The key features for us involve knowing time of day when rates are peak, monitoring grid current with a current transformer, and storing power in battery during off-peak to minimize import/maximize export at peak times.
So if you have SMA's Energy Meter (EM) and Data Manager (DM), you have a way to tell the SI about grid power. I am not aware of any way to tell the DM about grid tariffs, although this capability is available on the ennexOS portal, which may not be obviously useful for this type of feature requirement.
The most I have seen on the SI is the ability to define the most profitable month (in terms of solar yield + grid feed-in) and definition of seasons. But I have not dug into this as neither apply to me. Moreover, I don't think these will solve your problem anyway.
Have SMA told you whether they can support this feature in their "H" SI inverters at some point in the future?
For tax reasons, may want to ensure battery only charges from PV, never from grid (makes eligible for credits; Tesla is careful to do this.)
Right, Tesla do this because they do not market the PowerWall as an off-grid system. SMA's focus is off-grid (even though they do support grid-tied deployments), so their philosophy is quite different from Tesla's.
If we have different rate for import-export (e.g. pay $0.25, get credited $0.03, we want it to adjust its inverter/charger power seeking to maintain zero amps through grid connection.
I think this is something SMA can implement as part of their DM, but you would also need to install their EM and have a DM too. From my system, I don't see this capability today, but I do not see why it cannot be supported as part of a software update.
But again, given SMA really sell themselves as an off-grid vendor, I'm not sure how much appetite they would have for this, especially since not all markets have this from their grid providers (although, to be fair, the U.S. and Australia, which are big SMA markets, do).
These features are in other models like Sunny Boy Storage, which can also be set up for grid backup. But those still aren't stackable and don't have same surge capability. Also use high voltage battery. Sunny Island is still the best for larger systems & bigger motors, also DIY or otherwise affordable batteries.
Indeed.
The SB Storage also supports Li-Ion batteries only, so this is a limitation for LA installations. It is what comes closest to Tesla's PowerWall, which is why it can support similar grid-tied features.
The off-grid limit of GT PV wattage 2x SI wattage I would guess is a fuzzy limit.
The SI 8.0-H is 6kW rated, 11.5kW grid pass-through. The 6kW rating from the inverter is governed by other factors such as temperature and conversion losses. However, the 11.5kW pass-through performance is guaranteed under all conditions, as it's just a relay across the SI between loads and the grid.
Load addition would be limited to surge capability of SI (and whether it was already sourcing or sinking power), and load-dump
The limit I have hit with grid pass-through on the SI is the connection point from the grid provider. In South Africa, single-phase houses typically get 20A, 40A or 60A services. I get a 60A service, which is 13.8kW (230VAC service). If you parallel SI's and have a larger grid source (like 100A or 3-phase), you should be able to have good surge capability from just the grid, i.e., 11.5kW per SI in your system.
I would think battery charge limit (do battery charge limit settings prevent momentary surge higher? Lead-acid would tolerate, but lithium BMS would disconnect unless they have surge programmed in.)
The SI "H" inverters allow you to set charge and discharge current differently and separately. I have 7x 48V 100Ah Li-Ion batteries, so I set total charge current to 140A (0.2C per battery). However, I set discharge current to 665A (95A per battery). I never hit that discharge current in any situation - I just set it that way to allow for sudden surges. Each of my batteries can discharge 6kW peak for 60 seconds, so that is 42kW for 1 minute in total vs. 33.6kW rated.
But you are right, it all comes down to how relaxed the Li-Ion battery OEM is, and if they are restrictive, you may trip the battery sooner than you may trip the inverter during a surge. This makes is helpful to either choose the right battery OEM, or have access to the BMS where you can change thresholds based on your requirements.
There are also minimum battery size recommendations - 100 Ah per SI, 100 Ah per kW of GT PV. My AGM is 1/3 the recommended size based on PV, seems to be OK with my loads.
SMA permit a minimum of 50Ah for Li-Ion, but seem to recommend 100Ah for LA. But as you point out, they recommend 100Ah minimum for C10 capacity per battery per 1kWp of PV.
Technically, if your load is stable and smooth, I do not see an issue with you running a battery smaller than 100Ah. I think the issue will arise if you get a sudden in-rush of PV power, as the SI is slower to react via FSPC (Frequency Shift Power Control) vs. if charge current was coming from the grid. So a bigger battery ensures safety, because that sudden surge of PV power would have somewhere to go as the SI and SB ramp down PV production to match what the BMS is asking for.
The SI can react quicker if the charge current source is the grid, which is also why I've seen my SI automatically set a higher charge voltage when the grid is the charge source vs. when PV is the charge source (by up to 1.0V difference).
AC sources (grid or generator) are normally connected to AC2 input. SI synchronizes to them, then lets them be voltage/frequency source. When disconnecting, there is a drop-out.
So this has been markedly improved in the "H" series compared to the U.S. SI's.
The SI 4548-US-10 and 6048-US-10 have a switchover time of up to 35ms for low-resistance outages from grid or generator.
However, this has been improved for the "H" series inverters from 1.03.00.R, specifically the 6.0H-12/13 and 8.0H-12/13, to 20ms.
In practice, I have come across both scenarios, where high-resistance outages from the grid enjoy a 0ms - 10ms switchover (so UPS's notice, but nothing else in the house), as well as where low-resistance outages lead to a drop-out... but in that case, the inverter automatically restarts 60 seconds later.
I installed a voltage sensor to help the SI disconnect from the grid before they see a brownout, but that doesn't help all the time. I keep it anyway, just as an extra protection point, especially against unintended high voltage from the grid.
There have been numerous reports of SI customers across Africa of this issue during low-resistance grid outages, and a voltage sensor has been the recommendation. But as I mentioned, this is not a guaranteed solution in all cases. When it works, and when it doesn't, is unpredictable. I have to admit that Victron handle this better than SMA, probably because they have an in-built voltage sensor.
I think the issue lies somewhere between the SI and battery agreeing to quickly form a grid once the main grid fails, and one or both of them are unhappy by the low-resistance outage.
I'm thinking about whether monitoring grid frequency sag might be better than monitoring voltage sag. Any thoughts?
My idea is to feed AC source, not into AC2 input, but into a Sunny Boy just as you would from a hydro turbine. Rectify the AC and connect to "PV" input of SB. Needs to be electrically isolated, so use a transformer. Capable of delivering large fault current, so have soft-start mechanism (e.g thermistor, precharge resistor, or transistorized current-limit circuit. Possibly use arc welding transformer, which limits current by inductance.) The nicer way would be a power-factor correction circuit, but rectifier-capacitor is the crude way.
If you have an inverter generator, probably just tap into its DC rail.
What this does is it lets SI be the voltage and frequency source continuously, no glitch on transfer.
Hmmh, that is quite the solution, I have to admit
![Smile :) :)](data:image/gif;base64,R0lGODlhAQABAIAAAAAAAP///yH5BAEAAAAALAAAAAABAAEAAAIBRAA7)
. Not sure SMA would be pleased to hear you've turned their Sunny Boy into a rectified DC inverter, but well...
My workaround, for now, is to disconnect the main grid from being the main voltage source for loads by running the SI's as off-grid, even if they are grid-tied. I then set thresholds where the grid is called in if the battery is discharging at 4kW or higher for more than 45 seconds. This is ideal in periods of low sunlight where calling in the grid will help keep the battery charged so we have enough backup energy for emergency conditions.
This way, if the grid goes away, there is no drop-out because, well, it was not the voltage source at the time.
Have you tried monitoring the external grid via other mechanisms so that you give the SI some buffer where it does not have to "surprise" you during a grid/generator outage?