Greetings, 

I'm trying to figure out what could be encumbering production of a grid-tied installation very near the equator. The maximum power generated under a clear blue sky is 8.5kW (55% of capacity) against a variable torque motor load of 15kW driven by a 18.5kW VFD at 50Hz. The balance 5.5-6.5kW required is purchased from the grid.

Particulars of the Installation

• Inverter - Deye SUN-25K-G03 three-phase string inverter with 2 MPPT trackers running the latest firmware
• Panels - 14s1p Jinko JKM555M-72HL4 modules (7.77kWp, 696Voc, 574Vmp) on each MPPT, totaling 15.54kWp
• Orientation - 5 degrees tilt angle from the horizontal, 45 degrees East clockwise from magnetic north
• Cable run - 60m 4sq.mm PV1-F cable from the furthest panel on the rooftop to the inverter in the basement

There's an Eastron SDM630MCT 3-phase RS485 Modbus RTU smart meter installed for zero export management

The DC voltage displayed on the inverter when the motor is off is roughly 675Vdc which then drops to around 600Vdc when the motor is running. Would increasing the string voltage to 745.8Voc and 614.85Vmp by adding a panel to each string (15s1p) improve the overall efficiency? The maximum DC input voltage of the inverter is 1000Vdc.

Edited August 20, 20232 yr by Kilowatt Power

Or could the Grid Standard be throttling production? Which Grid Standard allows unlimited inverter power output for 240/415Vac distribution systems?

The Grid Standard is currently set to IEC61727. Other available options are:

INMETRO
EN50549
EN50438
IEC61727
CUSTOM
VDE_AR_N_4105
UTE_C15_712_1
RD_1699
CEI_0_21
G98_G99
AS4777
NBT-32004
AS4777.2 Australia A
AS4777.2 Australia B
AS4777.2 Australia C
AS4777.2 New Zealand
MEA
PEA
Norway
Switzerland
R25
 

24 minutes ago, Kilowatt Power said:

Or could the Grid Standard be throttling production? Which Grid Standard allows unlimited inverter power output for 240/415Vac distribution systems?

The Grid Standard is currently set to IEC61727. Other available options are:

INMETRO
EN50549
EN50438
IEC61727
CUSTOM
VDE_AR_N_4105
UTE_C15_712_1
RD_1699
CEI_0_21
G98_G99
AS4777
NBT-32004
AS4777.2 Australia A
AS4777.2 Australia B
AS4777.2 Australia C
AS4777.2 New Zealand
MEA
PEA
Norway
Switzerland
R25
 

The grid standard does not specify the maximum size . It normally just sets the limits on min and max volts at which point the inverters must disconnect. The standard also sets the frequency limits of operation. That is normally set by the grid supply authority. By using more panels mean you are increasing the voltage so for the same load the amps will be lower. Further more panels as indicated would just mean less power from the grid. 

Not sure which grid standard one would apply in the country where the system is installed but don't see a problem using the ZA grid standard or the EU standard. AUS also fine but I think they are stricter on the volts and freq limits. Their standard might also include a system where they have control over the inverter to switch it off if grid ties are moving the network into an unstable region. 

G83. 3 or G100 standard should  also be fine I think. Have not looked at other standards. 

My 2c on what I can remember which at times is dangerous. 

 

Edited August 20, 20232 yr by Scorp007

@Kilowatt Powerwhat time of the day in Nairobi are you carrying out this load? Time of day is very important here? You seem to have a good tilt Angle for your latitude. Near the equator close to zero or even 5° is good. However, your facing North East. So not great but time of day is important. 

My 2nd thoughts go towards the Meter: Is this installed correctly & is it given its own Modbus number & setup correctly?

Edited August 20, 20232 yr by Steve87

I’m not sure how the efficiency relates to the string current. If I was looking at the correct datasheet for that inverter, max DC current is 30A. The max current on those Jinko panels is around 13A. In reality they will produce probably around 10-12A. Way below the max. 
 

Not sure, if it’s not going to be better to reconfigure the strings 7s2p and bring the current higher and closer to the max MPPT spec. 

16 hours ago, Kilowatt Power said:

Would increasing the string voltage to 745.8Voc and 614.85Vmp by adding a panel to each string (15s1p) improve the overall efficiency?

Yes for sure it will but like @Steve87said the time of day you do measurements is key.

16 hours ago, Kilowatt Power said:

Orientation - 5 degrees tilt

So close to the equator the panels could face straight up but 5° is good assume a problem to change to 0° maybe just for a test run.

What is the surface temperature of the panel face when you run this load might also be affecting panel efficiency.

Edited August 21, 20232 yr by TaliaB

Pure speculation, but could this be moving into the realm of active power curtailment in the interest of grid stability, considering that your grid-tied installation is basically consuming reactive power only? Could it be that either the addition of a battery or power factor correction could improve the amount of self-generation? Again, just thinking out loud.

On 2023/08/20 at 8:56 PM, Steve87 said:

what time of the day in Nairobi are you carrying out this load?

The 15kW load runs from 8am to 4pm. Maximum solar production is at 12 noon when the sun is directly overhead.

On 2023/08/20 at 8:56 PM, Steve87 said:

You seem to have a good tilt Angle for your latitude. Near the equator close to zero or even 5° is good. However, your facing North East

North East happens to be the building azimuth angle. I would however want to believe that direction becomes less significant as the tilt angle approaches 0°. I read somewhere that  a horizontal roof has a tilt angle of 0° and an azimuth angle of 0°

On 2023/08/20 at 8:56 PM, Steve87 said:

My 2nd thoughts go towards the Meter: Is this installed correctly & is it given its own Modbus number & setup correctly?

I want to believe so. It's correctly displaying V, A, %THD, Hz, PF, kW, kVAr, kVA, kW and import/export kWh and kVArh. Most importantly it's communicating with the inverter and preventing grid feed-in. Shall peruse the manual once more and go through all settings just to be sure.

The only way to know if these panels are working 100% is to go to the System work mode timer & tick Solar export. That will open up the tap & you will see the full potential of the PV Array. I know you have no permission to export but a brief moment for testing will be ok & if the grid becomes unavailable the inverter will island by design by the fact the Grid Port will be dead. Look under the multi inverter page. You will find the option that allows for a meter to be added instead of the CT coil. Then you also need to give that meter a  Modbus number. Choose any Modbus number less than 17. Deye only allow up to 16. If on this page you have CT as an option then this is the issue. The meter is a bit dumb. It will read the correct metrics because it's a meter. But what is most Important is that you tell the Deye that you don't have a CT installed & that you have an Eastron meter. The meter will need a Modbus number that does not conflict with the inverters preset number. The Eastron within it's menu settings will have a Modbus assignment that can be changed. The way to verify that there is meter Comms established is that the Multi inverter phase mist have the Eastron meter as ticked & the Fault codes will not display "Meter Communication Error". I forget the Fault code associated. 

See attached pages from the Eastron meter. I think it comes preset as Modbus no. 1 so if the inverter is also number 1 then you have an issue. I would definitely have a look at the Meter setup before you look elsewhere. 

Stupid question: The 15kW Load, is this wired on the essentials load output or under non essentials?

Edited August 21, 20232 yr by Steve87

On 2023/08/20 at 11:28 PM, PowerUser said:

I’m not sure how the efficiency relates to the string current. If I was looking at the correct datasheet for that inverter, max DC current is 30A. The max current on those Jinko panels is around 13A. In reality they will produce probably around 10-12A. Way below the max. 
 

Not sure, if it’s not going to be better to reconfigure the strings 7s2p and bring the current higher and closer to the max MPPT spec. 

I stumbled upon an old Deye Inverter Catalogue with the below block diagram and efficiency curve for SUN-20/25K-G02

EDIT: Looks like 660Vdc is the sweet spot even though the MPPT operating range is specified as 200~850Vdc. I don't know if the formula Vdc = sqrt(2) x Vrms applies to the DC bus as is the case with rectifiers.

Edited August 21, 20232 yr by Kilowatt Power

13 hours ago, TaliaB said:

So close to the equator the panels could face straight up but 5° is good assume a problem to change to 0° maybe just for a test run.

The mounting structure is fixed. The 5° tilt angle was to allow for rain to naturally clean the panels. Having said that, I'll get the panels hosed down once again to completely eliminate the possibility of dust curtailing production.

14 hours ago, TaliaB said:

What is the surface temperature of the panel face when you run this load might also be affecting panel efficiency.

Wasn't measured. Shall try getting readings with the clamp meter thermocouple probe at high noon.

34 minutes ago, Kilowatt Power said:

I stumbled upon an old Deye Inverter Catalogue with the below block diagram and efficiency curve for SUN-20/25K-G02

Looks like the higher the DC input voltage the better. I don't know if the formula Vdc = sqrt(2) x Vrms applies to the DC bus as is the case with rectifiers.

I know my suggestion was a long shot and the efficiency shouldn't be as low as yours. Thanks for that efficiency curve. Haven't seen it before.

Do you have some sort of a monitoring app which can provide a full day V/A curve for each MPPT? That should at least tell you if you are lacking on the volts or amps. Something like this Solarman graph from my SunSynk inverter. I have ran a full load for the whole day and your site is pretty much doing the same from 8AM to 4PM. 

 

8 minutes ago, PowerUser said:

Do you have some sort of a monitoring app which can provide a full day V/A curve for each MPPT? That should at least tell you if you are lacking on the volts or amps.

Shall be able to share tomorrow's graphs now that WiFi is available in the basement.

1 hour ago, Steve87 said:

The only way to know if these panels are working 100% is to go to the System work mode timer & tick Solar export. That will open up the tap & you will see the full potential of the PV Array. I know you have no permission to export but a brief moment for testing will be ok

The 3-phase string inverter doesn't have the exact menu options as the hybrid but enabling unlimited export is still worth a try. I doubt if the utility meter will enter into tamper mode as a result. The worst that can happen is for the meter to count exported energy forward. 

1 hour ago, Steve87 said:

The meter is a bit dumb. It will read the correct metrics because it's a meter. But what is most Important is that you tell the Deye that you don't have a CT installed & that you have an Eastron meter.

The string inverter has options for a limiter or smart meter. Opted for the Eastron CT smart meter.

1 hour ago, Steve87 said:

See attached pages from the Eastron meter. I think it comes preset as Modbus no. 1 so if the inverter is also number 1 then you have an issue. I would definitely have a look at the Meter setup before you look elsewhere. 

Thanks. 

1 hour ago, Steve87 said:

The 15kW Load, is this wired on the essentials load output or under non essentials?

SUN-25K-G03 is a grid-tied inverter. Everything is on the grid side, after the main incomer MCCB. 

I did not realise this was a pure grid tied model. Makes total sense now. There is no split of circuits as there is only a single Grid port that is bi-directional...This particular machine probably no person here is SA will have experience with as our main requirements are Hybrids as we have a poor and unreliable grid for periods. But i hope you solve the puzzle soon. 

More stupid questions because i know this meter very well and have setup a Live logger making use of the exact same meter...Are you making use of the CT Coils ? Are you sure that the metrics are correct in terms of the CT direction as well as polarity? send a pic of the meter in action as well as the kW values for each phase...

6 hours ago, Steve87 said:

I did not realise this was a pure grid tied model. Makes total sense now. There is no split of circuits as there is only a single Grid port that is bi-directional...This particular machine probably no person here is SA will have experience with as our main requirements are Hybrids as we have a poor and unreliable grid for periods. But i hope you solve the puzzle soon. 

I have often wondered why the string inverter is not sold in ZA seeing that the hybrid is so popular. 

1 hour ago, Scorp007 said:

I have often wondered why the string inverter is not sold in ZA seeing that the hybrid is so popular. 

Scneider CL33 is avaliable in Sa in 30/33/50 kw models dc input voltage 200-1000v with 3 MPPTs. I have quoted on the 33kw model for a  OK Bazaar store(Shopright holdings). As i have not got much experience in these string inverters i am doing the 5 day installation course soon with Group Scneider in Midrand. I think the decentralized string inverters is the future for commercial installs making a lot of sense. The goverment should look at decentralized models and start moving away from centralized power distribution having losses in the form of heat by distribution of power through lines 100th of kilometers from the source.

Edited August 22, 20232 yr by TaliaB

CL-30-33-50-Datasheet.pdf

11 hours ago, Steve87 said:

Are you making use of the CT Coils ? Are you sure that the metrics are correct in terms of the CT direction as well as polarity? send a pic of the meter in action as well as the kW values for each phase...

Yes. 3 CT coils are installed with the same polarity on each phase. Had to change the direction of all 3 CT coils when I realised energy was being exported to the grid. Shall get pictures when I next visit the site.