Testing PowerWall off grid

The test is now complete.  I ran from about 5pm Tuesday 2/19/2021 continuously through about 10:30am Monday 2/25/2021 with the grid disconnected.  It was a highly successful test.  I'll put a synopsis here and leave the "stream of consciousness" stuff below for you to decipher if you wish.

The questions that needed answers are:

1) What happens if PV power can not be accommodated by local demand including PW charging? 

My PV production can reach about 25kw while the PW charging can accept only 5kw.  The short answer is that the PW shifts the frequency off of the standard 60hz and all inverters are expected to respond by suspending production.  That happens and the PV quickly shuts down and the PW can continue to supply less than 5kw of local demand from the PW battery.   The PW and PV can get into a cycle of PV being shut down and restarting 5 minutes later only to be immediately shut again.  Things go more smoothly if PV capacity is reduced to approximate the PW capacity to charge.

2) What happens if local demand exceeds the 5kw power capacity of the PW?

The PW shuts down and stops supplying local demand.  This should be avoided for several reasons.  One is that, due to inverters shutting down, the power will be down for about five minutes.  Another is that many devices such as clocks will have to be reset.  Another is that monitoring stops working and it is difficult to figure out what is going wrong or has gone wrong.  One fix for this is to adjust demands.  For instance, I have to watch the operation of my 5kw water heater closely to be sure that there are no other significant loads a the same time.

3) What happens when you try to charge a car?

Obviously car charging should be avoided unless there is good PV production.  It is risky to attempt to charge above the 5kw limit of a single battery PW.  However, up to 5kw charging is entirely feasible if one closely monitors other significant local loads.  I have a nattering nabob friend who has declared that car charging could not be done from a PW while off grid.  He is incorrect.  But, loads  such as a water heater must be managed to avoid conflicting with car charging.  PV production may be highly variable so it is best to insure that the PW battery can supply the loads if PV power disappears unexpectedly.  While experimenting charging my Tesla, I was surprised to find that Tesla has fairly recently integrated PW and Tesla car software to force the car to slow charging when it appears the PW may fail to supply demand.  I have more testing and study to do on that feature. 

The problems cited all become much easier if a PW has a generous number of battery units.  Each battery unit adds 5kw of supply capacity and 5kw of variable PW charging demand.

I did not address the possible problem of the PW battery becoming depleted.  If it does become depleted, one may well be left with no electric power whatsoever. for an indefinite time.  Once down and without grid power the PW will take 5 minutes to get the PV started again.  During that time, the PW will have to be supplying local demands even with good sun.  I imagine the PW has some reserve capacity which might be used for restarting especially if grid power can handle local loads immediately.

 Older parts of posting:

 I've had the PowerWall for about two years now and not done much testing in preparation of a long term grid down situation.  The reason is that I've been reluctant to lose the revenue from PV energy sales to my utility.  In fairly good production weather, that is around $10/day.  Sometimes near $20.

Due to a friend's questioning of off grid PW use during cloudy weather, I've decided to do a multi day test while we have a long period of overcast weather in the forecast.  The low production overcast weather will minimize the lost revenue as well as reduce the problem of the PW being unable to handle high PV production.  I switched off grid about dusk on Tuesday 2/19/2021.  The PW had a state of charge of about 95%.  This very cloudy morning, Wednesday 2/20/2021, the SOC bottomed out 25% about 9am.  At about 9:30, even though it is very cloudy, the PW is charging at about 2.5kw.  Since the charge power limit of my single battery PW is 5kw, something interesting will happen when my PV production goes above the ability of the PW ability to accept.  That is either 5kw or zero depending on whether the PW battery is charged or not.

In good weather while grid connected, I can sell large amounts of PV power to my utility.  Off grid, the PW is unable to deal with PV power above about 5kw; zero if the PW battery is fully charged.  Now, about 10am, the PW charge power is around 3.5kw and I can see trouble coming.   I'll first observe if the PW is able to shut down PV inverters when power is too high.  The PW attempts the shutdown by shifting the "islanding" frequency off of 60hz.

Yesterday, a very cloudy and grid connected day, I produced about 32kwh and sold the utility 26kwh.  There must be some measurement error there because the difference should be about 20kwh.  About  10-12kwh to charge the PW and about 10kwh for daytime house consumption.

 11am 2/20/2021.  Still very cloudy but production has been very near 5kw.  I first turned off the 5kw roof top which faces south and produces well for only about 4 mid day hours.  Then turned off another major production block  leaving my most recent projects which are mostly E+W facing and produces 8-9 hours.  PW is at 68% and charging at less than 3kw.  My hope is to get to about 4-5pm with a PW battery that is near 100%.  If that comes to pass, I'll likely leave only that production block on for tomorrow.  Then, tomorrow about noon, evaluate to see if I want to turn more PV back on to get charged by the end of the day tomorrow.

 4pm 2/20/2021.  About 22 hours into the test, things are going VERY well.  I babied it past the water heating but, since then, no worries whatsoever.  The PW became fully charged about 1pm so the PW has been turning PV off regularly.  The PV goes down quickly and gracefully.  After going down, it takes about 5 minutes for inverters to come back up.  The PW is maintaining about 98% charge.  The PV comes on briefly and goes up to above 2kw every 5-6 minutes.

7am 2/22/2021.  Since there is sun in today's forecast, this test is near it's end.  I've been off grid for two days and three nights.  The test is a near total success.  Here are some takeaways:

1) With a single battery PW, I will likely run out of energy if we have two days in succession that are FAR more overcast than the very cloudy weather we've recently had.  My estimate is that that is perhaps a once a year occurrence.

2) By frequency shifting, the PW very quickly and gracefully shuts down PV when needed.  That is, when PV power overwhelms the local demand including the possible 5kw  battery charging.  Even in this cloudy weather, battery charging is generally complete by about noon so the PW cycles PV on and off all afternoon.  During typical sunny weather, my full PV system supplies far more than the 5kw that can be used for PW charging.  Under those conditions, the PW would cycle PV on and off all during the day.

3) A problem is that it takes the PV inverters about 5 minutes to recover or come back on after a shut down.  That means that local demand can not exceed the 5kw PW supply limit during those 5 minute periods. 

4) The consequence of the PW being asked for more power than it can supply is PW shutdown.  Which leaves the house without power for a few minutes until the PW attempts to restart.  When the PW shuts down, the observer becomes blind.  He can not tell what the PW is doing; he can not tell what the PV is doing.

5) During the test, I avoided PW shutdowns.  But just barely.  My 5kw water heater operated during battery charged period so PV was being turned on and off every ~5 minutes.  Early in the day, when the PW was charging, the battery provided a buffer of available power.  That is, during the battery charge period more than 5kw loads can be accommodated.  That might be water heating or car charging.

6) For the test, I did not attempt to make use of the additional power available during the battery charging period.

7) Should we experience a long term grid outage in the near term, I would first lower my water heater demand well below 5kw which would remove the threat of PW shutdown due to high demand.  For the longer term, I plan to add either one or two PW batteries; that would increase my available PW power from 5kw to 10kw or 15kw.

8) I don't know whether the frequent cycling of PV inverters is damaging to the inverters.  In case it is, I would likely turn off large sections of my PV during a long term grid down.  That would both reduce the number of inverters being cycled and reduce the period of the cycling.  Better matching PV capacity to the demand.

9) I should mention that I have changed two demands for the test.  Normally, when grid connected, I allow 36v golf cart charging and 48v lawn mower charging during day light hours.  Neither was used during the test.  I do not normally charge cars at the house since the cost is much lower at another of my meters.  Each of the chargers is 1-1.5kw which does not threaten the off grid 5kw limit.  But, power demand would be too high with either charger plus the water heater.  So, I would likely set my charger timers to morning hours or other to not allow water heating at the same time.

4pm, Saturday 01/23/2021.  I thought I could stop.  But, I just can't.  Test still running.  Yesterday was pretty sunny and I decided to see how the PW behaved with full PV cranking.  Starting about noon, after the PW was fully charged, and about once an hour during high production, the PW shut down leaving me without power.   I watched carefully to see if I should turn off some PV.  But, it behaved fairly gracefully, taking 5-10 minutes each time to get back to normal.  So, I am adding 10) and will continue test at least until Sunday morning.  That will be 5 days and 6 nights by my reckoning.  On second thought, I think I'll do a little car charging tomorrow before going back on grid.  I have a nattering nabob full of shit friend who has opined that a PW could not be used for car charging while off grid.

10) With 20+kw of PV power, the PW can fail to get PV shut down promptly enough.  Leaving the house without power for a few minutes.  So, during real long term grid outages with my 5kw PW, I will have to disconnect some PV.

 

 

If you find any of the above confusing, post a comment and I will either edit or respond to the comment.