Adding more MPPTs to an inverter
First of all, I just wanna say how amazed I am by all the comments. You guys are incredibly smart and thorough, I never realised how useful this forum would be and I'm sorry for ignoring it for 3 days. Special thanks to @lvrasmoklet @jacekowski @Groots. Thanks to you guys I'm gonna rewrite this whole post to give you all the info you need.
Our system:
1 x Inverter: SYNK-12K-SG04LP3
21 x Panels: Jinko, (1000V) (550W)
3 x Batteries: VOLTA STAGE 1 - 5.12KWH
Commissioned: February 2024
Location: Cape Town, South Africa
Our roof is quite unnatural and therefore in order to fit 21 panels on it they had to be placed at different lattitudes and longitudes on the different sections of the roof.
However they are all at 60 degree inclines and there are only 2 different rotations.
5 Panels face North West (String 1) (purple on the diagram)
16 Panels face North East (String 2) (blue on the diagram)
Blue panels 11-16 are on a different section of roof that is less indented and lower than the top row of 10 panels, these panels experience the most shading due to other sections of the roof and cause the most trouble.
Situation:
Due to the unnatural nature of the house there are plenty of shadows from chimneys and other roof edges or trees that cause shading onto different panels at different times. It's currently summer, and in Winter the sun gets lower and the green circled tree will become even more of an issue.
We are not getting the power generation we think we should be getting from the number of panels that we have. Thanks to the comments this could be from one of two reasons:
The shading hitting Panels on String 1 is causing the voltage of the string to drop below the minimumm 200VdcIf the whole String is below that threshold of 200Vdc then inverter rejects it and treats it as 0Watts of power generation
The new panels that they just installed don't have bypass diodes and therefore are bringing the overall voltage of the string down when a little bit of shade hits one of the panels in the string.
Solutions:
Add SolarEdge Optimizers to the panels on the 16 Panel String 2. This would hopefully make them more effective and be cheaper than a new inverter. But I'd have to figure out how to wire them with the existing installation without the MPPT port on the inverter trying to undo what the Optimizer is doing.
Find a way to add 1 more panel to String 1 in order to make sure that if shade is over 1 panel on that string it doesn't go below threshold. 2.1) A panel can be rerouted from String 2 (which is at a different orientation). 2.2) A panel can be squished into the same roof section the other 5 panels are on, but would not be flush with the edges of the roof and need a robust mounting bracket (would be a wind hazard)
I thought the solution was to get more MPPTs on the inverter somehow, since it only has 2 MPPTs. This would allow the inverter to maximise the output of the panels the more I split them up. But I forgot to account for the minimum 200Vdc that the inverter needs per string. So I could only realistically add 2 more MPPTs to the panels on String 2 so that I have 4 strings in total. This would only increase my efficiency of String 2 Panels and wouldn't fix the issues with String 1.
When we complained about the inneficiency of the system; our installers told us the only way is to add more MPPTs to make the panels more efficient by buying another 12Kw 3 Phase inverter just to add 2 additional MPPTs. This would cost another R56 000 = $3000
Annexure:
As requested I have included the graphs you guys were looking for
Note: This specific house does not currently get hit by Loadshedding, so that doesn't affect the graphs
We are aware that more batteries are needed in order to go completely off grid over the night
Figure 1: 29th Feb Power Gen. According to this website there was clear sky this day high of 24°C
Side note: Why is there no way to see how much PV I am generating without me having to use it up... It's very frustrating that the inverter only measures the power going through it and not the potential power the solar is generating. Because if I had that graph, I could estimate how much electricity I can put back into the grid once it becomes legal to do so.
Figure 2: 29th Feb Current Graph
Figure 3: 29th Feb Voltage Graph
UPDATE ON FIXES
@lvrasmoklet @jacekowski @Groots, As promised (I admit, very late) the update on how the system was fixed:
Problem 1: Not generating as much as expected from the panels
Actual problem:
String 2 (blue) was actually 2 strings connected to 1 MPPT. It had 10 panels (String 2.1) and 6 panels (String 2.2) and they were connected internally within the inverter in parallel before being tracked by the MPPT.
Due to the way MPPT works it only considered the lowest power between those 2 strings when converting them. Since one of the strings had 4 panels more than the other string; those 4 additional panels were ignored by the MPPT.
It was now treating String 2 as a 6 + 6 parallel connection instead of as 10+ 6.
The fix:
Splitting the 2 strings within String 2 evenly.
We rewired the panels on the roof to be 8+8 on String 2 and got a better power output and less drastic drops when shade hits some of the panels.
Alternatives that couldn't be done:
Putting all 16 panels in Series instead of parallel was considered. But that would add up to more voltage than the 650V MPPT range.
Problem 2: Batteries taking forever to charge even though we have access to so much PV generation, therefore sometimes by the end of the day they wouldn't get to 100%.
Actual problem:
The batteries weren't setup correctly. One of the setup instructions for the 3rd battery weren't correctly done. The inverter doesn't actually see 3 batteries but instead treats all 3 as 1 big battery, but in this case it thought it was combining 2 batteries instead of 3 (Not 100% certain of my understanding here)
In the section where you enter battery info on the inverter: the battery was stated as 200aH instead of 300aH
For the charging threshold: it was set at 40A instead of 100A.
Solution:
Set our max battery charging to 63A and then increase it by 1A a week until the breaker trips. Then we will lose eskom supply, open the inverter with Allan keys and put the circuit breaker back on, lower the Amps by 2A and then leave it that way forever. (I was told I could do this 1A per week starting 5th April)
In practice, I forgot to do it, remembered it in June, set it to 65A and it tripped a few days later. In a panic I set it back to 63A, completely forgot about the warning they had told me so I called them to come fix the issue and it took 2 weeks of testing (Made harder by the fact that I unintentionally removed evidence of what I did to cause the problem) to finally determine that it was the internal circuit breaker (exactly like I was warned about in April).
Alternatives that couldn't be done:
I was told by the installer that the sunsynk inverter we have has a bug where if the Amps go over 65A then the internal circuit breaker trips and it stops receiving AC from grid. The batteries are rated for 100A, but unfortunately due to this bug it cannot be utilised.
Annexure A: Extract form inverter datasheet.
"MPPT Range (V): 200-650V.
No. of MPPT Trackers: 2
No. of Strings Per MPPT Tracker: 2+1."
Problem still to be solved in the future:
Battery overshoots the threshold and causes spikes in grid usage
Note: the battery SOC minimum is set to 30% whilst grid is available.
As can be seen in the image above. As the sun sets, the battery % slowly drops towards 30%. However it almost always overshoots and gets to 29%. Then the inverter remembers that the minimum is 30% so it uses the grid to charge the battery back up to 30%.This spike can also occur when the batter has been at 30% for a long period and due to natural phenomena it dips to below 30% and causes a spike in the grid usage.I don't think this is very good for the battery lifespan and I would prefer if there was a more relaxed threshold.
If the inverter could stop draining the battery to power loads at 30%
And the battery only charges from grid if it reaches 26%.
Then I wouldn't have to worry about this spike issue.
However it almost always overshoots and gets to 29%. Then the inverter remembers that the minimum is 30% so it uses the grid to charge the battery back up to 30%.This spike can also occur when the batter has been at 30% for a long period and due to natural phenomena it dips to below 30% and causes a spike in the grid usage.I don't think this is very good for the battery lifespan and I would prefer if there was a more relaxed threshold.
Unfortunately for you 3 different things will always prevent you from achieving that goal:
The inverter maths applied during control is only stopping loading/discharging the battery when the SOC is UNDER 30 (so whenever there is a value less than whatever your SOC setting is) and stops charging when the battery SOC reaches the SOC value set.
While the inverter may not eat your battery ito loads at that point, the battery itself IS consuming its own SOC in order to power the BMS and comms etc ................. thus SOC will slowly drop during "idle" use even if the inverter is not using the battery or doing anything to it other than to maintain its SOC (about 3% before charging is initiated again).
In practice your objection about the battery SOC consumption can be measured in a 24 hour period - I find that the battery eats itself to the tune of 0.3kWh per 24 hour period for a 5kWh battery if the inverter is not using the battery to power a load - iow you need that much to keep the SOC at whatever setting you chose.
Glad you are making progress though.
Would be interesting to see the current energy generation graph after the fixes to the PV system