MULTIPLUS 2 - PV AC OUT - MultiPlus/Inverters - Question: Battery full - Grid OFF - PV production?

DIY
, , , , ,
1

MULTIPLUS 2 – PV installation on AC OUT without a Modbus connection between the MultiPlus and the inverters – 15 kWh batteries – What happens if the battery is full and the grid is OFF? How will the MultiPlus respond?

An earlier topic with a technical solution that was ultimately found by Victron (thanks, Thierry): MULTIPLUS 2 - II - 48 - 8000 - non déchargement la nuit quand batterie pleine ... PV sur AC OUT

Hello everyone,

Configuration: MULTIPLUS 2 – 48 V, 8000

AC OUT: connection to an existing PV installation at the customer’s premises: 5,000 kWp (Sunny Boy 5.0) + 3,000 Wp via a VTAC. Surplus export installation, although this is not the priority because the feed-in tariffs are low (€0.05). Two current-measuring clamps for the solar production.

I have heard that only AC coupling with FRONIUS and EMPHASE is guaranteedA blue VICTRON measuring clamp on the grid connection.

Customer’s objective: to have an islanding system with PV, to optimise their PV production more effectively using battery storage and, in the event of a grid failure, to allow the solar installation to restart and charge the battery while islanded.

QUESTION: Without a connection between the MULTIPLUS 2 and FRONIUS inverters, or using SMA’s Modbus, the following critical situation could occur:

  • Battery full / grid switches OFF / PV production available (full sunshine at midday), with 5,000 W of instantaneous power.

→ I have been told that the risk is that, since the MultiPlus cannot stop the inverters (because there is no communication link with them), the inverters will continue producing and damage the MultiPlus.

Thoughts: this summer, when the demand for power is low, the batteries are full and production is high, one of the inverters sometimes cuts out (the one that is more sensitive to the maximum voltage) because the voltage reaches 248/249 V.

- → So, if the situation arises—battery full, the grid fails and switches OFF—the MultiPlus 2 switches to island mode, but with PV production being too high, I believe the overall voltage will rise and the inverters will inevitably enter safety mode in accordance with EN 50549?
The risk would be that one inverter does not comply with the standard and fails to enter safety mode…

What do you think? Am I missing something in this reasoning?

Thanks for your feedback. Fabrice**

2

How is the Multi configured (menu item “PV at output” enabled)?

If so, in inverter mode the MP shifts the frequency, causing connected controllers to throttle or shut down.

3

The inverters (SMA and VTAC) already present in the client’s installation are not connected to either the CERBO GX or the MultiPlus for control.
From memory, PV on the output is indeed enabled, but the MP cannot tell the inverters (SMA Modbus has not been set up and there is no communication with the VTAC) to limit their output or switch off.

But my reasoning was more along these lines: if the battery is full + the grid is OFF + PV is generating, the voltage on the line will inevitably rise because there will be too much power being generated, won’t it? And therefore, if the voltage rises, the inverters should go into safety mode, shouldn’t they? (voltage >253 V cut-off threshold)

4

no, it’s the battery voltage that will rise, and everything will fail due to battery overvoltage

without comms, grid-tied inverters have to be controlled by frequency shifting

5

Hello, thanks for this information. I thought I would have to go through Modbus or something similar, but this frequency-shift solution interests me.

Are there any recommendations from Victron? If so, I’ll look through the MultiPlus documentation first.

I’ll see how to do this for SMA and VTAC, as specific access to the inverters may be required, but here is the information I found quickly:

  • enable/adjust the P(f) reduction curve (gradual reduction above 51.5 Hz, disconnection around 52.3–53 Hz),
  • limit the maximum active power (useful for complying with the 1:1 rule with the MultiPlus II and the VTAC),
  • set the reconnection delay after the grid returns.

But it seems extreme to start at just 51.5 Hz, since the grid fluctuates between 50 +/- 0.1… so I’m wondering whether it might be better to start at a frequency such as 50.5 Hz?

Wouldn’t it be safer to disconnect at 52, or even 51 Hz, to maximise safety?

I’d be interested to hear your opinions, as this type of setting seems highly relevant for adapting installations at the end of an EDF OA contract into systems with batteries and PV on AC out.

fabrice

6

@LASEICHE

the P(f) function is defined by the grid code. In Germany the AR-N 4105 from 2018 is relevant. As the European grid is connected, same rules must be implmented in other local guidlines.

AR-N 4105 defines:
Above 50.2 Hz producers shall limit their production grudalarely to 60% at 51.5 Hz.
Above 51.5 Hz producers shall stop producing.

If your inverters are not older than 2018 or configurable, they follow this rules for certification.

As you want to have the option to connect to the grid, all inverters/producers must activate a compliant grid code and follow this rule.

For the case of grid loss, you can configure the MP2 to use the same ramp down for P(f) as the grid. The MP2 will first increase to 53 Hz to shutdown the inverters on AC_out and than lower frequencey to make them restart.

AR-N 4105 also defines ramp up times. Consequently the MP2 has time to find right balance between actual load and throtteling the inverters while battery is on 100% SOC.

If your system is pure offgrid (or charged via seperated charger to the DC bus), than you can remove the grid code from all inverters and configure your own P(f) values. On Fronius the MG50 (micro grid 50Hz) is such a case, where the frequency range is extended for smoother control. But this is not allowed while beeinng connected to the European grid.

MP2 is configured via VEconfigure.
With current firmware the ESS assistant includes the PV assistant.
On offgrid systems you need to load the PV assistant (without ESS).
I hope this helps a little.

The throttling vai Modbus is used only while conneced to the grid. Purpose is limit feedback to the grid.

7

It’s probably similar in France; take a look at the grid codes.

The associated grid password isn’t publicly available. For this, contact your dealer/installer, Fabrice.

9

@LASEICHE When setting this up test the frequency shift works and what range you want it in, some people set it up with the default but I found it didn’t work quite to my liking and reduction was too close to the normal 50Hz, I played around with the figure whilst at 100% and in the summer with the grid disconnected so I could ensure how it would operate when I wasn’t there.

10

Thanks for this feedback.

As the house’s entire electrical installation was already in place before the MP2 was installed, both inverters and all the loads are backed up by the MP2.

The house is connected to the grid, with both inverters operating in self-consumption mode with surplus export.

PV1 uses an SMA Sunny Boy 5.0 from 2016, with 5,000 Wp of panels and surplus export, so it complies with the standard applicable at the time. Due south, 10% pitch.

PV2 uses a 2025 Vtac with 3,000 Wp, due north, 10% pitch.

I assume both comply with the standards, but I’ll need to check the SMA Sunny Boy 5.0.

If I understand correctly, the installation is ultimately protected by frequency shifting: if the grid goes offline and the battery is full, the frequency will naturally increase, causing the inverters to shut down for safety reasons?

I’m unsure whether I ultimately need to configure not the inverters (just check the SB 5.0), but the MP2, because you wrote: “In the event of a grid failure, you can configure the MP2 to use the same reduction ramp for P(f) as the grid. The MP2 will initially increase the frequency to 53 Hz in order to stop the inverters on AC_out, then it will reduce the frequency to restart them.

The AR-N 4105 standard also defines the ramp-up times. Consequently, the MP2 has enough time to find the right balance between the actual load and limiting the inverters when the battery’s SOC is at 100%.”

But since it has been configured for the European grid, isn’t this ramp configured automatically? There’s nothing preventing me from checking via VEConfig.

So, in the end, with inverters that were initially connected to the grid, plus the MP2 now configured for Europe, does the installation appear to be protected by frequency shifting by default?

I look forward to your replies, and thank you for sharing your knowledge @TherealKT @Karl-Heinz @BjoernK
@Daza, I’m finishing my reply after seeing your feedback. Which country are you in? Are you also using AC_out? What configuration and frequency-shift curve did you set up (on the MP2, on the inverters, or identically on both)?

fabrice

11

For an introduction or further information

Good luck,

12

I’m in the UK and have 2x 3kW Solis S6 inverters on AC out (Multiplus2 10kVa), with a third soon to go onto AC out, I found the multiplus could balance better with it on AC out rather than AC in plus gives you AC solar without the grid input meaning island solar operation if the grid fails.

Can’t remember the frequency that I set mine to but I disconnected the grid and started to play around with the frequency changing the normal frequency which I increased to see what affect it would have on the Solis inverters, PS make sure your battery is 100% as this will affect the frequency shift as far as I’m aware. I know it was a lot of rebooting as you upload the New ESS configuration just be patient with the adjustments that you do.

13

Okay, same on my side, but that’s why I’m going to make islanding secure when the battery is full and the PV is generating. I need to look into frequency shifting. Thanks.

14

PS change the acceptable ie the one that it normally uses in a normal situation to the figure that you want it to start to react to, I raised the one that was normal frequency up, so I knew only the figure that I was changing in shutdown mode was affecting the frequency shift I started this below the standard shutdown figure to see if it was a 100% shutdown and lowered it to see what figure I needed for 100% shutdown then wrote it down and then started to change the normal frequency as it was cutting some production which was very close to normal grid frequency so changed that as well so my frequency shift starts later.

it helps if you’ve got like a VM3 or something that can track frequency so you can see that it’s hitting your frequency.

On another note do you have energy meters reporting to the Victron what power the solar is producing?

15

I have just carried out a real-world test by operating the inverters at minimum power (1 string out of 2 activated for the tests). I performed the test (2 protective shutdowns) with only the SMA inverter, using 1 string out of 2, followed by a third test in which the VTAC was shut down, again with 1 string out of 2.

Situation: Battery at 100% or nearly so, grid disconnected to operate in island mode.
Test with the SMA inverter, using only 1 of the 2 strings. Instantaneous power during the test: 700 Wp
Two successive frequency increases occurred. During these two increases, the SMA SUNNY BOY 5.0 inverter
stopped, and between the first and second protective shutdown, it gradually increased its power, then
cut out again / went into protective shutdown at a frequency of 51.5 Hz.
Cell voltage during the critical phase: 3.50/3.55 V on the MP2
Battery voltage before the SMA shutdown: 53.01 V on the MP2
Between shutdowns 1 and 2, there was a 500 W load consumption, resulting in a slight battery discharge;
the frequency fell back to 50 Hz and the batteries were at 50.84 V.*

*Approximately 6 minutes before the inverter restarted production, between Safety Tests 1 and 2.

Here are some curves extracted from the VRM.

As things stand, the system appears to be responding well: the SMA cut out at a frequency of 51.5 Hz with battery voltages of 53 V, while the VTAC cut out at the normal frequency, but the voltage at its terminals may have been around 248/249 V, which would explain the protective shutdown. I need to repeat the test for the latter and check the instantaneous voltage to determine whether this is what caused it to shut down or whether something else was responsible. However, it is much more sensitive than the SMA and cuts out before the frequency rises.

QUESTION: **Are there any important values or pieces of information you would like me to check to make sure that nothing may have been damaged by this test?
*

SMA and VTAC Safety Test.pdf (155.5 KB)

16

Yes, both solar installations have a current clamp connected to AUX & AUX2 on the MP2, allowing it to read the instantaneous power and adjust its behaviour accordingly—particularly for car charging with the Victron charger, which adapts to the solar production.

17

Nice :ok_hand:t5: the voltage seems high but it’s withing in France’s tolerances of 207 to 253V, did you start to see a decrease in the VTAC inverter or did it just shutdown?

Yes inverters from different brands seem to operate at different frequencies that’s why I try and keep them all the same make and model.

I would try the VTAC that went into protection again use the same shutdown you have now if it continues up it, but if it bleats about protection back it off slightly maybe the frequency tolerance is more sensitive hence why it goes to protected state.

But does it start back up on its own? If so could be a quirk with that inverter and I would consult the manual. But as long as both are shutting down when battery is at 100% then I would say all’s good, just check out the protective message in the manual

19

@LASEICHE

for 8kWp PV on ACout you should plan 24kWh battery.
“Per 1,5kWp you should have 4,5kWh battery” for offgrid usage.

If the control is not smooth enough, e.g. the SMA can only switch off and not throttle, substituting by an RS450/100 might be an option.
Victron PV chargers get throttle very smoothly and it enables black start.

I have 7kW Hoymiles PV inverters on AC out which get throttles quite rough by frequently. But my MPPT 150/45 with 1.7 kWp get controlled quite fast and fine.

20

We need to get over capacity-based rules of thumb.
Capacity matters but what is more relevant is the ability to sink or deliver power - dcl, ccl, transient ratings.

You can have the biggest battery in the world but if it cannot provide or absorb the necessary current, it’s pointless.
Manufacturer sizing is based on this, not Ah or kWh.

21

There, I think it was the voltage across its terminals that caused it to limit its output and enter protection mode. There was no frequency variation at the MP2 terminals, just a rise in the battery voltage.

It’s the furthest installation on my network, and I know that in summer, when there is little load and the batteries are fully charged, the voltage across its terminals can sometimes be higher than the grid voltage … and it shuts down. In those cases, I disable one of the two strings to limit the power.

I’ll run some more tests and keep you informed.

22

The aim was to see whether, in a situation where the grid is lost and the battery is full, the system would be at risk: that’s what I was told … but I couldn’t understand how that would work physically. Even if these are sudden outages rather than gradual ones, I’m thinking that at least the system won’t deteriorate… I’ll optimise the inverters if anything changes.

I’m grid-connected, but with an MP2 to optimise self-consumption and ensure security of supply.

Thanks for your feedback