Multiplus II Grid connection and grid code issue – Australia

DIY
, , ,
1

Hi everyone,

This might be a long post to convey all the information needed. For disclosure, I am a licenced electrician, however this Victron system is a hobby for myself, and my day-to-day job does not involve installing and commissioning them.

The installation and backstory:

  • 3x MPII 10k in three phase 230/400V configuration.
  • DC coupled solar only (2x MPPT 450s)
  • Ekrano
  • Battery Shunt
  • Custom LifePo battery bank – BMS’s currently not wired / configured to talk to Victron system currently.

The wiring configuration is not ideal given the location of the MPs in relation to the incoming grid / MEN connection and the outgoing circuits from here. It is not laid out how I would prefer it to be, but given existing infrastructure, limited room in the main switchboard etc I designed it this way.

A basic description: The house contains the main switchboard where the DNSP termination and meter is located. From here a grid supply to ‘Shed 1’ is provided where the system is situated. The MP’s then supply their inverter output to a change-over switch where the entire properties loads are supplied from (either the MP’s or grid). After the change-over switch I have three main submains: A 3phase supply on a separate cable is sent back to the house to feed all the house circuits, A 3phase supply to ‘Shed 2’ and a 3phase supply the distribution board for local loads in ‘Shed 1’. The Victron power meter is situated in the main switchboard adjacent to the retailer meter.

In Australia, we can only have 1x MEN (Neutral-Earth Bond) throughout the installation which must be located at the main switchboard where the DNSP terminates.

This means that the MPs are down stream of the MEN and in my case, they are in another building.

We can have multiple MENs under certain circumstances and installations but for the sake of this discussion on a residential property, there is only 1x allowed.

I will provide a drawing of how it is laid out and wired and post it shortly.

The system has been working very well completely off grid for a year now with no issues.

Current story and issues that unfolded last night:

Coming into Australian winter it was time to get the system onto the grid for when solar input is insufficient, especially now that I have the entire property running off the MP’s inverter to be off-grid as much of the year as possible. With priority of self-consumption from solar and the battery, only using the grid in certain conditions:

  • SOC <15%
  • Charge batteries to 90% during a daily time slot for 'cheap or free power from the grid
  • Never feed into the grid

I Setup ESS assistant in each of the MP’s. with the below settings:

  • System uses LiFePo4 with other type BMS
  • Battery Capacity 1120Ah
  • Sustain voltage 50.00V
  • Dynamic cut off – 0.005C – 50.00V, 0.25C – 49V, 0.7C – 48.2V, 2C – 47.5V
  • Restart offset – 1.2V
  • PV inverters – No

MP main settings

General

50Hz, Shore Limit Input – 40A. Overruled by remote - unticked, Dynamic Current limiter - Unticked, External current sensor connected – Unticked, Enable battery monitor – Ticked, SOC when bulk finished – 95%, Battery Capacity – 1120Ah, Charge Efficiency – 0.95

Grid

Grid code: Australia A: AS/NZS 4777.2:2020 Standalone App. M (no N-bypass), AC Input related settings – Above selected gridcode plus LOM B (compliant)

Inverter

Inverter output voltage – 240V, Ground relay – Ticked but greyed out, DC Input low shut-down – 47V, DC Input low restart – 51V, DC input low pre-alarm – 48.5V, Do not restart after short circuit – Unticked, Powerassist – Unticked, Shut-down on SOC – Unticked, Enable AES – Unticked

Charger

Enable charger – Ticked, Weak AC input – Unticked, Stop after excessive bulk – Unticked, Lithium batteries – Ticked, Configured for VE Bus BMS – Unticked, Disable VSense – Unticked, Charge curve – Dropped current to 75A.
Virtual Switch – disabled

Assistants - ESS

Once I believe I had the settings correct, I switched on the grid to the MPs, they began chattering their relays several times in sequence then appeared to sync to the grid for a brief period of time before disconnecting again. System then appeared to attempt re-sync every 1-minute which lines up with ESS settings.

Looking at IP issues log, input voltage was flagged as too high on the unit which I found bizarre as it is verified as acceptable.

Looking further into it, I noticed that the grid code was set as Australia A AS4777:2020 M (No N-Bypass). Now as I do have a physically wired Neutral bypass where the MP input and output neutrals are tied together, I thought this was odd and that I must have selected the wrong grid code when initially setting up the units and that this might be the reason the MPs are not syncing to the grid.

Changed grid code to Australia A AS4777:2020 M (With N-Bypass) after finding the correct password and sent the change to the units. Upon the MP’s powering up again I experienced some very odd, and unfortunately damaging behaviour. When they powered up a ‘surge’ for lack of a better word or understanding has occurred at the house damaging certain equipment. The MPs were on for about 5 seconds before I immediately isolated the output isolator from the MPs after I noticed some of the outdoor lights rapidly flickering, and upon further inspection damaged smoke alarms in the house and an RCD on L3 in shed 2 had tripped.

The damage: Multiple smoke detectors in the house were damaged, along with 1x or more external PIR light detectors and several phone chargers.

I am yet to investigate this in detail and understand the damage – I have a feeling that there might be a phase cross issue across the lighting circuits where the smoke detectors are on separate phases (and RCDs) yet have an interconnection wire between them.

NOTE – The grid input to the MP’s was isolated at this time. Grid was energised and live to the meter board (at the house) powering only the Victron Grid meter and Grid surge diverter. It is worth noting that the grid Neutral is still ‘live’ and physically connected to the MPs input and output Neutral – All neutrals throughout the property must be permanently connected with none switched as required by Australian Wiring rules.

I changed the grid code back and resulted back where I was, albeit with some device damage.

Looking at the settings again today, it appears that the ground relay is still in used – it is ticked but greyed out, my understanding is that the MP should disable the ground relay when using an Australian grid code where the neutral earth bond (MEN link) is done as a permanent link in the main distribution board.

I need to get some guidance on what issue(s) I am potentially facing here. I completely understand I may have a hidden wiring issue and/or entered incorrect settings. However, as I am in a steep learning curve on this I may miss considering other aspects.

What I believe is wrong in order of likelihood:

  • Ground relay not disabled
  • Configuration / setup
  • Imbalance / harmonic issue between the main switchboard and the MPs (Grid Neutral and MP Neutral are essentially shared)

Questions:

  • What do the two grid codes specifically do to the Multiplus input / outputs? In some way, changing the grid code has caused a short circuit or surge of some description from phase to neutral or even phase to phase.
  • Ground relay operation, I need to ensure this is completely disabled under all circumstances, whether on grid or off grid.
  • What may I have configured wrong?
  • What should the correct operation of syncing to the grid look and sound like? Logic says to me that the MP grid relays should all close at the same time upon syncing and that’s it.

I will start doing some basic checks now:

  • Incoming grid voltage
  • Incoming grid phase rotation
  • Incoming grid phases are wired to the correct MP L1-L1, L2-L2, L3-L3
  • Neutrals are connected and all tied together.
  • MP Active and Neutral input not reversed.

Regards,

2

Update - I think I have solved it.

Decided to start from square one and check every connection, including the connections between the main switchboard (meter) and the switchboard in the shed where the MP’s are. After finding that all the above board connections were good, I started looking into the earthing and neutral connections - I couldn’t get continuity between the neutral and earth bus bars in the switchboard (this should have come back with an impedance of well under 1 ohm as both the neutral and earth are linked by the MEN point in the main switchboard - this stumped me for quite a while.

Looking further into it, I found that the wrong neutral conductor had been connected, essentially meaning no neutral at all between the two switchboards. This was done either by myself during installation or likely in the past few weeks where I had the grid service changed from single phase to three phase. I can see where there was some confusion on the neutral labelling.

So what was occurring; The system had been operating quite normally, a few quirks which make more sense to me now that I understand the problem at hand. As it is a three phase system, the neutral return current is fairly low when the phases are fairly balanced, so missing a neutral and having that small amount of current return through the earthing system didn’t raise any immediate alarms - if it was a single phase system you would know immediately upon power up that something was not quite right. This issue was very well masked by the fact I had the wrong grid code selected in the MP (No Neutral Bypass) which enables the ground to neutral relay. As there is no MEN intentionally installed in this distribution board, this was the return path for the neutral currently.

Current path: From the loads in the main switchboard (house), neutral return current passed from the main neutral bar, through the MEN into the main earth bar, then into the earthing conductor to the shed switchboard earth bar, then to the MPs via their earths and into the ground relays which are tied to the neutral.

The surge event: Upon updating the grid code from without N bypass to with N bypass meant that the ground relay is now disabled. This meant that the neutral return path to the MP’s was now lost resulting in 230/400V supplying single phase devices with no neutral. Upon the MPs powering up again this was evident with the exceptionally weird behaviour and damage to some equipment as they no longer had a viable neutral return path. I believe some devices may have received 400VAC where a return path was trying to be made.

The loss tally looks to be; 5x smoke alarms in the house, 3x phone chargers, an outdoor PIR sensor and one of the bathrooms 3in1 units (its one of the remote control types, not hard wired). I am adamant that the surge protection devices I have played a significant role in it not being worse.

Moving on; I quickly put the grid on the MPs and they accepted the grid and began sustaining the batteries. I’ll need spend the weekend playing around with the settings and ensure it behaves as I want.

Regards!