Advertisement
If you have a new account but are having problems posting or verifying your account, please email us on hello@boards.ie for help. Thanks :)
Hello all! Please ensure that you are posting a new thread or question in the appropriate forum. The Feedback forum is overwhelmed with questions that are having to be moved elsewhere. If you need help to verify your account contact hello@boards.ie

Solar PV Monitoring/Automation Thread

1383941434471

Comments

  • Moderators, Education Moderators, Home & Garden Moderators Posts: 8,193 Mod ✭✭✭✭Jonathan


    Just sitting on top of the heatsink on rubber tipped standoffs. Will take a pic next time I'm up in the attic.



  • Registered Users, Registered Users 2 Posts: 16,102 ✭✭✭✭Ha Long Bay



    Thanks was just looking at the stats today and mine getting a lil on the warm side.





  • Registered Users Posts: 52 ✭✭Sligobuck


    I use this Honeywell fan on the floor on its lowest setting and it keeps the Sofar 6K Inverter just above 40C on days like today.




  • Registered Users, Registered Users 2 Posts: 1,431 ✭✭✭Big Lar


    The fan would work fine I imagine if it fixed to a heavy weight on top of the inverter to counter balance the fan, problem is that I have the inverter in the utility and the wife already gives out a bout the noise of the inverter fan.


    I was actually in a house the other day and saw these radiator reflectors and was wondering if I could put some fins against the wall behind the inverter would I get more passive cooling, the radiant heat from the inverter would heat the plate and cause the air to rise, I think that this may work well for me as I have the inverter 50mm out from the wall already


    @Jonathan Have you any idea of what temps you were seeing before the fan install ?



  • Moderators, Education Moderators, Home & Garden Moderators Posts: 8,193 Mod ✭✭✭✭Jonathan



    Turns out this day last year was also a cracking day. Temps peaked at 50°C; 17°C higher.




  • Advertisement
  • Registered Users, Registered Users 2 Posts: 2,547 ✭✭✭bullit_dodger


    Might be better over here than the quotes thread. but @HotSwap can you explain this a little more?

    or just recently I got a great suggestion here and I charge up to 80% in the morning; at which point I set the system to feed excess PV to the grid (having the space in the battery prevents the system from clipping during the peak times) and once the peak solar has passed I change the mode of the system back to use excess solar PV to charge the battery.

    This seems "a bit odd". What I mean is, if it's likely that your going to be experiencing clipping.....why charge your battery at all? If your expecting clipping chances are it's going to be a super day, no?



  • Registered Users Posts: 824 ✭✭✭HotSwap



    The charging I’m talking about above (up to 80% is actually coming from solar) say from 8am-10am. Then it hits 80% and I stop charging the battery and feed it to the grid until 3pm.

    But I am charging from grid at night also because is it’s cheeper (prob not the right word- more financially beneficial in the end) to fill the battery between 2-4 am for 12 cent rather than use excess solar that I can get 21 cent for exporting during the day. So the more I put into the system at night; the sooner I am exporting the following day. I know it seems backwards; but I aM considering charging the battery fully each night even during these times of plentiful solar.

    so I’m charging at night based on this rule:


    target_soc: >-    {% set solar_forecast = states(entity_solar_forecast)|float %} {% if    solar_forecast > 25.0 %} 75 {% elif solar_forecast > 20.0 %} 80 {% elif    solar_forecast > 15.0 %} 85 {% elif solar_forecast > 10.0 %} 90 {% elif    solar_forecast > 5.0 %} 95 {% else %} 100 {% endif %}


    ok so 7.5 kWp of panels on a 6kwh inverter.

    It clips all the time; on days like yesterday the highest it got was 6337:

    if the battery was full that would of been a straight line capped at 6kwh.


    take the 18th; there are moments when I spiked to 7251w:




  • Registered Users, Registered Users 2 Posts: 2,547 ✭✭✭bullit_dodger


    Ahh yeah, I kind of see what your doing now. Makes sense.

    I'd have a slightly different opinion myself on that one. I'd max charge the battery at night and not mess about with a 80% stop. Why? Well any clipping you experience will be 100's of watts at most for a hour or two. While dropping to 80% of a (what?) 10Kwhr battery is 2kw give or take.

    I also think your clipping isn't a result of having an non-full battery, I think your clipping is because the panels isn't as hot as other days. If I take 3 days this week (including the glorious yesterday) and chart them against each other I see.....

    Today (in red) I saw occasions where it was generating more than the glorious blue skies yesterday, esp when the sun came from behind a cloud. I'm not sure that battery state not being full is actually stopping you from clipping. It's just co-incidental.

    Certainly with my inverter battery status makes zero difference in clipping/non-clipping for me.

    Btw - it's probably pennies either way - but it's interesting to do the optimizing for fun :-)



  • Registered Users Posts: 824 ✭✭✭HotSwap


    I don’t have any clipping in them two examples. Clipping would be where the inverter has nowhere to put the power from the panels and limits their production. It would show up as a perfectly straight horizontal plot.

    Your 100% right on the temp impacting the peak output yesterday; but it’s not clipping in my understanding of the term.

    in the huawei system (with a 6kw inverter) cos it’s a high voltage battery; if I have 7.5 kw coming from the panels; 6kw of that can get inverted into AC and the other 1.5kW can get pushed directly into the battery as long as it isn’t already fully charged.


    the 80% charge cap is only temporarily imposed during the afternoon where I can get more than 6kw from the panels. as soon as this passes I allow the battery to charge fully from the panels.

    it might only be 500W on a good day (I’ll try and calculate it at some point when I have more data). But it’s energy that I would not of generated had the battery been sitting at 100%; so i think it’s worth it.

    I think I can charge to 90% rather than 80%; as I think 500-1000 w of non clipped excess (dunno what I’m saying here) is prob more than optimistic.

    and I will prob do away with my nightly charge cap based on the following days forecast and just go to 90% every night.

    one part of me would like to try and go a month without importing anything from the grid. But another part just wants to max out the returns. The second part prob wins :)



  • Registered Users, Registered Users 2 Posts: 2,547 ✭✭✭bullit_dodger


    Yeah, I think I get what your saying. It makes sense - though since it's just that small bit at the top end of the sine wave (assuming your south facing) 90% is probably plenty

    That said it's be interesting to see a generation curve on a cloud day where you know your battery is full - at least to prove the hypothesis in the negative. I have a sneaking hunch that you'll still see those 7000Kw spikes figures that you have in the 2nd chart above. If you saw those with the battery full that would disprove the theory....



  • Advertisement
  • Registered Users Posts: 824 ✭✭✭HotSwap


    It’s not possible; the power has nowhere to go; this is what happens when the battery is full and the sun pops out from behind a cloud on an otherwise mainly overcast day:



  • Registered Users Posts: 1,084 ✭✭✭silver_sky


    I'd like to setup an entity to track the daily peak DC power in homeassistant.

    Anyone done this? how would I do it?



  • Registered Users, Registered Users 2 Posts: 2,547 ✭✭✭bullit_dodger



    Yeah, it does look like that's the case alright. I don't suppose you have access to raw data values for that chart? I'm just wondering if there's possible a charting issue there? I saw a bug in the GivEnergy app doing something similar where there were values "cropped" off that chart because the scale was wrong. As in the chart when up to (say) 4000 and yet there were some actual values at 4122 or something. The same chart on the webportal displayed fine (they fixed that bug)

    Course you could be 100% correct - I'm just curious.



  • Registered Users Posts: 824 ✭✭✭HotSwap


    Ha. You’re really having a hard time believing this one. This is just how things work with a high voltage DC battery. Here are the same days directly from the huawei app; I can export the data if I need to yeah.


    And here are the battery SOC for the same days:




  • Moderators, Home & Garden Moderators Posts: 6,211 Mod ✭✭✭✭graememk


    Happens on the solis hybrids, too, and I think possibly on the sunsynks as well.

    I've seen screenshots of the solis running at 7-8 kw from panels, with the battery being filled and still less than 6kw leaving the inverter.

    But I get your point though, don't have a full battery at the highest production time,

    On the straight inverter mine tops out at 6.1kw though.



  • Registered Users Posts: 824 ✭✭✭HotSwap


    I understood it was only possible with a high voltage battery. That’s great! I would like to get some experience with other brands and low voltage battery’s as I do feel that while I was willing to pay a bit more; I don’t think others should have to unless they are really into it.



  • Registered Users, Registered Users 2 Posts: 2,547 ✭✭✭bullit_dodger


    No - no trouble believing you at all mate. As I said....

    Yeah, it does look like that's the case alright.

    Your battery makes no odds to me - how you fill it or not. The reason I asked is that I work with data every day and I try not to jump to false conclusions. The chart with the 6000 scale could potentially be simply truncating higher values. Its not - and that's cool. The problem when you don't check these things is that you start solving problems which don't actually exist to begin with. So I was/am (if anything) trying to help you make sure that your maximizing your kit!

    Came across this for anyone interested in the difference between low/high voltage batteries.

    High Voltage or Low Voltage Batteries | Energy Storage | Deege Solar



  • Registered Users Posts: 824 ✭✭✭HotSwap


    I agree 100% your right to confirm. There are a lot of people that think they have all the info and a lot of misinformation gets pushed about the place.

    I think a lot of people have a very deep understanding of a specific setup (usually their own home system) but very few people have knowledge of many different systems.

    Ive only been running this automation a few days now and I already changed it 2 times. I have just now changed it so disable that charge limit during the night; so I will go to 100% or as close to it as I can get in the 2-4am window. And also changed that “pause” of charging to be 95% until 3pm.



  • Registered Users, Registered Users 2 Posts: 16,102 ✭✭✭✭Ha Long Bay


    So just a quick update on this the fan you suggested arrived today.

    Up and running now connected to a Sonoff Wi-Fi switch so was able to automate it in Home Assistant to turn on when the temp gets over 30 degrees.

    It seems to be very efficient in reducing the inverter temperature.



  • Registered Users, Registered Users 2 Posts: 309 ✭✭JayBee66


    Have any Solarman/Home Assistant users noticed a discrepancy between what their Solarman and Home Assistant apps are reporting?

    Both are using the same logger data but Solarman is 7KW less than Home Assistant.

    I can only assume that Solarman is dropping some data. Whereas, my Home Assistant (running on my local network) is capturing all logger data.

    Note - Home Assistant has only been running since the last week in March.



  • Advertisement
  • Registered Users, Registered Users 2 Posts: 1,470 ✭✭✭DC999


    That's only a 1.3% difference on the 530kWh used. Could be down to rounding even as they both use the same data source you say. That one app rounds it each day and adds the rounded total day on day. So the 2 apps differ over time.

    You can check if the % remains the same as time goes on.



  • Registered Users Posts: 1 cylstudio


    Hello everyone,

    I am delighted to have found this forum and spent around two hours reading through a thread from page 1 to 41. Although I have no experience with coding, I have been following various YouTube videos and have successfully used the pysolarmanv5/solismod to control my Solis 3.6 inverter with a DLS-W logger. I have two Pi set up, one running HA continuously and the other running solismod, pysolarmanv5, and mqqt.

    I am facing an issue during the setup, where I keep receiving the "No Socket Available" error when I execute the script solismod/main.py.

    I believe this was due to the fact that my second Pi was connected to an AP router, which was connected to a switch. However, I have now connected my second Pi to the same switch as the first Pi, which is running the HA system. There has been no "No Socket Available" error for also 24 hours.

    Currently, I have to manually check if the script is still running by using "sudo ps -ef | grep main.py". I am concerned that if the script is shut down due to the "No Socket Available" error, then my automation will not work.

    To avoid the script shutdown due to network issue, power cut...etc, I have used "sudo crontab -e" to run solismod/main.py at @reboot.

    My question is, is there a way to set up a script or something that will continuously check if the solismod/main.py is running? If the script has stopped running or exited, then it will automatically restart the script. Also, is it possible to set up a status indicator on HA dashboard to check the status?

    I apologise if my description is unclear. As I mentioned earlier, I have no knowledge of coding, Python, or Pi OS. I am thrilled that I have come this far, and everything is working as expected. This is the last thing I need to ensure that the system is stable. Thank you for your time, and I hope someone can help.

    Best regards, Madison



  • Registered Users Posts: 60 ✭✭dauhee


    you could try try using socket_timeout with a value greater than 60 and also auto_reconnect=True


    They might buy you a little resiliency but should really try fix the network issue which you may have done. What I did was put in a "while True" at the top of the code and error handle the code inside so it will try forever, reglardless of error and do a littletime.sleep(60) to give it a breather on each fail



  • Registered Users Posts: 60 ✭✭dauhee


    modbus.read_input_register_formatted(register_addr=33173, quantity=2)

    Just wondering if anyone knows how to get the daily exported amount (running total). I thought the above would work but it just gives the lifetime export amount

    any assistance would be much appreciated



  • Moderators, Education Moderators, Home & Garden Moderators Posts: 8,193 Mod ✭✭✭✭Jonathan




  • Registered Users Posts: 60 ✭✭dauhee


    yeah thats it exactly, 5kw Solis Hybrid, thanks



  • Registered Users Posts: 60 ✭✭dauhee


    OK I'm talking out loud here but running this to get what is coming from/going to the grid:

    modbus.read_input_register_formatted(register_addr=33173, quantity=2)

    gave me correct values when I dump to grid i.e. amount of watt going to grid, however when I consumed, it gave me a value close to the threshold of an int, so what I'm doing is if its "> 4294967296 - my max panel capacity", then I subtract it away from 4294967296

    so seems OK now. It may be already covered somewhere but I'm just catching up with this mouth watering thread

    PySolarmanV5 really is a gift from a god


    I'm wondering is there any potential issue running it constantly at a hight frequency i.e. every 2 seconds? Just in case it can cause early equipment failure I'm only keeping it to every 60 seconds for now. I do notice that the logger stick goes to sleep and takes a few seconds to wake up when it hasn't been polled for some time, maybe its just for powersaving

    Post edited by dauhee on


  • Registered Users Posts: 60 ✭✭dauhee


    OK I see earlier in this thread that register for exported today is mis-labelled in the solis PDF and its 33175



  • Moderators, Education Moderators, Home & Garden Moderators Posts: 8,193 Mod ✭✭✭✭Jonathan


    High praise indeed. 😂 Glad it is useful.

    In general, it might be easier to just let pysolarmanv5 handle the negative values for you by using the signed parameter.

    e.g. modbus.read_input_register_formatted(register_addr=33079, quantity=2, signed=True)



  • Advertisement
  • Registered Users Posts: 60 ✭✭dauhee


    Ah brilliant. Oddly enough, I am getting an intermittent disparity now where live consumption on register 33147 is not exactly the mirror of grid on register 33130 i.e. consumption of 614watt but grid could be -619watt. When I did my overcomplicated hacky thing with 4294967296 they were always exactly mirrored. Signed int is the way to go though, less code for me to muck up



Advertisement