Announcement

Collapse
No announcement yet.

PVOutput with Rainforest Eagle and Enphase

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

  • PVOutput with Rainforest Eagle and Enphase

    Hi all,

    My 4.7kW system was installed last Friday and I just got my Enphase Enlighten account today. I had previously set up my account (with donation) on PVOutput and started uploading via auto-loader of Rainforest direct (instead of Wattvision). Now I have it also getting my solar production while "testing" and waiting for PG&E to finish up approval.

    My question is around Primary vs Secondary devices. The only choice for the Eagle is as primary so I have the direction set for Net and then the Enphase is Secondary and set for Solar. Anyone with an Eagle auto-loading direct now or is everyone still using WattVision? I currently have the Eagle uploading to Bidgely but I'm not married to that. Does it really matter if solar is the secondary device?

    Also, for the net consumption, should I be putting in a net delay as well? What's the best way of figuring out the best value?

    Thanks,
    Derek

  • #2
    The difference between setting the inverter as primary or secondary seems to affect the timing of data sampling slightly, but overall, does not really affect the accuracy of the output. As for the net delay, you might just try playing with that setting and see what you get. If you set it too short, you might get some obviously bad calculations on gross consumption, but mostly, when I've played with that value I've found the response to be minor. LogdIn is the goto guy around here for setting up PVOutput... he has an Enphase system, and may be able to offer more insight.

    As for the Rainforest Eagle... I still use Wattvision (or my own software polling the eagle directly). The Eagle direct uploader, by default, only loads instantaneous demand, not average power. Trying to determine the 5 min energy consumption based on a single instantaneous demand measurement leads to a lot of error that can stack up over the course of the day. The "net history" option has PVOutput look at the kWh counters reported by the Eagle, instead of the demand. In principle, it would produce a better 5 min average, but the kWh counters don't have a consistent update rate and you will sometimes get an indication of 0 power consumption if it doesn't update by the next 5 minute status interval. If you keep your status interval 10 min or longer, it helps avoid the 0's. The "net history" option does give very good end of day totals, though.

    Wattvision reports the average demand over each 5 min period based on the instantaneous demand sampled every ~8 sec. The 30-40 point average does a decent job of representing the demand over that period, and the accumulated error in the consumed energy is not usually not too bad. I have recreated what Wattvision does with my own software so I'm not dependant on the cloud, and find the calculations to be pretty accurate. I could correct the accumulated error periodically by using the (unpredictably updating) kWh data, but haven't found the errors to be big enough to be worth fixing.
    CS6P-260P/SE3000 - http://tiny.cc/ed5ozx

    Comment


    • #3
      Thanks for the info sensij.

      I have Enphase Enlighted and a Rainforest Eagle, and was just thinking of switching to direct to eliminate the extra step. It sounds like I should just leave it on Wattvision.

      on thing I don't like about using Wattvision, is the data timing is slightly off form the Enphase enlighten data. this does not really matter when the solar output is smooth, but on partly cloudy days when solar is oscillating, the power consumption bounces around too, which mean it's subtracting the Wattvision data from the incorrect sample of the enlighten data. Any ideas on how I could fix that?

      Comment


      • #4
        Originally posted by stevesreed View Post
        Thanks for the info sensij.

        I have Enphase Enlighted and a Rainforest Eagle, and was just thinking of switching to direct to eliminate the extra step. It sounds like I should just leave it on Wattvision.

        on thing I don't like about using Wattvision, is the data timing is slightly off form the Enphase enlighten data. this does not really matter when the solar output is smooth, but on partly cloudy days when solar is oscillating, the power consumption bounces around too, which mean it's subtracting the Wattvision data from the incorrect sample of the enlighten data. Any ideas on how I could fix that?
        Yes, data synchronization is hard and PVOutput.org doesn't make much of an attempt to help. I've casually suggested in its forum to add a time offset that could be used to better align the data from different sources, but haven't pushed on it and am not sure it would really solve the problem to the level that we would like. Ultimately, I ended up writing by own data collection scripts that do a much better job of synchronizing the data... you can see in my PVOutput chart that the consumption numbers are much cleaner. Compare my main system with the test system to see the difference... the test system is still using the SolarEdge and Wattvision uploaders.

        A Raspberry Pi would probably be the way to do your own collection, if you were interested. Otherwise, installing a TED or other CT based meter (I like EKM) that monitors consumption and generation with a single device and gives you true consumption (as opposed to back calculating it from generation and net consumption) would be another way to get a clean PVOutput consumption chart.
        CS6P-260P/SE3000 - http://tiny.cc/ed5ozx

        Comment


        • #5
          Originally posted by sensij View Post
          Yes, data synchronization is hard and PVOutput.org doesn't make much of an attempt to help. I've casually suggested in its forum to add a time offset that could be used to better align the data from different sources, but haven't pushed on it and am not sure it would really solve the problem to the level that we would like. Ultimately, I ended up writing by own data collection scripts that do a much better job of synchronizing the data... you can see in my PVOutput chart that the consumption numbers are much cleaner. Compare my main system with the test system to see the difference... the test system is still using the SolarEdge and Wattvision uploaders.

          A Raspberry Pi would probably be the way to do your own collection, if you were interested. Otherwise, installing a TED or other CT based meter (I like EKM) that monitors consumption and generation with a single device and gives you true consumption (as opposed to back calculating it from generation and net consumption) would be another way to get a clean PVOutput consumption chart.
          I might try a local collection script at some point. That does sounds like the best way to get cleaner data.

          One problem with that approach, is it's not easy to get other people setup when they want to join PVOuput.. Having off the shelve solutions makes it easier to help other people get started.

          Comment


          • #6
            Originally posted by stevesreed View Post
            I might try a local collection script at some point. That does sounds like the best way to get cleaner data.

            One problem with that approach, is it's not easy to get other people setup when they want to join PVOuput.. Having off the shelve solutions makes it easier to help other people get started.
            Agreed. I don't like saying... here, this is what you can do, without really offering a path to go do it. A Raspberry Pi script would be fairly transferrable, only a few parameters would need to change from one system to the next. Forum member logdin has already worked out the local envoy data collection, and although it would take some work to add the eagle and data synchonization, I've got a reasonably good routine in a different programming language that could provide a starting point.

            Maybe it would be worthwhile to sell preprogrammed Pi's for the different systems... Enphase+Eagle, SolarEdge+Eagle, etc. For, say, $100, how many people would buy that over the free (but imperfect) service PVOutput offers?
            CS6P-260P/SE3000 - http://tiny.cc/ed5ozx

            Comment


            • #7
              I know it's slightly off topic, but I have a concern over putting both production _and_ export data online. I noticed this when analyzing my own consumption data- it becomes very easy to spot patterns of when people are home and not. Although PVOutput doesn't show exact addresses, it is potentially a data point that someone looking for empty houses might investigate.

              I'm fine showing solar production, but not consumption or any combination that can be used to calculate consumption.

              Comment


              • #8
                It depends a bit on the exact consumption pattern, but yes, many people have this fear. If you upload consumption data and want to hide it (or share it only with other trusted users you've identified), you have that option if you become a donator (<$10 / year).
                CS6P-260P/SE3000 - http://tiny.cc/ed5ozx

                Comment


                • #9
                  I have mentioned many times how difficult it is to synchronize consumption and generation data from two different sources well enough to make accurate gross consumption calculations. I missed it when it was released last year, but PVOutput now supports a user selected time shift to help the auto-uploaders match up better. In 5 min increments, the utility of it might still be limited, but maybe it will develop from there. I don't know whether the time difference between two devices remains consistent enough to rely on a shift like this to align them.

                  https://getsatisfaction.com/pvoutput...-time-shifting
                  CS6P-260P/SE3000 - http://tiny.cc/ed5ozx

                  Comment


                  • #10
                    Originally posted by sensij View Post
                    The Eagle direct uploader, by default, only loads instantaneous demand, not average power. Trying to determine the 5 min energy consumption based on a single instantaneous demand measurement leads to a lot of error that can stack up over the course of the day. The "net history" option has PVOutput look at the kWh counters reported by the Eagle, instead of the demand. In principle, it would produce a better 5 min average, but the kWh counters don't have a consistent update rate and you will sometimes get an indication of 0 power consumption if it doesn't update by the next 5 minute status interval. If you keep your status interval 10 min or longer, it helps avoid the 0's. The "net history" option does give very good end of day totals, though.
                    sensij - thanks for this info - very useful.

                    I have just noticed that my Eagle consumption flowing to PVOutput is way off compared to what is reported by PG&E website, but if I look at the daily consumption on Rainforest Cloud, it matches exactly. I am assuming that the Cloud consumption data is also based on kWh counter.

                    So now I am trying to fix my PVOutput setup to try to correct this - when you mention "net history" option is this the same as setting the History Data option to Enabled? Just want to make sure I am doing the right thing - maybe the PVOutput interface on this has changed since 2015.

                    Thanks!

                    Comment


                    • #11
                      Originally posted by SunJoker View Post

                      sensij - thanks for this info - very useful.

                      I have just noticed that my Eagle consumption flowing to PVOutput is way off compared to what is reported by PG&E website, but if I look at the daily consumption on Rainforest Cloud, it matches exactly. I am assuming that the Cloud consumption data is also based on kWh counter.

                      So now I am trying to fix my PVOutput setup to try to correct this - when you mention "net history" option is this the same as setting the History Data option to Enabled? Just want to make sure I am doing the right thing - maybe the PVOutput interface on this has changed since 2015.

                      Thanks!
                      FYI, Sensij hasn't been heard from around here for ~ 1.5 years or so.

                      Comment


                      • #12
                        Originally posted by jimqpublic View Post
                        ...when analyzing my own consumption data- it becomes very easy to spot patterns of when people are home and not.
                        You probably already have neighbors, mailmen, delivery drivers, emails, social media posts, out of office replies and even your own utility providers picking up clues as to when you aren't home. I doubt the data on PVOutput is a source that crooks are using to tie it all together for a night of crime.

                        Dave W. Gilbert AZ
                        6.63kW grid-tie owner

                        Comment

                        Working...
                        X