3,752 KWh/past 12 mo). I have been told that I can modify my NEM permitted system by replacing my original 6 kw inverter and with a 7.6 Kw inverter and add more panels. A 7.6 Kw inverter can be legally oversize by 1.35% to 10.26 Kw. Since I already have 7.54 install that would allow me to install 2.72 Kw more panels. Using 0.325 Panasonics, I would be able to add 8 more panels (2.6 Kw). However, the inverter will have to be "derated" to the original wattage of my original inverter, 6 kw. That would allow me to generate more power but also result in more clipping. Rather than try to create multiple bar graphs to model this system's output, I generated an hourly output .csv file from PVWatts for my original system 7.54 Kw @ 322 azimuth and one for my proposed new array 2.6 kw @ 232 azimuth. The I copied the "AC system output" last column of data (8780 rows) from the 2nd file and pasted it beside the same "AC system output" column of data in my 1st file. Then, I created a new column that contained the 'sums' of the hourly outputs. Then I created another new column with the IFS formula, which basically said if the 'sum' value was greater than 6000 (w) make the new value 6000. If the sum value was less than or equal to 6000, make the new column data the same as the 'sum' column data.
        Of course this is all from PVWatts estimates but my July 2018 data production was rather close to my PVWatts estimated July output.
        The PVWatts estimated annual outputs:
        Original 7.54 kw = 10,737
        New 2.6 kw array = 4,248
        Total output without derating = 14,985
        Total output derated at 6 kw = 14,313
        Theoretical Clipping loss = 672
        Net gain from addition = 3,576
        %Loss (672/4,248) = 15.8%

        I realize that my 'derating calculation' was on hourly values and the inverter would be doing instantaneous deratings so I would presume (never ASSUME) that this averages out over one hour. If my calculations are correct based on PVWatts estimates, it would seem to be worthwhile to add the 8 more 325w panels and use the SE7600H-US inverter that I got for $299.
        At $0.34/kwh and climbing, 3,576 would be worth $1,216/yr.
        I welcome any comments or suggestions.
        Last edited by Steeler.Fan; 09-07-2018, 07:49 PM.

        Comment


        • #5
          Now that wasn't so bad was it ? Bet you learned a lot too.

          Given the idea that the weather for any one hour that PVWatts uses will have little bearing on the actual weather for the actual corresponding hour, means that the amount of power that actually gets clipped and the times of that clipping will vary, what you did will probably be close enough.

          My guess is that the max. value of any one clipping event will be higher or lower than your spreadsheet shows, and the times of clipping and the duration of those events will be different, but over a year's timeframe, and over a bunch of years, it'll probably get close.

          Analogous to the difference between climate and weather, any model's output is what you expect, the array's output is what you get. Just like predicting the weather upon which it's highly dependent, estimating an array's output is not an exact science. What you did and the output of the model you got is probably close enough.

          Comment

          Working...
          X