Announcement

Collapse
No announcement yet.

New PV installation, still in progress

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

  • New PV installation, still in progress

    Hello!

    I live in a country where the political situation has in a way forced me to install an off-grid PV system. Energy costs have been going up and up and in the near future we'll have lots of power outages because of many factors, including no maintenance of generation equipment (which is run by the government). Just this year the KWh cost will go up at least 17% and last year the increase was around 10%.

    Since a few months ago I started sourcing components, but the most difficult part were the batteries. I could only find sealed "mixed use" batteries sold as "deep cycle" batteries. That's what all solar stuff companies carry around here. They are not exactly cheap and offer very limited Ah. The best thing I could find were Outback 200NC, which are true deep cycle batteries, but have really low capacity and high price tag (178Ah@C/20 and $475 each, advertised as 200Ah, which is right, but @C/100!). Next best were Ritar DC-12-200C which on paper are better than the Outbacks at the same price, but found very few and mixed opinions on that brand.

    I finally found someone that could get me Trojan SPRE 06 415 for $385 each delivered to my house, and that's what I'll go for. These are rated for 377Ah @ C/20.

    I don't have a really big budget, so I compromised on some components.

    So what I have is:
    - 9xSG330P Peimar 330W Poly panels (2970Wp)
    - Tristar TS-MPPT-60 charge controller (got it second hand off eBay, just 60 days use for $380 with meter)
    - 8xTrojan SPRE 06 415 (on the way)
    - Spartan Power SP-TS4500PLUS transfer switch.
    - 2xAPC SMT2200 that I'll use as inverters (here's the compromise). These I already had. I got them some time ago at an auction for $90 each.

    Additional stuff that I ordered, still not arrived:
    - MC4 connectors, Multi-Contact brand
    - Water Miser XL battery caps
    - Hydrometer

    Things I'll order soon:
    - IotaWatt for logging and monitoring of energy usage.

    So my monthly energy use varies by season. From March to April the consumption is a tad below 500KWh a month. Those are the hottest months. From November to January the energy usage is the lowest, around 350KWh a month. The rest of the year consumption varies between 350 and 420KWh.

    There's a lot of sun here. During the longer days the sun goes up around 5:30am and goes down around 6:30pm. During the shortest days the sun goes up around 6:30am and down around 5:15pm.

    From October to end of November there's a lot of rain, so there's little sun. At the same time, it's not hot and energy use is lower.

    All in all I expect to get at least some energy on cloudy days.

    During the day I have a minimum load of around 350VA. It can go to around 650VA constant. There'll be peaks for short periods of time (microwave oven, toaster, some tool, washing machine, etc.).

    During the night the load for up to a minimum of around 700VA, it can go up to 1200-1300VA when using the A/C with a peak of around 1900VA.

    As for the APC SMT2200 UPSs, they are pure sine 2200VA/1980W and have a 48V input. I tested one with the microvwave oven AND toaster simultaiously, and it can handle the load. I'll put one "Inverter" for high consumption loads, the other one for the rest of things. I know these are not meant to work continuously, but I already have a smaller SUA1000 unit working as a UPS, and it has no problem providing power for about 20hours at around 30% load. I tested this during power outages to power my critical computer stuff.

    I'll modify the APC units to be ableto better handle extended battery use. First mod, cooling: temperature controlled fan controller with 3 fans blowing over the transformers and MOSFETS. The case already has a fan, I'll replace it with a better, more silent one to extract hot air. I'll also put heatsinks on the transformers. Second mod: Schottky diode on the positive cable to the batteries, so that it won't charge the batteries off the power grid at night. I will put a bypass switch so I can charge the batteries off the UPS when required.

    The transfer switch will connect to the main panel, and it'll provide connection to the inverters. When the batteries go below a set voltage, the SMT2200s will get power from the grid and they'll act as voltage regulators. When battery voltage reaches a set voltage, the UPSs will act as inverters.

    So far only the panels have been installed on the roof:



    I still have not decided where tu put the batteries, I'll elaborate on that later along with some other doubts that I have.

    So, I'm open to comments/suggestions.



  • #2
    Before you get too far along with plans, please give up on the idea of Schottky diodes for current blocking, unless you are talking about 10A or less for battery charging.

    Your useable sun hours will be between 9am and 4pm, in the very best summer conditions. Winter, I'd expect less. Use PVWatts to get a detailed report of your expected sunlight

    Energy budget - Your loads look to be a lot larger then your batteries. have you looked at golf cart batteries ?
    Golf courses often have electric carts and they replace batteries by the pallet. Maybe order via them.
    Powerfab top of pole PV mount (2) | Listeroid 6/1 w/st5 gen head | XW6048 inverter/chgr | Iota 48V/15A charger | Morningstar 60A MPPT | 48V, 800A NiFe Battery (in series)| 15, Evergreen 205w "12V" PV array on pole | Midnight ePanel | Grundfos 10 SO5-9 with 3 wire Franklin Electric motor (1/2hp 240V 1ph ) on a timer for 3 hr noontime run - Runs off PV ||
    || Midnight Classic 200 | 10, Evergreen 200w in a 160VOC array ||
    || VEC1093 12V Charger | Maha C401 aa/aaa Charger | SureSine | Sunsaver MPPT 15A

    solar: http://tinyurl.com/LMR-Solar
    gen: http://tinyurl.com/LMR-Lister

    Comment


    • #3
      From the sunrise/sunset times the OP provides for summer/winter, it appears the OP is located relatively close to the equator. That would make for less solar potential difference as f(calender season). Seasonal weather conditions such as monsoon/dry season will probably have more influence on energy production from a PV system. Either or any way, the PVWatts model may provide some guidance with respect to solar potential provided weather and irradiance data near the OP's location is available. The international PVWatts data base of locations is a bit more spread out however, but if available it'll be a good tool for preliminary design.

      Comment


      • #4
        Forgot to mention, the A/C won't be connected to the PV system. I don't have a 240V split phase inverter anyway.

        I don't expect the system to cover for 100% my actual consumption. The goal is to pay less on energy to the state, and in the event of a blackout be able to have power for essential things like the fridge and lights, charge phones, etc.

        Why the is the diode a bad idea? Can you elaborate? There are diodes rated for up-to 400A and 100V. The UPSs charging voltage is about 59V, charge current I haven't measured, but should be below 10A (original batteries are 17-18Ah and recharge in about 3 hours).

        As for the batteries, those Trojan 6V 377Ah @C/20 are the best thing I could find, and are specific for PV applications. Where I live the golf courses are a very rare thing, so I don't think it's much of an option.

        Here's what PV Watts says about my location:

        IMG_20190725_095151.jpg

        My estimations were around 14KWh on the best days, and that's in line with that table.

        Comment


        • #5
          A diode at 10A and Vf of 0.5V gives 5 watts of heating that must be dissipated. That's not able to be done in a sealed box, so you have to have venting.
          Since you are going to use a "bypass switch" forget the diodes and simply use a inline switch for charging, or not.
          Powerfab top of pole PV mount (2) | Listeroid 6/1 w/st5 gen head | XW6048 inverter/chgr | Iota 48V/15A charger | Morningstar 60A MPPT | 48V, 800A NiFe Battery (in series)| 15, Evergreen 205w "12V" PV array on pole | Midnight ePanel | Grundfos 10 SO5-9 with 3 wire Franklin Electric motor (1/2hp 240V 1ph ) on a timer for 3 hr noontime run - Runs off PV ||
          || Midnight Classic 200 | 10, Evergreen 200w in a 160VOC array ||
          || VEC1093 12V Charger | Maha C401 aa/aaa Charger | SureSine | Sunsaver MPPT 15A

          solar: http://tinyurl.com/LMR-Solar
          gen: http://tinyurl.com/LMR-Lister

          Comment


          • #6
            Thanks for pointing that out Mike. I knew the basic operation of a diode and had a basic understanding of the different diode types, but did not know or take into account the heat it'll produce nor did I expect for the power loss it'll produce, which would be significant for the amperage of the inverter (~45A at full load from the batteries).

            So here are some of the doubts I have:

            * The batteries should be located in a cool dry place, and because they are FLA that place has to have good ventilation. Batteries' life is cut in half for every 10

            Comment


            • #7
              So far I've been doing the AC rewiring of the house in a subpanel. That part turned out more complex than expected and was a considerable cost. But it's advancing.

              I also have placed the batteries in a room.

              IMG_20190818_181615.jpg

              Comment


              • #8
                In happy to report that my PV system has been working for 5 days now. Solar harvest is in line work my calculations so far, 14.28kWh on average.

                The first day I was extremely happy, with a total harvest of 16.44kWh. Since then, the harvest has been going down by around 1 kWh a day. It's been cloudy and August is supposed to be one of the worst months of the year for solar production.

                Production starts at around 5:30am with a few Watts and by 9am there's 1800W (like 80% of the maximum generated power). The maximum steady power generated is between 2200 and 2300W. By 5pm there is no power generated or just around 5W.

                Going by the panels' labels the maximum STC power is 2970W, so I'm getting 80% of that.

                On the not so bright side (pun intended), my batteries are not getting fully charged save for two days. So I'll have to limit the DoD these months or I'll have to install one or two more strings when I got the money.

                Yesterday I observed something unusual. At around 10:30am the sky was cloudy and harvested power was just around 700W. Suddenly the sun came out with full brightness and generated power jumped to 3050W and stayed there for a couple of minutes until the clouds messed with my happiness again.

                IMG_20190827_104910.jpg

                How is it possible that the panels generate more power than what the label says?

                Comment


                • #9
                  Originally posted by Glock24 View Post
                  How is it possible that the panels generate more power than what the label says?
                  Its a phenomenon called edge of cloud effect. Ive had it described to me as such, a thin cloud acts like a magnifying glass concentrating light on the panels that exceeds STC conditions briefly. Hence the power increase. Another possibility is that thin clouds reflect light rather than absorb it so therefore your panels see both direct light as well as reflected light. Again, a power increase.

                  A third possibility is that your panels were cooler than normal because of the clouds. Since temperature affects a panels efficiency a temperature lower than STC conditions could account for the observed power increase as well.

                  Hopefully others will chime in and describe the effect better than I.
                  Last edited by Suprasoup; 08-28-2019, 01:45 PM.

                  Comment


                  • #10
                    Originally posted by Suprasoup View Post

                    Its a phenomenon called edge of cloud effect. Ive had it described to me as such, a thin cloud acts like a magnifying glass concentrating light on the panels that exceeds STC conditions briefly. Hence the power increase. Another possibility is that thin clouds reflect light rather than absorb it so therefore your panels see both direct light as well as reflected light. Again, a power increase.

                    A third possibility is that your panels were cooler than normal because of the clouds. Since temperature affects a panels efficiency a temperature lower than STC conditions could account for the observed power increase as well.

                    Hopefully others will chime in and describe the effect better than I.
                    I would think of more like you're getting direct AND indirect light. Makes sense if you think about it. On a clear day you're only getting direct light from the sun (~1,000w/m^2). On a cloudy day only filtered indirect light (~300w/m^2). If there's a break in the clouds and a clear view of the sun you get both (~1300w/m^2).

                    Comment


                    • #11
                      It might be the clouds acting as a magnifying glass or the panel getting direct and indirect light. It's definitively not that the panels got below 25C as it's very hot here. Ambient temp was about 30-31C that day.

                      For comparison here's what solar harvest looks under "normal" conditions with good sun.

                      IMG_20190828_193149.jpg
                      Last edited by Glock24; 08-28-2019, 09:38 PM.

                      Comment


                      • #12
                        Originally posted by Glock24 View Post
                        How is it possible that the panels generate more power than what the label says?
                        Several ways:

                        1.) Panels are often spec'd as so many STC W + a few %, - zero. After burn-in most of the over tolerance goes away.
                        2.) Cold(er) weather will increase efficiency by maybe ~ 0.5 % for every degree C. the panels/array operate at. Example: If the array operates at, say, 24 C, the nominal efficiency will increase by something around 0.5 % or a little less from it's STC efficiency which is measured at 25 C. As the array operates under bright sun, the efficiency will drop as the array warms from the sun to something like 25 - 30 C. above the ambient air temp.
                        3.) The Plane of Array (P.O.A.) irradiance may sometimes be greater than the STC irradiance. That can happen when irradiance is reflected off the surroundings such as from buildings, vegetation or clouds. Cloud reflection is, in theory, and as a practical matter except that clouds are more fleeting, no different than building or ground reflected radiation.

                        Sunlight reaching an array comes from 3 sources: Direct beam (the light that comes directly from the and casts a shadow)., diffuse, and reflected. The last two are called albedo.

                        Under very clear skies, direct beam irradiance accounts for something like ~ 80 % or so of the incoming irradiance an array sees. Diffuse irradiance from the sky starts out as beam irradiance at the top of the atmosphere and gets scattered by air molecules and other stuff in the atmosphere like water vapor, dust, smog, etc.

                        Under clear skies and a very clean atmosphere, the diffuse irradiance usually amounts to ~ 15 - 20 % of the total irradiance an array sees. Most good models usually treat "most" of the diffuse irradiance as coming from an apparent origin "near" the sun, meaning most sunlight scattering is "forward" in nature. That irradiance's spectral distribution however, tends to be of the longer wavelengths that PV devices don't utilize very well (shorter wavelengths are scattered more, hence clear skies are blue). All that's interesting but off topic.

                        Under complete cloud cover, all the irradiance is diffuse. In hazy or dusty atmospheres, the diffuse to beam ratio will be higher than for clear/clean skies.

                        The rest of what an array sees is reflected from the surroundings - including clouds. That reflected irradiance may be either beam - if it starts out as beam and is specularly reflected from a mirror like surface, or diffuse - if reflected from a diffuse surface such as a light colored building or a cloud.

                        Add all those sources of irradiance and it can amount to more than the 1 kW/m^2 irradiance level that is the standard used to for STC panel efficiency measurements.

                        All the above means that under certain conditions and at certain times, a panel or an array may produce more than it's rated STC output because the input is, at some times, greater than the 1 kW used to measure the panel efficiency and output.

                        For a discussion of albedo, see Wikipedia.

                        Take what you want of the above. Scrap the rest.

                        Comment


                        • #13
                          It's been two weeks since my system started working, and yesterday was my worst day. It was cloudy/rainy for most of the day and it only reached peak performance for brief moments.

                          The total harvest was 9.3kWh, my best has been 17 and average is around 13.

                          IMG_20190907_100624.jpg

                          Also yesterday I noticed again the momentaneous spyke of 3200W I mentioned in one of my pervious posts. I guest I'll be seeing that on cloudy days.

                          With cloudy days like yesterday I also noticed something rise that raised some doubts.

                          When 30A are going into the batteries the voltage will slowly climb until it reaches the absorb voltage target. But then it becomes cloudy and only 5-10A are going into the batteries the voltage drops quickly.

                          Assuming the sun shines 24h a day, with just 5A of charging power will the batteries never reach absorb voltage?

                          Comment


                          • #14
                            Originally posted by Glock24 View Post
                            .............

                            When 30A are going into the batteries the voltage will slowly climb until it reaches the absorb voltage target. But then it becomes cloudy and only 5-10A are going into the batteries the voltage drops quickly.

                            Assuming the sun shines 24h a day, with just 5A of charging power will the batteries never reach absorb voltage?
                            The only thing that would make that happen is that your house loads are greater. I am not familiar with the self drain of your batteries but self drain could affect that as well.
                            if I recall you have intermittant acess to the grid and that would be the time to either put your loads on the grid and charge your batteries.
                            9 kW solar. Driving EVs since 2012

                            Comment


                            • #15
                              Originally posted by Ampster View Post
                              The only thing that would make that happen is that your house loads are greater. I am not familiar with the self drain of your batteries but self drain could affect that as well.
                              if I recall you have intermittant acess to the grid and that would be the time to either put your loads on the grid and charge your batteries.
                              My loads are around 10A. With good sun I get 40A from the charge controller. When it's a bit cloudy I get 15-20A. So the 5-10A are going to the batteries.

                              According to the spec sheet, the batteries' self discharge rate at the temperature here is around 10% every two weeks. I don't think that's an issue with the low power charging.

                              Comment

                              Working...
                              X