Announcement

Collapse
No announcement yet.

Advice Needed!! Am I being robbed in proper materials

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

  • #31
    Originally posted by AzRoute66 View Post
    sensij Would you mind taking another look at the tilt? I went to PVWatts for Miami, and changed only the tilt parameter and total energy kept going up as I stepped from 13 degrees to 23 degrees.
    What happens in PVWatts if you set the GCR to 1.0, which we know will have self-shading?
    CS6P-260P/SE3000 - http://tiny.cc/ed5ozx

    Comment


    • #32
      Originally posted by AzRoute66 View Post
      sensij Would you mind taking another look at the tilt? I went to PVWatts for Miami, and changed only the tilt parameter and total energy kept going up as I stepped from 13 degrees to 23 degrees. I originally did this as 13 'sounded' too shallow. I ran it for Roof Mount and Open Rack. I'm wondering if the panels make a difference - aren't his panels the gizmos that collect light from both front and backsides? At any rate, I would appreciate any explanation of the difference.
      Without self shading considerations, the 23 deg. tilt will produce slightly more annual output than a 13 deg. tilt. With self shading considerations and the same 8 ft. pitch, the 13 deg. tilt will produce a smaller penalty on the diffuse portion of the insolation from self shading and perhaps make up for most or more than the slight reduction in annual output caused by the lower tilt.

      One other consideration is that a 13 deg. tilt would allow the 8 ft. row pitch to be tightened up a bit.

      Other considerations which are harder to quantify include the possible effects that a lower tilt will have on panel fouling as well as the rate and nature of the fouling. In general, lower tilts will probably foul faster than higher tilts and/or require more frequent cleaning to maintain performance, particularly in what's probably considered a marine environment.

      As for the added output from the rear of the panel, how much that will add to annual output is still undetermined, particularly for self shaded arrays, and SAM has no way to account for such effects, shaded or unshaded, at least not at this time. While I believe backside irradiance contributions are a small to insignificant portion of the total incident insolation on a solar panel for most applications, in theory at least, a higher panel tilt would tend to increase the backside irradiance on a panel. More side/side testing is needed.

      Comment


      • #33
        Originally posted by sensij View Post

        What happens in PVWatts if you set the GCR to 1.0, which we know will have self-shading?
        Why are you setting the GCR = 1 ?

        64" Panel / 96" Row-To-Row Spacing = 0.666

        Their design has about 3' empty space from the rear of one row, to the front of the next row.

        Comment


        • #34
          Originally posted by sensij View Post
          What happens in PVWatts if you set the GCR to 1.0, which we know will have self-shading?
          I really don't know. I didn't run it with GCR because:

          1. I had already determined that the row spacing was sufficient for non-shading with another tool.
          2. GCR is only applied to single axis trackers (according to the GCR 'info' button), of which this is not.

          I would gladly run it again, in fact I WILL run it again at some point to confirm that it is not applied to fixed mount, but if it does change the numbers I would have no idea how to interpret them.
          Last edited by AzRoute66; 10-14-2017, 01:04 PM.

          Comment


          • #35
            Originally posted by AzRoute66 View Post
            I really don't know. I didn't run it with GCR because:

            1. I had already determined that the row spacing was sufficient for non-shading with another tool.
            2. GCR is only applied to single axis trackers, of which this is not. (according to the GCR 'info' button).

            I will gladly run it again, in fact I will run it again at some point to confirm that it is not applied to fixed mount, but if it does change the numbers I would have no idea how to interpret them.
            I thought you were a "teach me to fish" kind of guy. If GCR is only applied to single-axis trackers, that suggests that self-shading analysis in PVWatts is limited to single-axis trackers, as well (since GCR is a critical input parameter for self-shading as the NREL tools model it). If you want self-shading analysis from an NREL tool, SAM if the right choice. As we've already covered in this thread, there are self-shading effects on diffuse irradiance that simple line-of-sight of the sun calculations don't address.

            NEOH setting GCR to 1.0 is just an exercise, to show PVWatts is not the right tool for fixed array self-shading analysis.
            Last edited by sensij; 10-14-2017, 01:14 PM.
            CS6P-260P/SE3000 - http://tiny.cc/ed5ozx

            Comment


            • #36
              Originally posted by sensij View Post
              I thought you were a "teach me to fish" kind of guy.
              Well, I have always thought of myself as such, but when I am standing by the riverbank with only a cane pole in my hands, I don't expect a discussion of downriggers. It was clear to me from the start that you have to confirm self shading elsewhere, which I did. I really didn't understand why you would want me to play with GCR at all. That said, I will check that the GCR value is correctly ignored in PVWatts just for completeness, on my way to SAM to see if I can reproduce your results with that.

              Regardless of how that turns out, if it is true that 13 degrees is a more productive fixed tilt than 23 I will run it with 'normal' panels, and if it comes out the same way I suspect I will be at a total loss of as to why - as fixed tilting to latitude probably became mantra for a reason. At that point we can talk about downriggers. Be back sometime this weekend, probably with a SAM question or two.

              My only possible theory on this right now is that perhaps he is so far south that winter production is completely swamped by summer, if so I will have learned something indeed. I wouldn't expect that above 23.5 degrees... Hmm, perhaps I'll swing by optimalpaneltilt first and pick up some bait.

              Comment


              • #37
                Originally posted by AzRoute66 View Post

                My only possible theory on this right now is that perhaps he is so far south that winter production is completely swamped by summer, if so I will have learned something indeed. I wouldn't expect that above 23.5 degrees... Hmm, perhaps I'll swing by optimalpaneltilt first and pick up some bait.
                If you use stink bait, you'll only catch bottom feeders.

                Try this. Page number 30 (pdf page 34).
                CS6P-260P/SE3000 - http://tiny.cc/ed5ozx

                Comment


                • #38
                  This should be stinky enough. At least enough for sensij to tell me what I mucked up. Pay no note to absolute energy. I cheated and made 8 rows of five panels because I am lazy and no good. I ran it from 9 degrees to 27 degrees so the 13 degree column is run 3, the 23 degree column is run 8. The only two areas I noted changes in loss were 'POA shading' and 'DC Module Modeled Loss' (which could be anything, I'm still reading... and there is a lot of college football that needs my supervision today). The total energy for 9 and 11 degrees was minutely less than for 13 degrees.



                  So, based on this I now concur with sensij that 13 degrees would be my number of choice. My advice to the OP is to have the contractor install the 23 degree sticks (just to have them just in case) and go back and replace them with some 13 degree sticks in his copious spare time. Plus, at the risk of annoying the contractor before he starts working, I might ask him to confirm this analysis and to ask him why he didn't suggest 13 degrees to start.

                  I now think about 'blue sky' diffuse radiation in a new way. Also, SAM is starting to appeal to me enough to put up with having to somehow find and specify an appropriate inverter. Thanks sensij - I don't say that enough do I...

                  This might be the first time I have seen a situation where I didn't really want to replace a fixed tilt with a goat powered chain driven adjustable tilt. 13 degrees fixed is fine - at 23 degrees I might still go buy a goat.

                  Edit: Re-linking the referenced SAM output below:
                  Capture3.JPG
                  Somehow it got 'disappeared' from the original... (10/19/17)
                  Last edited by AzRoute66; 10-19-2017, 02:54 PM.

                  Comment


                  • #39
                    Originally posted by AzRoute66 View Post
                    My only possible theory on this right now is that perhaps he is so far south that winter production is completely swamped by summer, if so I will have learned something indeed. I wouldn't expect that above 23.5 degrees... Hmm, perhaps I'll swing by optimalpaneltilt first and pick up some bait.
                    I'd go back to the drawing board on that one.

                    As it turns out, the closer to the equator an application is, the less difference there will be between summer and winter output. In the limit(s), the seasonal difference in system output based on solar geometry == 0 at the equator and a max. at the poles.

                    Also, It may be worthwhile to keep in mind that SAM, while about as good as it gets for everyday serious use, is till a model, and has limitations not immediately apparent even to those familiar with it. For example, for the shading model SAM use the isotropic diffuse model for the diffuse irradiance. While that's probably done to keep things workable for shading estimates (and I'd not want to be the one to attempt a diffuse shading model using an anisotropic model for the diffuse irradiance), the easier to use and understand isotropic model is not usually the first choice for estimating the entire diffuse irradiance on a surface if a closer match to measured reality with respect to distribution and behavior of the diffuse is a goal.

                    As a practical method of showing why anisotropic is a better fit, get in the shadow of the bulb portion of a streetlight on a sunny day and look in the direction of the sun. The greater brightness around the sun you'll see is burning proof of the anisotropic (and therefore somewhat directional) nature of the diffuse irradiance. Most of that circumsolar diffuse BTW, is due to forward scattering by the atmosphere and the stuff in it. which is why the sky near the solar disk appears brighter than the rest of the sky.

                    The difference between using an isotropic model vs. the anisotropic models - of which there are several anisotropic, 2 of which SAM offers as options for irradiance calcs. that do not involve self shading - may well be slight with respect results for diffuse shading, but so probably is the output loss for this application with respect to diffuse shading, and may amount to fly specs in the bigger scheme of things.
                    Last edited by J.P.M.; 10-15-2017, 08:21 AM.

                    Comment


                    • #40
                      Originally posted by sensij View Post
                      NEOH setting GCR to 1.0 is just an exercise, to show PVWatts is not the right tool for fixed array self-shading analysis.
                      PV-Watts assumes only one (1) ROW for "Fixed (Open Rack)" Array Type.
                      There is no "Self Shading" calculations for "Fixed (Open Rack)"
                      Even if the GCR field is changed in PV Watts, it changes nothing for "Fixed (Open Rack)".

                      So, we have to use the 9am-3pm Dec 21st Rule or SAM to verify.




                      Comment


                      • #41
                        Originally posted by J.P.M. View Post

                        The difference between using an isotropic model vs. the anisotropic models - of which there are several anisotropic, 2 of which SAM offers as options for irradiance calcs. that do not involve self shading - may well be slight with respect results for diffuse shading, but so probably is the output loss for this application with respect to diffuse shading, and may amount to fly specs in the bigger scheme of things.
                        Just to be clear, the isotropic model is only used to make the shade *fraction* a workable number. The actual irradiance calcs still follow from the model chosen, but are adjusted by the fraction calculated.

                        For example, under the model conditions I've been running, on 12/20 during the noon hour (a clearer day in the miami tmy3 file than 12/21), the shade factor on ground diffuse irradiance is calculated to be 0.340095, while the factor on sky diffuse irradiance is 0.921313. After those factors are applied, the diffuse irradiance available to act on the array is 145.758 W/m2 (relative to 161.291 if calculated with the Perez anisotropic model and self-shading ignored).

                        If we switch the array to be north facing instead of south facing, the calculated ground diffuse shade factor is .307544 and the sky shade factor is still .921313, but the diffuse irradiance available to act on the array is just 91.8647 W/m2 (relative to 102.968 W/m2 if calculated with the Perez anisotropic model and self-shading ignored).

                        This isn't to suggest that SAM be treated as gospel, but it seems to me that the isotropic assumption used to calculate shade fraction is not likely to be a fatal flaw, nor give much reason to doubt that something less than 23 degrees (and in the neighborhood of 13 deg) is probably the optimal tilt angle for this system when considering self-shading.

                        This kind of discussion about the merits of the model is possible because SAM is so detailed in their documentation and open about the intermediate calculations used in the models. If equivalent documentation exists for commercial modeling software like Aurora, it doesn't seem to be so easily accessible by the public.
                        CS6P-260P/SE3000 - http://tiny.cc/ed5ozx

                        Comment


                        • #42
                          I am still unclear on the definition of 'DC module modeled loss' output. I've searched the document for that term. I suspect it is an overall aggregation of some sort.

                          In the TRM, the intro to section 8 states: "after specifying the tracking and orientation options on the System Design page, on the Shading page, for the subarray's Shading mode input, choose self-shaded. I think it is happening, I just don't see it and it adds to my confusion.
                          Last edited by AzRoute66; 10-15-2017, 06:32 PM. Reason: Stupid apostrophe mangling crap.

                          Comment


                          • #43
                            Originally posted by AzRoute66 View Post
                            I am still unclear on the definition of 'DC module modeled loss' output. I've searched the document for that term. I suspect it is an overall aggregation of some sort.

                            In the TRM, the intro to section 8 states: "after specifying the tracking and orientation options on the System Design page, on the Shading page, for the subarray's Shading mode input, choose self-shaded. I think it is happening, I just don't see it and it adds to my confusion.
                            If the DC Module modeled loss value is the one I think you are looking at, it is discussed in this thread:

                            https://sam.nrel.gov/node/70830

                            It is affected by self-shading as a consequence of the change in irradiance on the module, but it is not directly giving much insight into the self-shading calculation.

                            If you want to see an output more directly showing the effect of self-shading, look at "POA Shading Loss (%)"
                            CS6P-260P/SE3000 - http://tiny.cc/ed5ozx

                            Comment


                            • #44
                              Originally posted by sensij View Post

                              Just to be clear, the isotropic model is only used to make the shade *fraction* a workable number. The actual irradiance calcs still follow from the model chosen, but are adjusted by the fraction calculated.

                              For example, under the model conditions I've been running, on 12/20 during the noon hour (a clearer day in the miami tmy3 file than 12/21), the shade factor on ground diffuse irradiance is calculated to be 0.340095, while the factor on sky diffuse irradiance is 0.921313. After those factors are applied, the diffuse irradiance available to act on the array is 145.758 W/m2 (relative to 161.291 if calculated with the Perez anisotropic model and self-shading ignored).

                              If we switch the array to be north facing instead of south facing, the calculated ground diffuse shade factor is .307544 and the sky shade factor is still .921313, but the diffuse irradiance available to act on the array is just 91.8647 W/m2 (relative to 102.968 W/m2 if calculated with the Perez anisotropic model and self-shading ignored).

                              This isn't to suggest that SAM be treated as gospel, but it seems to me that the isotropic assumption used to calculate shade fraction is not likely to be a fatal flaw, nor give much reason to doubt that something less than 23 degrees (and in the neighborhood of 13 deg) is probably the optimal tilt angle for this system when considering self-shading.

                              This kind of discussion about the merits of the model is possible because SAM is so detailed in their documentation and open about the intermediate calculations used in the models. If equivalent documentation exists for commercial modeling software like Aurora, it doesn't seem to be so easily accessible by the public.
                              To your first point, agreed. What you write is pretty much what part of my post was trying to get at. I thought I was being clear. Guess not.

                              Another point I was trying to make using the isotropic diffuse reference was as an example of why models are not gospel, as well as perhaps an inference that a little knowledge can be a dangerous thing. By pointing out that SAM's use of isotropic diffuse rather than the probably more accurate and certainly more widely used and accepted anisotropic models for calculating diffuse shading only, while making things a lot more manageable, is something to keep in mind.

                              A third point I was trying to make was that the small differences in incident diffuse irradiance from using different models for estimating diffuse irradiance, and the resulting differences in output from using different models for the diffuse irradiance are probably not worth worrying about for several reasons (as you have shown here).

                              A fourth point I was trying to make, or a comment, was that perhaps all the discussion about the relatively small difference (if any) of the merits of a 13 vs. 23 deg. tilt may have on lower diffuse fraction due to shading was shifting discussion away from what are perhaps more important other design considerations such as possibly tightening up row pitch made possible from a 13 deg. tilt, or the efffects on performance that a lower tilt may have due to greater fouling that often accompanies such lower tilts.

                              Finally, the idea that SAM is as detailed as it is can be 2 edge sword. Like most all software, it'll let you do things that can squeeze more out of an application or design. But, it'll also allow designs or results from a less than astute designer that cannot be built or worse, wasteful, or worst of all, allow dangerous designs when used by those not familiar with how it works or the principles involved. I never used and I never let anyone I supervised use a piece of software that did things or designs that I, or they, could not do by hand. I guess that'll always be a bug up my butt.

                              Comment


                              • #45
                                Originally posted by sensij View Post
                                If the DC Module modeled loss value is the one I think you are looking at, it is discussed in this thread: https://sam.nrel.gov/node/70830
                                Yes, that is pretty much what I thought.

                                It is affected by self-shading as a consequence of the change in irradiance on the module, but it is not directly giving much insight into the self-shading calculation.
                                Yes, I got there by trying to breakdown 'sky' (Sdss, 8.10) and 'ground' (Srss) diffuse shading in the TRM.

                                If you want to see an output more directly showing the effect of self-shading, look at "POA Shading Loss (%)"
                                Yes, I presented that chart into post #38 but felt it was Sdss + Srss + [Whatever]. The Whatever is what I thought I might find in the 'DC Module modeled loss' parameter.

                                Any insight into the "Shading Mode input = Self Shading" TRM reference given? (I mentioned above but didn't explicitly ask) I don't see anything like that on my SAM Shading page. Don't put any additional time into it, I think I'm losing interest. It only caught my eye as half of the flat roofs in this area have snow white cool-coat and the other half have gravel or asphalt - got to be a difference between those two when you start splitting hairs to the extent SAM models do.

                                Comment

                                Working...
                                X