How I started planning
Get data. In my opinion, getting the most and most accurate data is an important first step in planning a solar installation. For 2 years, I would take a picture of the power meter twice a day, once in the morning (as close to sunrise as I could) and once in the evening (as close to sunset as possible). Using this data, and data from the grid operator's invoices, I created a spreadsheet to find out how much power I used per day (estimated from invoices) and how much power I used during daylight and dark hours.
example spreadsheet :
|date||Daylight consumption||Night time consumption||Daily consumption|
Of course, this is a tedious process, but it will give you pretty indicative values with which to calculate the size of the system you will need, and the more data it contains, the more precise your calculations will be.
Plan your sizing
The values shown above would indicate that I would need a system that could produce about 7.8 kWh of power during a typical day (6.5 kWh + 20% to account for cloudy days and slightly higher consumption than anticipated). The data also implies that if I wanted to run the system 24/7, I would need at least 4.2 kWh of batteries to run the system at night and the panels would have to produce 13 kWh of power in a typical day (7.8 kWh to run the house and 5.2 kWh to charge the batteries). It's important to know how much sun you typically get in your area, so using NASA insolation data (how much sun hits a particular area) for my location, I determined that I live in an "ok" place for solar, so how many panels would I need? Next step is to know how many sun-hours your location receives (you can also get this data online) and I found out that I get an average of 5 hours of power producing sun per typical day. Using the previous values of 13 kWh and the output of a random panel, let's say 400W i would need 13 000 W / 5 h = 2 600 Wh / 400 W = 6.5 panels (since I think it's not common to be able to order half a solar panel, let's round up to 7 panels and add 1 for good measure ... ;) )
OK, so now we know that we need 8 solar panels and 4.2 kWh of batteries, do we have space to place the panels, where they will face as close to south as possible, get no or very little shade throughout the day, and can be angled to the a nice "sun facing" angle?
I calculated the angle using values from Charles R. Landau's excellent solarpaneltilt.com like this ...
(note these values are for the northern hemisphere)
Summer tilt (March to September) = (Latitude in ºN) x 0.93 - 21º
Winter tilt (September to March) = (Latitude in ºN) x 0.785 + 19.2º
Year round tilt = average of Summer and Winter tilt
So for my location at 37.8ºN I got 14.2º for summer tilt, 48.9º for winter and 31.5º for my fixed angle tilt.
Space, the final frontier ...
My house has a roof section that faces more or less south, but it's a triangle and doesn't have a lot of area. So I had to look for a different solution since I could probably only get at most 4 or 5 panels to fit there. As luck would have, my south facing wall is two floors high, has no windows, and has nothing blocking the sun (like trees and other houses etc...)! The total width available was about 8 meters, which allowed me to purchase 8 panels that are about 1 meter wide each ...
I had a lot of difficulty finding a solar supplier that sold supports for hanging panels on a wall (in Europe). Apparently, this is not a common request. But, I'm a little stubborn, so I started looking for ground mounts. My reasoning was if a mount can hold panels on the ground, it should be able to hold them on a wall...
|with a 45% rotation ... wall mount!|
Notes on hanging panels on a wall ...
Use the longest screws you can (based on the thickness of the wall) and I recommend using chemical bushing to fill in hollow bricks (in addition the the metal / plastic bushing). The objective is to keep the screws in the wall no matter how much pressure the weight of the structure exerts or how much wind pushes the panels. Ideally, you'll get an engineer to calculate the screw length 7 diameter for this.
Measure the angle you need on the ground first. Assemble one of the triangles on the ground, adjust the angle, then measure the distance between the screw holes in the brackets, that way, you can put in the top bracket, hang a weight from it (to get vertical level), then measure along the string to get the location of the bottom bracket.
If, like me, you don't have a lot of space away from the wall to hang the panels, after you've placed all the brackets, assemble the long end of the triangle leaving them hanging vertically.
Then assemble and tighten the horizontal rails, and hang your panels, tightening all the panel brackets.
After all the panels are hung and tightened, lift the whole assembly into position, place the short end of the triangles, then tighten them all (use a torque wrench to tighten to the specifications that came with the supports).
Now, with the panels in their final position, you can get underneath them and wire them together. (don't forget to secure the wiring to the rails after wiring them to avoid the wires chaffing and reduce the possibility of the MC4 connectors disconnecting)
Selecting an inverter and batteries
Selecting an inverter
I chose the Goodwe GW-5048-EM inverter because of the characteristics of this inverter.
It uses mostly convective cooling which makes it quiet enough to have in the house.
I chose the 5KW version to account for the maximum output being able to handle the most powerful appliance in the house which is the stove that is 7 kW maximum ( at full power ) and since it has 2 MPPTs (solar sting inputs), I can add another string of panels in the future without major difficulties.
It's compatible with Pylontech US2000B batteries, which I'll get into in a minute.
Price, for the characteristics of this inverter, the price made sense, it contains all the components I needed/wanted in a small self contained unit : 50A charger, 2 MPPTs, Grid output limiter, quiet, WiFi control / stats, easy to configure and best of all a dedicated 1.5 kW UPS output.
For the panels, I chose Ameritech AS-6M144-HC Monocrystalline Half cell 440W panels. The main reasons for selecting these panels was :
Long warranty (20 year warranty + 30 year linear power output warranty)
High efficiency of up to 20.37%
Price, they were on sale @ 159 € each which is 0.36 € /W + shipping
I chose to buy 3 Pylontech US2000B 2.4 kW (2.2 kW usable) batteries, which although highly praised in reviews as being high quality and reliable, are also relatively inexpensive (I bought them at about 1 000€ each delivered). I also researched the longevity of these batteries on the outstanding website Lithium Ion Battery Test Centre in Australia, which ultimately influenced my choice. The reason behind buying 3, was that my average nightly consumption was about 6 kWh, so 3 units will give me 6.6 kWh of usable power to use during the night.
Todo : Wiring / positioning
Choosing where to install the inverter / batteries ...
Selecting cable cross-section, selecting safety equipment ...
Selecting what equipment to connect to the UPS output ...
I'll get to this ASAP
How my setup is going ...
These are some notable things I've discovered / learned about my solar setup (subject to change with experience) ...
When the batteries reach 20% charge whilst discharging, the battery output is lowered to 0 W to preserve the DOD that was set up on the app during commissioning (Normally 80%).
Although the batteries have 2.4 kW they have a buffer of 0.2 kW which means they have 2.2 kW of usable power. Since the inverter limits the DOD to 80%, that means they each have only 2 kW of output giving a total of 6 kW for my 3 battery setup.
The UPS output of the inverter should be direct to the protected loads (both phase and neutral) otherwise there will be RCD tripping due to the output having different harmonics than the Main output / grid .
In the General mode, batteries will only charge from PV (not from the grid) but they will start charging as soon as the PV starts and has enough output.
If you are not injecting into the grid (which requires bi-directional power meter, additional licencing / fees / and authorizations) and your consumption becomes less than your production, your PV output will be much lower than expected, because, since you're not exporting, the inverter doesn't need to produce more power than required for your consumption.
To be verified, but the PV Power on the SEMS site is the total of W of the panels (i.e. 8 x 440W = 3520) and not the Isc x Vsc of the panels.
Even rainy / cloudy days will produce output, but don't expect your batteries to charge at all. My system on a stormy day produced 3.5 kWh of a theoretical maximum of 17.6 kWh, which to me was great. But remember, due to your location, tilt, weather, panels, inverter ...etc, your mileage will vary.
Bad things may happen, so plan for the future ... I have insurance for the panels, so if they should happen to fling off of the supports, I have insurance for that ... maybe you should too ... Most home insurance policies will have optional provisions for this, ask your insurance company for a quote and make sure of what is covered in your policy (i.e. if your panel flies into the street, or into your neighbours' house ...).
The SOC on the SEMS apps/website will jump from ~89% to 100% in just 15 minutes (which is too fast), but the explanation for this is (from https://powerforum.co.za/topic/4866-pylontech-soc-jump-while-charging/) : "SOC is always a numerical estimation of a physical system, and there is always a bit of drift. It is normal for the SOC to be off by some margin, and for a "reset" to cause a jump at the top. The one thing we know for sure about batteries is that 1) the voltage rises and 2) the current drops when they are full, so most battery monitors will look for that event, and then reset, which corrects any drift there might have been since the last time the battery was full. These jumps will be smaller if the battery is regularly charged fully."
Satisfaction. I've had the system running since March 9th 2021, and so far I'm very satisfied that everything is up and running correctly, and I'm already starting to save on my bills, and I have more peace of mind that I won't be losing hours of work due to grid failure thanks to the UPS function.
Cost / profit. This system was expensive (for me), and will probably take a long time to pay for itself (at least 8 years). I'm applying for government subsidies at the moment 25% for the batteries from the local government, and a new incentive program should be available from the federal government that subsidizes up to 70% for the rest of the installation this month or the next (April). If both these subsidies apply, I should be able to pay off the entire system in about 4-5 years.
Is this for you? This system was planned and installed in a relatively sunny island in the Atlantic. But the reasoning and calculations should work anywhere. I am not directly sponsored by any of the brands mentioned here nor do I represent them. There are MANY different approaches. manufacturers, and products on the market to create a system like this. Always check your local / regional laws and zoning regulations to make sure you have or can get permission / licencing to put solar on your home. Also check with your grid operator to make sure there isn't any technical / legal impediment to you connecting any type of generation equipment to the grid! My advice is to research A LOT, ask for quotes from reputable professional local installers especially if your not comfortable with the manual labour involved, and don't under any circumstances do ANY of the electrical wiring unless you know what you are doing ... even if you're a competent electrician and are accustomed to working with AC, DC is ALOT more dangerous and possibly lethal when it comes to the Amperages and Voltages involved (my panels at full power will produce ~400V @ 11.16A or 49.8V @ ~89.28A DC), this is not a 9 V battery! It actively wants to kill you! Hire a professional! And finally, This is not a single person job, (at least the panels aren't), The panels I have weigh about 25 kg each, that, while not very heavy, are almost impossible to hold in position and fasten to the support structure with just one person. Ideally you will have at least 3 people, one that will tighten the fasteners, and two that will position / hold the panels. I recommend 4 people.