Somewhat interestingly/frustratingly, clear, concise (and even vaguely) technical information on either version of the PowerWall 2 (yes, there are two versions - the "DC" and "AC") has been a bit difficult to find, so in my research, what have I found?
This page or its contents are not intended to promote any of the products mentioned nor should it be considered to be an authoritative source.
It is simply a statement of opinion, conjecture and curiosity based on the information publicly available at the time of the original posting.
It is certain that as time goes on that information referenced on this page may be officially verified, become commonplace, or proven to be completely wrong.
Such is the nature of life!
The "DC" PowerWall 2:
- Data sheets (two whole pages, each - almost!) for both the DC and AC versions of the PowerWall may be found here at this link - link.
Unlike its "AC" counterpart (more on this later) this version of the PowerWall 2 does NOT appear to have an AC (mains) connection of any type - let alone an inverter (neither are mentioned in the brochure) - but rather it is an energy back-up for the solar panels on the DC input(s) of the hybrid inverter. "Excess" power from the panels may used to charge the battery and this stored energy could be used to feed the inverter when the load (e.g. house) exceeds that available from the panels - when it is cloudy, if there is a period in which the load exceeds the output of the PV array for a period of time or there is no sun at all (e.g. night).
Whether or not this version of the PowerWall can actually be (indirectly) charged via the AC mains (e.g. via a hybrid inverter capable of working "backwards" to produce AC from the mains) would appear to depend entirely on the capability and configuration of the hybrid inverter and the system overall.
But, you might ask,why would you ever want to charge the battery from the utility rather than from solar? You might want to do this if there were variable tariffs in your area - say, $0.30/kWh during the peak hours in the day, but only $0.15kWh at night - in which case it would make sense supplant the "expensive" power during the day with "cheap" power bought at night to charge it up.
Whether or not this system would be helpful in a power outage is also dependent on the nature of the inverter to which it is connected: Most grid-tie converters become useless when the mains power disappears (e.g. cannot produce any power for the consumer - more on this later) - and this applies to both "series string" (e.g. a large inverter fed by high-voltage DC from a series of panels) and the "microinverter" (small inverters at each of the panels) topologies. Inverters configured for "island" operation (e.g. "free running" in the absence of a live power grid) or ones that can safely switch between "grid tie" and "island" modes would seem to be appropriate if you use the DC PowerWall and you want to keep your house "powered up" when there is a grid failure.
The "AC" PowerWall 2:
- Data sheets (two whole pages, each - almost!) for both the DC and AC versions of the PowerWall may be found here - LINK.
(The basic diagram of Figure 1, above, shows how such a system might be connected. This diagram does not specifically represent a PowerWall, but rather how any battery-based inverter/charger system might be used to supply back-up power to a home in the past and future.)
Because its power would be connected "indirectly" via AC power connections to the PV system it should (in theory) work with either a series-string or microinverter-type system - or, maybe even if you have no solar at all if you simply want to charge it during times of lower tariffs and pull the charge back out again during high tariffs. (The Tesla brochure simply says "Support for wide range of usage scenarios" under the heading "Operating Modes" - which could be interpreted many ways, but I have not actually seen an "official" suggestion of a use without any sort of solar.)
What might such a system look like - schematically, at least?
How might this version of the PowerWall operate? First, let's take a look at a diagram of how any sort of battery/inverter/charger like this might be configured for a house.
Referring to Figure 1, above:
Shown to the right of center is a switch that opens when the utility's power grid goes offline, isolating the house and the inverter/charger from the power grid and included in that is a voltage monitor (consisting of potential transducers, or "PTs") that can detect when the mains voltage has returned and stabilized and it is "safe" to reconnect to the grid. The battery-based inverter/charger is connected across the house's mains so that it can both pull current from it to charge its battery as well as push power into the house in a back-up situation.
The "Net current monitoring" current transducers ("CTs") might be used to allow the inverter/charger to "zero out" the total current (and, thus power) coming in from and going out to the power grid (under normal situations) such as when its battery is being charged and extra power is being produced by the PV system, but to control the charge rate just so that only that "extra" power from the PV system is being used to assure, as much as possible, a net-zero flow to/from the utility. The "House Current monitoring" is used to determine how much current is being used by the entire house while the "PV current monitoring" is used to determine the contribution of the PV system.
By knowing these things it is possible to determine how much excess/deficit their may be in terms of the production of the PV system with respect to actual usage by the household. Not shown is the current monitoring that would, no doubt, be included in the inverter/charger itself. Some of the shown current monitoring points may be redundant as this information could be determined in other ways, but are included for clarity.
Finally, a local network connection is shown for both the inverter/charger so that they may communicate with each other, perhaps for control purposes, as well as communicate via the Internet so that statistics may be monitored and recorded and to allow firmware updates to be issued.
How it might operate in practice:
As can be seen in Figure 1 and determined from the explanation, we can see that the PV is connected to the input/output of the inverter/charger (which could be a PowerWall - or any other similar system) via the house wiring which means that there is a path to the PowerWall to charge its battery, and the same path out of it when it needs to supply power along with means of monitoring power flow.
With a system akin to that depicted in Figure 1, consider these possible scenarios:
- Excess power is being produced by the PV system and put back into the grid and the PowerWall's battery is fully-charged. Because the battery is fully-charged there is nowhere to put this extra power so it goes back into the grid, tracked by the "Net Meter" in the same way that it would be without a PowerWall.
- Excess power is being produced by the PV system and put back into the grid and the PowerWall's battery is not fully charged. It will pull the amount of "excess" power that the PV system would normally be putting into the grid and charge its own battery at that same rate resulting in a net-zero amount of power being put into the grid.
- More power is being consumed by the user's household than is being produced by the solar array. Depending on the state-of-charge and configuration of the PowerWall, the power wall may produce enough power to make up for the difference between what the PV system is producing and the user needs. At night this could (in theory) be 100% of the usage if the system were so-configured.
- It would be theoretically possible to configure it so that even if there was no solar, but a higher daytime than nighttime power rate, to charge overnight from the mains and put out power during the day to reduce the power costs overall.
All of the above scenarios are to be expected - and they are more-or-less standard offerings for many of the battery-based products of this type - but what if the AC mains go down? For the rest of this discussion we will ignore the "DC" version of the PowerWall as its capability would rely on the configuration of the user's hybrid inverter and its capabilities/configuration when it comes to supplying backup, "islanded" AC power.
As mentioned before, with a typical PV system - either "series string" (one large inverter) or distributed (e.g. "microinverter") - if the power grid goes offline the PV system becomes useless: It requires the power grid to be present to both synchronize itself and present an infinite "sink" into which it can always push all of the "extra" power power that it is producing. Were such units to not dangerous voltages could be "back-fed" into the power grid and be a hazard to anyone who might be trying to repair it. It is for this reason that all grid-tie inverters are, by law, required to go offline - or, at least, disconnect themselves completely from the power grid during a mains power outage.
The "AC" version of the Tesla PowerWall's system includes a switch that automatically isolates the house from the power grid when there is a power failure. Once this switch has isolated the house from the grid the inverter built into the PowerWall can supply power to the house - at least as long as its battery lasts.
What about charging the battery during a power outage?
Here is where it seems to get a bit tricky.
If all grid-tie inverter systems go offline when the power grid fails, is it possible to use it to assist, or even charge the PowerWall during a grid failure? In other words, can you use power from your PV system to recharge the PowerWall's battery or, at the very least, supply at least some of the power to extend its battery run-time?
In corresponding with a company representative - and corroborated by data openly published by Telsa (see the FAQ linked near the bottom of this posting) - the answer would appear to be "yes" - but exactly how this works is not very clear.
Based on rather vague information and knowing the behavior of the components involved it would seem to need to work this way:
- The power (utility) grid goes down.
- The user's PV system goes offline with the failure of the grid.
- The PowerWall's switch opens, isolating the house completely from the grid - aside from the ability to monitor when the power grid comes back up.
- The inverter in the PowerWall now takes the load of the house.
- When the back-up power system's inverter goes online, the PV system again sees what looks like the power grid and comes back online.
- As it does, the back-up power system monitors the total power consumption and usage and any excess power being produced by the PV system is used to charge its battery.
- If the PV system is producing less power than is being used, the back-up power system will supply the difference: Its battery will still be discharged, but at a lower rate.
What is there is extra power being produced by the PV system?
Grid tie PV systems expect the power grid to be an infinite sink of power - but what if, during a power failure, when your backup up system is standing in as the power grid, you are producing 5kW of solar energy and your house is using only 2kW: Where does the extra 3kW of production go if it cannot be infinitely sinked into the utility grid, and how does one keep the PV system from "tripping out" and going off line?
To illustrate the problem, let us bring up a related scenario where we have a generator instead of some sort of battery-based back-up power system system.
There is a very good reason why owners of grid-tie systems are warned against using it to "assist" a backup generator and using that generator as a substitute for the power grid. What can happen is this:
- The AC power goes out and the transfer switch connects the house to the generator.
- The generator comes online and produces AC power.
- If the AC power from the generator is stable enough (not all generators produce adequately stable power) the PV system will come back online thinking that the power grid has come back.
- When the PV system comes back online and produces power, the generator's load decreases: Most generator's motors will slightly speed up as the load is decreased.
- When the generator's motor speeds up, the frequency goes high. When this happens, the PV system will see that as unstable power and will go offline.
- When the PV system goes off, the power is suddenly dumped on the generator and it is hit with the full load and slows back down.
- The cycle repeats, with the PV system and generator "fighting" each other as the PV system continually goes on and offline.
- The AC power goes out, the transfer switch connects the house to the generator.
- The generator comes online and produces power.
- The PV system comes up because it "sees" the generator as the power grid, but its producing, say, 5kW but the house is, at the moment, using 2kW.
- The PV system, because it think that it is connected to the power grid, will try to shove that extra 3kW somewhere, causing one or more of the following to happen:
- The generator to speed up as power is being "pushed" into it, its frequency wil go high and trip the PV system offline, and/or:
- If the PV system tries to push more power into the system than there is a place for it to go (e.g. the case, above, where the solar is producing 3kW more than is being used) the voltage will necessarily go up. Assuming that the generator doesn't "overspeed" and trip-out and the frequency doesn't go up and trip the PV system offline, the PV system will increase the voltage, trying to "push" the extra power into a load where there is nowhere for it to go:
- As the PV system tries to "push" its excess power into the generator, it will increase the output voltage. At some point the PV system will trip out on overvoltage, and the same "on-off" cycle mentioned above will occur.
- It is possible that the excess power from the PV will "motor" the generator (e.g. the input power tries to "spin" the generator/motor) - an extremely bad thing to do which will probably cause it to overheat and eventually be destroyed if this goes un-checked.
- If it is an "inverter" type generator, it can't be "motored", but the excess power will probably cause the generator's inverter to get stuck in the same "trip out/restart" cycle or simply fault out in an "overload condition - or it might even be damaged/destroyed.
What if we have excess power and nowhere to put it?
The question that comes to mind now is "What does the PV system do when the PowerWall's battery is fully-charged and there is no-where to put extra energy that might be being produced?" Where we again bring up the situation where our PV system is producing 5kW but we are using only 2kW leaving an extra 3kW to go... where?
The answer to that question is not at all clear, but four possibilities come to mind:
- Divert the power elsewhere. Some people with "island" systems utilize a feature of some solar power systems that indicate when excess power is available and use it to operate a diversion switch to shunt the excess power in an attempt to do something useful like run an electric water heater, pump water or simply produce waste heat with a large resistor bank. Such features are usually available only on "island" systems (e.g. those that are entirely self-contained and not tied to the power grid) and with large battery banks.
- If it is possible, simply disable the PV system for a while and drain, say, 5-10% of the power out of the back-up power system battery before turning it back on and recharging it. This will cause the PV system to cycle on and offline, but it will do so relatively slowly and it should cause no harm.
- Somehow communicate with the PV system and "tell" it to produce only the needed amount of energy. This is a bit of a fine line to walk, but it is theoretically possible provided such a feature is available on the PV system.
- Alter the conditions of the power being produced by the back-up power system inverter such that it causes the PV system to go offline and stay that way until it needs to come back online.
Let's eliminate #1 as that will not apply to a typical grid-tie system, so that leaves us with:
#2: Disable the PV system:
Of these three possibilities #2 would seem to be the most obvious and it could be done simply by having another switch on the output of the PV system that disconnects it from the rest of the house, forcing it to go offline - but this has its complications.
For example, in my system the PV is connected into a separate sub-panel located in the garage: If one were to disconnect this branch circuit entirely, the power in the garage would go on and off, depending on the state-of-charge of the PowerWall or other battery-based back-up power system. Connecting a PV system to a sub-panel is not an unusual configuration as it is not uncommon to find them connected to sub-panels that feed other systems, say, the air conditioner, kitchen, etc. (e.g. wherever a suitable circuit is available) so I'm guessing that they do not do it this way - unless they do it at the point before the PV system connects to the panel.
As noted above, one would disable the PV system once the battery had fully charged but enable it again once the battery had run down a bit - say, to 90-95%. This way, one would not be rapid-cycling the PV system and the vast majority of the back-up power system's battery storage capacity would be available.
While this, too, should work, I suspect that it is not the one that is used as the drawings in the brochures don't show any such connection - but then again, they don't show the main house disconnect that would have to be present - but it would probably work just fine if the PV system were to gracefully come back online when it was time to do so (e.g. no user intervention to "reset" anything.)
#4: Alter the operating conditions to cause the PV system to go offline:
Then there is #4, and one interesting possibility comes to mind - and it is a kludge, but it should work.
One of the parameters that could be altered would be the frequency at which the back-up power system's inverter operates (say, 2-3 Hz or so above and/or below the proper line frequency) and force the PV system offline with that variance. Even though this minor frequency change is not likely to hurt anything (many generators' frequencies drift around much more than this with varying loads!) things that use the power line frequency as a reference - such as clocks - would drift rather badly unless the frequency were "dithered" above and below the proper frequency so that its long term average was properly maintained.
Edit - 20170719:
In digging around, I have determined that "dithering" the frequency is, in fact, one of several ways that is used by an battery-backed inverter to disable a PV inverter when the PV is producing more power than can be accommodated by the load and/or battery charger. This system, called "Frequency Shift Power Control" (FSPC) by at least one manufacturer (e.g. SunnyBoy) is designed to do this very thing.#3: "Talk" to the PV system and control the amount of power that it is producing:
A description of this technique may be found in section 6 of the document at this link.
Whether or not this is one of the control methods used by the Power Wall is not known at this time.
That leaves us with #3: Communicate with the PV system and "tell" it (perhaps using the "ModBus" interface) to produce only enough power to "zero" out the net usage.
The problem with this method is that it would depend on the capabilities of the PV inverter system and require that they support such specific remote control functions. While it is very possible that some do, this method would be limited to those so-equipped.
#3 and #2: "Talk" to the PV system to turn it on and off as needed:
Included in #3 could be a variant of method #2 and that would be to send a command to the inverter via its network connection to simply shut down and come back online as needed to keep the battery between, say, 90% and 100% charge as mentioned above.
This second variant of #3 seems most likely as there it is likely that there is some sort of set of commands capable of this that would be widely implemented across vendors and models.
* * *
What do I think the likelihood to be?
I'm betting on the second variant of #3 where a command is sent to the PV system to tell it to turn off - at least until there is, again, somewhere to "send" excess power.
* * *
Having said all of this, there is a product FAQ that was put out by Tesla that seems to confirm the basic analysis - that is, its ability to run "stand alone" in the event of a power failure and the charge be maintained if there is sufficient excess PV capacity - read that FAQ here - LINK.
Additional information may be found on the GreenTech Media web site: "The New Tesla Powerwall Is Actually Two Different Products" - LINK. This article and follow-up comments seem to indicate that there were, at that time, only a few manufacturers of inverters, namely SolarEdge and SMA (a.k.a. SunnyBoy) with which Tesla was installing/interfacing their systems, perhaps indicating some version of #2 or #3, above. Clearly, the comments, mostly from several months ago, are also offering various conjectures on how the system actually works.
I'm investigating getting a PowerWall 2 system to augment my PV generation and provide "whole house" backup and have been researching how it works.
While I have occasionally asked questions of representatives of Tesla, nothing that they have said is anything that could not be easily found in publicly-released information on the internet and as of the original date of this posting I haven't signed anything that could possibly keep me from talking about it.
However it is done, it should be interesting!
* * *
Finally, if you can find more specific information - say from a public document or from others' experience and analysis that can add more to this, please pass it along!
This page stolen from "ka7oei.blogspot.com"