2x Powerwall on 2-phase UK supply: sensing PW SOC individually

[TyneTes]

I would greatly appreciate it if anybody has the answer to a bit of knotty problem I am trying to sort.

SitRep. Had 1x Powerwall + 1x Gateway 1 + 1x solar array on a single phase UK electricity supply for about 2 years. All working fine.

Have recently upgraded to 2-phase grid supply with our DNO (Northern Powergrid). We are very rural, they do not support full three phase supply up here in rural Northumberland. Upgrade done to accommodate 2x Air Source Heat Pumps which will shortly replace our oil boiler.

Gateway 1 now upgraded to Gateway 2 Config is now split as follows.

The L1 phase has 1x 5kWp solar array + 1x PW2 + 1x zappi + core domestic loads, which are designated to be supported during power outages by the Powerwall on the L1 phase. In near future 1x ASHP will be added to this phase.

The L2 phase has 1x 5kWp solar array + 1x PW2 + non-core domestic loads, which are not designated to be supported during power outages by any Powerwall. In near future 1x ASHP will be added to this phase.

Both phases are controlled by the one, newly installed, Tesla Gateway 2 unit. The incoming grid supply and solar production are all being highly accurately recorded by the SMETS1 polyphase meter on the grid connection and the SolarEdge app recording inverter production. That is to say, the system is functioning entirely as expected. I'm generally happy with it.

The one thing I would like to be able to achieve is to individually monitor the SOC of each of the Powerwall batteries. The Tesla app simply aggregates total Power Flow and total SOC across both phases. It is doing this entirely accurately. However, I would like to know individually what energy each of the two PWs has stored at any given time, so that I (or a home automation system) can intelligently connect/disconnect loads to either the L1 or L2 phases, depending on the SOC of the specific PW on that specific phase.

Does anybody know a way of reliably doing this in the Tesla app? Or by some other means? Any advice gratefully received. Thanks in advance.

 

[jjrandorin]

The tesla app only aggregates everything. In order see what the individual totals of each powerwall has, you have to access the tesla api to get that information out of the system.

There is a sticky thread on that topic at the top of this subforum, here:

Discussion: How to get status information from your Powerwalls

Since we have a warranty that specifically covers degradation over 10 years, is there any way to actually monitor it? When I log in to the gateway, i don't see it. I know there are a few undocumented json paths for more info, but I don't know them all. Just wondering if there is a way other...

teslamotorsclub.com

 

[TyneTes]

Thanks for pointing me in the right direction. Will read with interest.

 

[sashton]

@TyneTes I would be very interested in your experience of a split phase supply and a gateway2. We currently have a number of PW2 and a GW2 on a single phase supply but have reached the DNO prescribed generation limit.
The solution being offered by WPD is a second supply 180 degrees out of phase fed off the same PMT.
It is reassuring to hear the GW2 supports two phases. Our plan is to leave the PWs and house loads on PH1 and just install an additional ~25kWp on the second phase.
A polyphase vector sum SMETS2 meter should provide a reading being the sum of the two phases.
...that's the plan anyway.

 

[ACarneiro]

@TyneTes I would be very interested in your experience of a split phase supply and a gateway2. We currently have a number of PW2 and a GW2 on a single phase supply but have reached the DNO prescribed generation limit.
The solution being offered by WPD is a second supply 180 degrees out of phase fed off the same PMT.
It is reassuring to hear the GW2 supports two phases. Our plan is to leave the PWs and house loads on PH1 and just install an additional ~25kWp on the second phase.
A polyphase vector sum SMETS2 meter should provide a reading being the sum of the two phases.
...that's the plan anyway.

Our of interest, why not go with a three phase supply and a three phase inverter for your total solar array?

 

[sashton]

Our of interest, why not go with a three phase supply and a three phase inverter for your total solar array?

Simply down to cost. We are currently limited to 23kW gen / 17kW export on a single phase supply. To expand our capacity WPD offered split phase for £6k (48 kW gen / 40kW export) but three phase would require a replacement PMT and a second 11kV conductor back to the substation and would cost £32k. This would raise the limits to 72 kW gen / 55kW export.
We only need the lower ost option.

 

[TyneTes]

Our of interest, why not go with a three phase supply and a three phase inverter for your total solar array?

Our of interest, why not go with a three phase supply and a three phase inverter for your total solar array?

Our problem here (deeply rural Northumberland) is that we only have 2 phase in the village. 3 phase not technically possible for the DNO.

So, 2 phase supply installed. Roughly 5kWp solar and 1x Powerwall on each phase. Plus 1x ASHP per phase, each drawing max 4 kW. Plus zappi on L1 for the MX. Plus a pottery kiln and a 3-pin granny lead charger for a local-use-only Nissan Leaf on L2.

Two issues so far.

1) MAJOR issue: Octopus have not yet found a solution for SMETS2 smart meter on a TWO phase supply (as opposed to 3 phase). Looks like we’re heading for TWO single phase meters, with two MPANs. Only this way will Octopus be able to continue supplying the Go Faster tariff which is essential for the economics of the heavily electrified setup we’re pursuing here (cars, heating, hot water, micro-generation, storage to time-shift off peak grid imports and solar generation).

Over the seven year term of the Renewable Heat Initiative subsidy, being forced to use a ‘dumb’ tariff as opposed to our Octopus smart TOUT would cost us between £29,000 and £61,000 extrA, depending on the delta in kWh pricing (!!). And given that all the kit and two phase upgrade is costing us over £40,000 that would be a material hit to the payback calls.

2) Minor issue. The Tesla Gateway routes ONLY one of the two phases (L1 in our case) through the connectors which provide back up in the case of grid outage.

We elect to keep the HOUSE circuit on the backed-up supply (L1).

Using the app, we then set a given percentage as backup. Say, for the sake of argument, 50%.

Because the gateway is backing up ONLY one of the two phases, and because ONLY one of the two Powerwalls is on that phase, you’d expect. 50% backup setting to produce about 6.8 kWh of reserved energy.

However the 50% setting produces about 13.6 kWh reserve, i.e. 50% of BOTH batteries. Which is weird, because the L2 circuit on which the battery sots is routed through the second phase connectors on the Gateway. Which are explicitly identified in the installer’s instructions as NOT providing backup.

Note: we noticed this odd behaviour when Storm Watch recently switched on 100% backup reserve. The Gateway reserved 100% of the capacity of BOTH batteries, even though only ONE is wired to connectors in the gateway which is capable of providing backup.

 

[miimura]

If you're stuck with two separate meters, you should ask if you can wire each Powerwall with its own Gateway. The benefit is that both phases will be backed up, but they will be two separate systems that cannot share power with each other. However, today if one PW is on the phase that is not behind the Gateway, it's essentially useless during an outage anyway. I think this will work because there are no loads connected to multiple phases like they would in a true three phase system. The concern would be the phase relationship between the two separately backed up phases going through their own Gateway. I'm pretty sure a split phase 240V supply in the UK doesn't need rigorous phase consistency when the grid is down. In North America, our split phase is 120V/240V and 240V loads must have the 120V lines exactly 180 degrees apart. The Powerwalls natively supply power this way when configured for North American grid.

 

[sashton]

Our problem here (deeply rural Northumberland) is that we only have 2 phase in the village. 3 phase not technically possible for the DNO.

So, 2 phase supply installed. Roughly 5kWp solar and 1x Powerwall on each phase. Plus 1x ASHP per phase, each drawing max 4 kW. Plus zappi on L1 for the MX. Plus a pottery kiln and a 3-pin granny lead charger for a local-use-only Nissan Leaf on L2.

Two issues so far.

1) MAJOR issue: Octopus have not yet found a solution for SMETS2 smart meter on a TWO phase supply (as opposed to 3 phase). Looks like we’re heading for TWO single phase meters, with two MPANs. Only this way will Octopus be able to continue supplying the Go Faster tariff which is essential for the economics of the heavily electrified setup we’re pursuing here (cars, heating, hot water, micro-generation, storage to time-shift off peak grid imports and solar generation).

Over the seven year term of the Renewable Heat Initiative subsidy, being forced to use a ‘dumb’ tariff as opposed to our Octopus smart TOUT would cost us between £29,000 and £61,000 extrA, depending on the delta in kWh pricing (!!). And given that all the kit and two phase upgrade is costing us over £40,000 that would be a material hit to the payback calls.

2) Minor issue. The Tesla Gateway routes ONLY one of the two phases (L1 in our case) through the connectors which provide back up in the case of grid outage.

We elect to keep the HOUSE circuit on the backed-up supply (L1).

Using the app, we then set a given percentage as backup. Say, for the sake of argument, 50%.

Because the gateway is backing up ONLY one of the two phases, and because ONLY one of the two Powerwalls is on that phase, you’d expect. 50% backup setting to produce about 6.8 kWh of reserved energy.

However the 50% setting produces about 13.6 kWh reserve, i.e. 50% of BOTH batteries. Which is weird, because the L2 circuit on which the battery sots is routed through the second phase connectors on the Gateway. Which are explicitly identified in the installer’s instructions as NOT providing backup.

Note: we noticed this odd behaviour when Storm Watch recently switched on 100% backup reserve. The Gateway reserved 100% of the capacity of BOTH batteries, even though only ONE is wired to connectors in the gateway which is capable of providing backup.

I have a written reply from Agile Phil (Octopus) that they can supply a Polyphase SMETS2 meter which function as a two phase vector sum meter on a 180 degree split phase supply.

 

[sashton]

If you're stuck with two separate meters, you should ask if you can wire each Powerwall with its own Gateway. The benefit is that both phases will be backed up, but they will be two separate systems that cannot share power with each other. However, today if one PW is on the phase that is not behind the Gateway, it's essentially useless during an outage anyway. I think this will work because there are no loads connected to multiple phases like they would in a true three phase system. The concern would be the phase relationship between the two separately backed up phases going through their own Gateway. I'm pretty sure a split phase 240V supply in the UK doesn't need rigorous phase consistency when the grid is down. In North America, our split phase is 120V/240V and 240V loads must have the 120V lines exactly 180 degrees apart. The Powerwalls natively supply power this way when configured for North American grid.

I have no issue with one phase on a 180 degree split phase (230v - 0 - 230v) supply not being backed up. It is whether the GW2 will function correctly with 2xPW/2 on the backed up (house load phase) and 1xPW/2 on the non-backed up phase (workshop so the PW/2 here is just to improve self-consumption)?

 

[gasbag]

I have a written reply from Agile Phil (Octopus) that they can supply a Polyphase SMETS2 meter which function as a two phase vector sum meter on a 180 degree split phase supply.

did you achieve a polyphase vector sum SMETS2 meter with octopus? I am twin phase in cornwall trying to expedite my migration from shell t octopus...
I have 1x powerwall and some solar on each phase. newly connected and minimal sunshine since plus shell have managed to fail to the the meter talking to them (the previous single phase meter managed fine. the new meter is
Meter:
EDMI ES-30B ESME Polyphase
Comms box is: Toshiba SKU2 cellular & mesh
Antenna is: WNC T1 aerial (or T2?)

I have asked for a T3 aerial?
Also, can the meter do vector sum and load balance?
Finally (assuming that you have succeeded... are you willing to share (sanitised if ned be) the letter from octopus?
BW RS