Powerwalls during PG&E PSPS 2020-10-25

[bmah]

We just came out of a 40-hour PG&E PSPS event that was originally supposed to last around 24 hours. Powerwalls and solar worked great. Our solar system is slightly undersized, it was not designed with this use case in mind. But it was good enough to keep the lights on, fridge cold, and HVAC system warm. Not the longest outage we've seen (that was 61 hours) but it was definitely non-trivial. Based on usage patterns in this and other PSPS events, I think we could run for about 4-5 days, maybe longer if we could shed more load.

I just took a screenshot of the Grafana dashboard I use for monitoring the Powerwalls (it uses some custom golang code that I wrote to poll the gateway on its local interface). At the bottom of the dashboard is a 2-day time history of the state of charge (right side axis) and power (left side axis). Might be of interest to some people here.

Bruce.

 

[miimura]

Nice. Lowest SOC appears to be about 40%. That's comfortable. The only fly in the ointment would be if there was an actual fire that caused smoke to reduce your solar production. During a Winter storm, clouds reducing production would be the major concern for extended runtime.

 

[bmah]

Nice. Lowest SOC appears to be about 40%. That's comfortable. The only fly in the ointment would be if there was an actual fire that caused smoke to reduce your solar production. During a Winter storm, clouds reducing production would be the major concern for extended runtime.

Agreed. I got concerned over the weekend when we had a lot of overcast extending into the early afternoon, but on the actual outage days it worked out OK.

Bruce.

 

[SMAlset]

Bruce curious what type of heat source you have, gas or electric (assume not baseboard water heat—I had that in Illinois condo)?

Our house inside was 66F this a.m. when we woke up and turned our heat on for the first time. Our gas furnace will be functional in a power shut off with PWs to keep the electric to the starter and blower fan available. For us that should make Summer/Fall with the A/C on the largest power draw time during the year and less of an issue with reduced sun angle and charging hours during the winter.

 

[JPP]

Did your ISP go down?? Most ISPs seem to only be able to keep their local switches and amps up for like 24 hours.

 

[bmah]

Bruce curious what type of heat source you have, gas or electric (assume not baseboard water heat—I had that in Illinois condo)?

Our house inside was 66F this a.m. when we woke up and turned our heat on for the first time. Our gas furnace will be functional in a power shut off with PWs to keep the electric to the starter and blower fan available. For us that should make Summer/Fall with the A/C on the largest power draw time during the year and less of an issue with reduced sun angle and charging hours during the winter.

We have a gas furnace (two actually). So as you pointed out, the fan requires electricity, but not the heat itself. This would have been very un-fun on a hot day in the summer because our AC compressors aren't backed up.

Did your ISP go down?? Most ISPs seem to only be able to keep their local switches and amps up for like 24 hours.

AT&T fiber (resold by sonic.net) is our primary ISP, Comcast is our backup. Comcast was only up for a few hours at most. AT&T fiber is a GPON network, which is basically passive from the subscriber's house all the way back to the ... central office? Anyway with that type of network, they only need to power a few larger installations with generators/batteries, rather than a bunch of small amps like with a cable TV (HFC?) network.

It was pretty much the same way for the two PSPS incidents we had last year...our fiber stayed up the whole time, but all our neighbors on Comcast went down.

I think the cell towers in the area were powered, although I wasn't paying close attention to that.

Bruce.

 

[JPP]

AT&T fiber (resold by sonic.net) is our primary ISP, Comcast is our backup. Comcast was only up for a few hours at most. AT&T fiber is a GPON network, which is basically passive from the subscriber's house all the way back to the ... central office? Anyway with that type of network, they only need to power a few larger installations with generators/batteries, rather than a bunch of small amps like with a cable TV (HFC?) network.

It was pretty much the same way for the two PSPS incidents we had last year...our fiber stayed up the whole time, but all our neighbors on Comcast went down.

I think the cell towers in the area were powered, although I wasn't paying close attention to that.

Bruce.

Nice. We only have Comcast here--no fiber. So during PSPS we lose internet and VoIP phones after a few hours. And ATT cell service is spotty--before the towers go off line. My backup internet is HughesNet (big dish on roof), with about 20 megs down/5 up, a really small data cap, big latency, and $70/month. Can't wait for Starlink.

 

[jboy210]

How does your app know when you took a shower and had to blow dried your hair? Is it sentient and has a camera?

On a more serious note, I was thinking of you when they said Lafayette was impacted by the PSPS. Glad to see the PWs did their job. Do you think you could have made it with just 2 PWs?

 

[Dave EV]

We just came out of a 40-hour PG&E PSPS event that was originally supposed to last around 24 hours. Powerwalls and solar worked great. Our solar system is slightly undersized, it was not designed with this use case in mind. But it was good enough to keep the lights on, fridge cold, and HVAC system warm. Not the longest outage we've seen (that was 61 hours) but it was definitely non-trivial. Based on usage patterns in this and other PSPS events, I think we could run for about 4-5 days, maybe longer if we could shed more load.

I just took a screenshot of the Grafana dashboard I use for monitoring the Powerwalls (it uses some custom golang code that I wrote to poll the gateway on its local interface). At the bottom of the dashboard is a 2-day time history of the state of charge (right side axis) and power (left side axis). Might be of interest to some people here.

View attachment 602916

Bruce.

Great data!

Looks like you have a pretty high baseload of energy in your house at 537W. Over 40 hours, that's 21.5 kWh or about a third of your total energy consumption. Knock that down and you should be able to go quite a bit longer in a power outage.

 

[miimura]

Great data!

Looks like you have a pretty high baseload of energy in your house at 537W. Over 40 hours, that's 21.5 kWh or about a third of your total energy consumption. Knock that down and you should be able to go quite a bit longer in a power outage.

I am 600W-700W of background consumption. In an outage I knew would be going for a while, I would shut down my RAID media server and desktop PC. I would also discourage use of the 65" plasma TV and home theater receiver, which cost about 650-700W. The bedroom 55" LCD is probably <50W.

 

[jboy210]

We are closer to 600W (NAS, 3 Access points, UDM Pro). Then every morning the water heater recirc pump and bathroom floor heater starts up. They can get us up to 2 KW.

 

[SoundDaTrumpet]

Great data!

Looks like you have a pretty high baseload of energy in your house at 537W. Over 40 hours, that's 21.5 kWh or about a third of your total energy consumption. Knock that down and you should be able to go quite a bit longer in a power outage.

@bmah Nice data. My baseload is 500W-600W with the 4 of at home doing work/school. My 2 PWs consume about 60-100W themselves, so there is not much to trim. The home network and IoT consume a good amount.

 

[h2ofun]

We have a gas furnace (two actually). So as you pointed out, the fan requires electricity, but not the heat itself. This would have been very un-fun on a hot day in the summer because our AC compressors aren't backed up.



AT&T fiber (resold by sonic.net) is our primary ISP, Comcast is our backup. Comcast was only up for a few hours at most. AT&T fiber is a GPON network, which is basically passive from the subscriber's house all the way back to the ... central office? Anyway with that type of network, they only need to power a few larger installations with generators/batteries, rather than a bunch of small amps like with a cable TV (HFC?) network.

It was pretty much the same way for the two PSPS incidents we had last year...our fiber stayed up the whole time, but all our neighbors on Comcast went down.

I think the cell towers in the area were powered, although I wasn't paying close attention to that.

Bruce.

This is why I have trying to get options for all options. I converted my propane heaters/air to milni split heat pump heaters/air. So now 100% electricity. When I lost power the last 2 days, I turned on my whole house 22K generator, and ran my ENTIRE house, including 4 mini split heads for heat since my house was 59 degrees. (I could have ran all 9 heads if I wanted.) Not sure I could do this easily, cheaply with PW's at all or for any length of time. This is why if I can get the PW's, I still have my generator ready for when the PW's will not cut the need I have in my house to keep the boss happy

 

[jjrandorin]

I am 600W-700W of background consumption. In an outage I knew would be going for a while, I would shut down my RAID media server and desktop PC. I would also discourage use of the 65" plasma TV and home theater receiver, which cost about 650-700W. The bedroom 55" LCD is probably <50W.

I would leave my Nas on (thats how I would likely watch TV), but I definitely would not be using my main living room TV + Surround sound. That TV is an energy hog. My bonus room TV is much more energy efficient. My normal "energy run rate" us around 1 to 1.2 kWh an hour, but in the middle of the night my home runs at 500-600w and I could likely run at 600-700w during the day by just using the bonus room TV.

@bmah very interesting data, thanks for sharing, and glad you came through ok!

 

[holeydonut]

That's really cool - does the Gateway (the black thing, not the Energy Gateway) have an open API to allow access to this data? I think this type of info would go a long way to managing energy consumption.

 

[BrettS]

That's really cool - does the Gateway (the black thing, not the Energy Gateway) have an open API to allow access to this data? I think this type of info would go a long way to managing energy consumption.

Unfortunately it does not. The zigbee gateway is really more of a router. It allows the inverter and the powerblaster to communicate to tesla over the internet, but it doesn’t collect or process the data at all.

 

[jjrandorin]

That's really cool - does the Gateway (the black thing, not the Energy Gateway) have an open API to allow access to this data? I think this type of info would go a long way to managing energy consumption.

the black gateway doesnt have any data accessible to it. I believe @bmah is pulling data from the API to a custom dashboard he created for himself.

 

[holeydonut]

What system/database/thing has the API that @bmah is accessing? I don't have anything set up yet so I don't really know where data is recorded and housed.

 

[bmah]

What system/database/thing has the API that @bmah is accessing? I don't have anything set up yet so I don't really know where data is recorded and housed.

I have some custom code that polls the API documented in this thread:

Powerwall 2 Gateway API Documentation

The API endpoint is available on the Powerwall Gateway's local LAN interface (basically whatever address it gets on your WiFi or Ethernet). I have a Gateway 1, not sure if the Gateway 2 is similar. This code is pwimport in this Github repository...it's admittedly not the best-written golang code, and it wasn't really written with distribution in mind.

bmah888/gotesla

Anyway, it sucks the meter data (plus some others) into an InfluxDB database, and then Grafana draws the pretty pictures based on that. All of this runs in containers in a Linux VM on one of my home servers (FreeBSD and Virtualbox).

Segue to the question of base power consumption that @Dave EV raised...

That server (small desktop Supermicro box) is one of the reasons my base power consumption might appear high, although it is apparently not high compared to some other posters. Another reason might be that I have a lot of stuff on my home network...lots of Ubiquiti switches and access points, multi-camera home security system with a DVR. We kept TVs, another VM server, and desktop computers off. Most of our lighting is LED, and we just didn't use the few remaining incandescent lights. The most critical appliance is our refrigerator. Probably the easiest thing we could tweak is our HVAC, but we just replaced our thermostats and I'm still developing intuition about how these work.

Glad I could start some good discussion, keep it coming...

Bruce.

PS. Also interesting was that this PSPS was much more fine-grained than the ones we had last year that took out almost the whole Bay Area (and I do respect the people in the North Bay and further north who have done the PSPS dance many more times). So the western part of Lafayette was dark, but part of the business district was up, and I think the eastern and southern areas weren't affected at all.

 

[roblab]

3 powerwalls here. Solar will charge PW IF IF IF the grid is on, otherwise, no. Lotta good that does! PWs shut down from using up all the batteries, and we live in the dark.