It’s been 9 months since the first installs of Tesla’s Powerwall2 in Australia, and now you can enable programmatic Time-of-Use power-cost tariff awareness for your PW2. This means that you can have your PW2 “save” it’s capacity during cheap-power off-peak periods when it’s likely that the PW2 will discharge to empty before the peak period starts, and the PW2 can even charge from the grid during cheap-power off-peak periods, when the predicted solar generation for the following day is going to be too low to meet the peak power demands.
There’s been two new developments that allow this, and which one you choose depends on your level of technical ability, as they both appear to reach a similar result – the Tesla App Time-Of-Use tariff-awareness feature can be enabled in the App on request, or you can run a user-developed service on a Windows PC to control it yourself.
The Tesla App way
First up, the Tesla App has been upgraded for a few people at random here in Australia, according to posts at Whirlpool and TMC. I came across a Tesla engineer who commented on my PVOutput setup and asked them to enable this for me, and it was done a few days later. The detail of how the App TOU feature works is here.
The App changed in a number of areas when TOU is enabled:
There is an additional “tick” in “Customize” to enable TOU awareness (they call it “Time Based Control”), and it opens up a choice of two options:
Balanced makes the PW2 save capacity during off-peak time, for later use, but it will only charge itself when solar generation is greater than home consumption.
Cost Saving is like “Balanced”, but it also allows charging the PW2 from the mains during the off-peak periods. Use this mode if you have a cheap Off-peak power tariff.
There is an associated Edit Price Schedule screen that lets you set the TOU periods so the PW2 logic knows when it can charge or go to standby, and when it must power the home. It can deal with weekends having all off-peak. No shoulder option though, so I am treating shoulder rates as off-peak.
In practice, I have observed these modes both stop the PW2 from delivering power during off-peak periods, such that the PW2 still retains 10% capacity above the blackout “reserve” come the following morning peak period.
In the day graphed below, you can see the purple grid-supplied periods are present even when the PW2 still has capacity to deliver – the green PW2 capacity line flattens out instead of continuing the downward discharge trend.
From the left, the PW2 is put to standby around midnight with a big load still required to charge my car, and then a few hours later, it begins to deliver power as the morning approaches and presumably it “knows” that a reasonably sunny day will be coming up and the capacity no longer needs to be “saved”.
Prior to enabling this feature, the PW2 would always discharge overnight until it hit the reserve backup percentage I specified for blackouts, as seen on the day graphed below.
As winter approached, there were days where the home was powered from the grid in the early morning at shoulder rates before the solar kicked in. I have not had any rainy days since this functionality was enabled, so I have yet to see the PW2 charging in off-peak, but someone else has mentioned it will do so, and at up to 5KW.
The other change to the App is that the Energy Usage graph now incorporates the off-peak/peak times as part of the display.
And, the display now splits your overall usage ratios down to the peak/offpeak as well. For this period, the entire peak-tariff usage was self-powered, and the offpeak was drawn from the PW2 until it was depleted, and the car charged to completion from the grid.
The Windows PowerwallService way
Not everyone is getting this App update though, and perhaps Tesla is rolling it out slowly to make sure that any early show-stopping bugs only affect a few users. Maybe enabling the feature in the App requires them to do some personalisation for your setup, similar to what is done below, that is labour-intensive, and there’s only so many programmer-hours available (how many engineer-hours are required to configure the thousands of PW2’s being installed in Puerto Rico this month?). Anyway, you can DIY if you can edit a config file in Windows, following the instructions below.
How does this work?
A user at Whirlpool has written some code that you can install on an always-on Windows PC to command your PW2 to achieve a similar result to the above App-based control.
Broadly, the service sends your solar array details to an external solar-output estimation provider, who combines them with local weather forecasts, to return a tailored estimate of your future energy production to the service. You define your estimated usage and the amount of power your PW2 needs to retain, the TOU periods and tariffs you have and so on in a .config text file. Then, the service runs to calculate the amount it needs to save for the following day, taking in to account the estimated future solar generation, and it will use your Tesla cloud login to switch the PW2 modes from “self-powered” to “standby” or “backup” (charging).
Your PW2 is connected to Wifi.
An always-on Windows PC on the same network as the PW2.
How to get this working.
Grab the latest .ZIP of the files from the Github page here.
Follow the install instructions from here. The service has an .EXE that gets called to do the work, and a .config file that you can edit with Notepad++. The parts you need to edit in the file are clearly marked, and the meaning of the user-defined variables are well documented in each case.
You will need to register an account at solcast.com.au to get an external prediction for the following days of solar output from your system. You edit the .config file to give it your Solcast credentials and system stats, and it returns a value for the following 3 days-worth of solar predictions. You need to look up your location in Google Maps to get the latitude/longitude of your solar array, and estimate it’s tilt and orientation towards North (azimuth).
The PW2 minimum backup amount and your guesstimates of the “save” you want to achieve get defined in the .config here:
You get to define the TOU rates and times – the rates can be useful if you want to use the data at PVOutput later, provided you want to install a database server on the PC as well.
Start the service as an Administrator, and once the service is running, open the Event Viewer (Click Start, type “Event”) to see it’s actions under Windows Logs->Applications.
The pic above is the Solcast result for the anticipated solar generation over the next three days.
This is the Powerwall service reporting that the capacity is above the estimated “save” target, so no action is required.
Now the service has determined that the capacity is below the estimated “save” and it is commanding the PW2 to charge.
You will need to tweak the “Save” and “Consumption” values until it’s a good fit for your usage – they do not get updated any other way, except by the .config file. Tesla’s App option, however, claims that it learns from your usage and solar over time, so perhaps they can take into account the daily consumption and adjust the “save” actions to match the consumption patterns more closely.
Is all this effort worthwhile?
For people on same-price-all-day (Anytime) rate tariffs, this feature has no benefit. However, it’s unlikely that many Australian PW2 owners would be on Anytime tariffs, as they would be far better-off on TOU, since they will never pay for power at Peak rates if their PW2 is charged adequately enough every day, and the Off-Peak and Shoulder rates *should be* less than Anytime rates.
If you happen to be lucky enough to get off-peak power for a cost close to, or below, the Feed-In tariff, and you use more power than the PW2 + Solar can provide on cloudy days, you will benefit from this, for sure.
For users with higher Off-Peak tariffs, the benefit reduces as the cost goes up, unless there is a decent spread between the Off-Peak and the Peak tariff rates. You will end up discharging your PW2 completely less often, but if your PW2 discharged to empty during Peak time in the past, this feature should prevent that from occurring.
I have only been billed for Off-Peak to charge my car, and a tiny amount of Shoulder, since installing the PW2 last year – no Peak usage to speak of, and the Off-peak rate is under the Anytime rate in my area – $0.31 Anytime vs $0.26 Off-Peak, before discount, so remaining on the Anytime tariff would have been silly.
This is my billed usage for one month – Jan 2018 – where I used 688KWHr total, or around 23KWHr a day. Of that 688KWHr, only 133KWHr was drawn from the grid (85% was overnight car charging), with the rest coming from the PW2 and the solar array.
I paid for 114KWHr @ $0.20, 2.5KWHr @ $0.39, and 17KWHr @ $0.25, totaling less than $30 for a month’s power usage – less than $80 / quarter in total. Yes, there’s a connection charge on top of that, but the solar system export more than covers that. If I didn’t charge a car every night, I would likely be in credit with my retailer.
Now, if only I could find a power retailer who has the lowest Off-Peak tariff, and the highest Solar FiT rate – their Peak and Shoulder tariffs don’t matter, since I won’t ever use enough of them to notice.
Is a PW2 better for the environment, than just exporting your excess power?
Some people argue that the most environmentally-friendly use for excess daytime solar generation is to send it to the grid, powering your neighbour’s homes in exchange for a Feed-in Tariff, instead of storing it in a PW2 for later use. Treating the grid as a battery, as it were.
This view assumes that the solar power exported is always “greener” than the (mostly-coal-fired) grid power that would otherwise be used by those neighbour’s homes, and that the grid will have a reduced demand for power in my area as a result of my export, which might translate to less coal being burnt, and less CO2, ash and so on. They also often mention that the losses of charging a PW2 (around 12% of a KWHr is lost per round-trip) are wasting power.
However, the issue I have with this argument is believing that the coal-fired steam-generators, hundreds of kilometers away from Sydney, are going to have their rate of coal consumption adjusted just because I sent a few KW of power back out to the street.
There’s literally no way to reverse the coal-burning-to-heat-water process, there’s no programmatic way to communicate the export of my paltry few KW to the generator infrastructure, my state’s grid does not have a gigantic Tesla battery that can reverse it’s power input to output, and if demand on the grid changes minutely, the grid operators use fine-grained “surplus spinning reserve” and “frequency control” hardware that will move the grid power levels up or down to compensate, instead of forcing a change to a coarse control, such as the coal consumption rate. And, excess solar export in the street often causes voltages to rise instead, which does not help the neighbours, and effectively amounts to “wastage”.
Thus, the excess solar presents a problem for the legacy grid we have, and as such, if I store it in my PW2, I will be drawing no power from the grid later-on in the day at Peak time, and given the one-way flow of coal-fired power generation that the grid was designed for, this will reduce the amount of coal needing to be fed to the boilers.
As to the efficiency argument: without the PW2, the coal-fired power I would have to draw after sundown has to come from hundreds of kilometers away, meaning there’s plenty of losses getting it to me, at least 10%.
Anyway, if you are going to point out the PW2’s 88% round-trip efficiency figure, the no-PW2-alternative “grid as virtual battery” option is way worse. The FiT payment I get to “charge” the grid is $0.12 / KWHr, but the cost to “discharge” it from the grid later starts at $0.20 and goes up to $0.39 / KWHr, for a round-trip efficiency of 60% at best, and 30% at worst.