Oversized System - Best Routine?

[Redhill_qik]

I switched to TBC and trying to figure out what % split I should be using for backup vs time-based control. Also, how does the system work compared to Self-Powered?

When I switched to TBC, it looks like all the PV was sent straight to the battery and whatever the house needed, it pulls it from the grid but I would sometimes see it flip to where the PV was covering the house, sending some back to the grid and then the majority going to charging the battery. Does the system do this to keep the grid consumption at net 0 until the batteries are full and then it will export it all to the grid? What will happen when PV system is gone? it will always pull from the batteries or is it going to pull from the grid?

TBC tries to minimize the costs, but not agressively, so mostly discharging during peak and rarely during other times. The behavior is very dependent on the utility pricing and the built-in price numbers usually cause more discharge than is ideal as the system doesn't understand NEM rules, if you are a net exporter vs exporter and NBCs. So I use different numbers to get the behaviour that I want which is discharge to house load during peak with solar being exported to my reserve level (max credits), with solar going to the house first in the morning with excessive to the Powerwall for recharge (min NBCs).

 

[HardHitter]

TBC tries to minimize the costs, but not agressively, so mostly discharging during peak and rarely during other times. The behavior is very dependent on the utility pricing and the built-in price numbers usually cause more discharge than is ideal as the system doesn't understand NEM rules, if you are a net exporter vs exporter and NBCs. So I use different numbers to get the behaviour that I want which is discharge to house load during peak with solar being exported to my reserve level (max credits), with solar going to the house first in the morning with excessive to the Powerwall for recharge (min NBCs).

On October 21, I switched my settings from Self-Powered to Time-Based Control. So just looking at the last 3 weeks of data, this week still with 2 more days left I get the following table:

1. Oct 10 -16
   1. Solar Generated: 294.8kWh
   2. Home Usage: 156.6kWh
   3. PowerWall Discharge: 72.9kWh
   4. Net Grid Use: -125.5kWh
2. Oct 17 - 23
   1. Solar Generated: 345.7kWh
   2. Home Usage: 150.8kWh
   3. PowerWall Discharge: 62.8kWh
   4. Net Grid Use: -184.6kWh
3. Oct 24- 30 (Note this week is still missing 2 days of data 29th and 30th)
   1. Solar Generated: 253.6kWh
   2. Home Usage: 166.8kWh
   3. PowerWall Discharge: 28.4kWh
   4. Net Grid Use: -81.7kWh

In the limited data that I have, it seems as though Self-Powered mode seems to provide a better Net Grid Use result of exporting into mid to low 100kWh/week where the first full week of Time-Based Control (Oct 24 - 30) will seem to fall shy on the Net Grid Use export results. Time-Based Control drains my batteries down to around only 75-80% since its only discharging from 3:00 PM - Midnight, where Self-Powered will have my batteries down to around 65-70% by the time solar starts charging them the next day.

I guess I can do something like 1 month Time-Based Control and another month Self-Powered to get better data to compare against, but what do you take away from this data? Is Time-Based Control really better than Self-Powered for our use case?

My take away from this is that Time-Based Control will keep ~10% advantage on battery charge but you'll be pulling from the grid generating NBC and you'll end up with lower net grid export in comparison to Self-Powered mode where you sacrifice having 10% less battery reserve by the time solar starts charging but you'll never generate a NBC and seem to have higher net grid export.

 

[RKCRLR]

I'm a net producer on PG&E EV2A and I've found that the best balance between cost savings and Powerwall wear for my use is Time Based Control with peak set to 4-9pm and off peak for all other hours. This yields NBCs greater than MDCs by ~$5/month on average. Last true-up I had ~$50 net surplus credit. My goal is the elusive NBCs=MDCs and zero net surplus.

Tesla warranties the Powerwalls for 37.8MWh of aggregate throughput in certain applications. While the warranty aspect doesn't apply for our use it gives an indication of what Tesla considers the design life. If you divide 37.8 MWh by the 13.5 kWh capacity you get exactly 2800 cycles.

At some point Powerwall age will become the determining factor of life but we don't know where that point is. If you believe Powerwalls will last more than the 10 year warranty life then you also need to estimate how long you think they will last if treated right. For me, 20 years seems like a reasonable goal and it also coincides with the current NEM grandfathering period.

If you divide 37.8 MWh by 20 years you get 1.89 MWh/year (140 cycles/year) per Powerwall. Or a ~38% average daily discharge. So under these assumptions if you limit your Powerwall use to this it would last 20 years. And the if your Powerwall use is less than this then worrying about excess wear due to a net surplus becomes less important. Of course, there a lot of assumptions about Powerwall life here but it is a starting point. And it ignores the impact of storage capacity degradation at the end of its life but if anyone has better data to base Powerwall life on I'd like to see it.

 

[BGbreeder]

I'm a net producer on PG&E EV2A and I've found that the best balance between cost savings and Powerwall wear for my use is Time Based Control with peak set to 4-9pm and off peak for all other hours. This yields NBCs greater than MDCs by ~$5/month on average. Last true-up I had ~$50 net surplus credit. My goal is the elusive NBCs=MDCs and zero net surplus.

Tesla warranties the Powerwalls for 37.8MWh of aggregate throughput in certain applications. While the warranty aspect doesn't apply for our use it gives an indication of what Tesla considers the design life. If you divide 37.8 MWh by the 13.5 kWh capacity you get exactly 2800 cycles.

At some point Powerwall age will become the determining factor of life but we don't know where that point is. If you believe Powerwalls will last more than the 10 year warranty life then you also need to estimate how long you think they will last if treated right. For me, 20 years seems like a reasonable goal and it also coincides with the current NEM grandfathering period.

If you divide 37.8 MWh by 20 years you get 1.89 MWh/year (140 cycles/year) per Powerwall. Or a ~38% average daily discharge. So under these assumptions if you limit your Powerwall use to this it would last 20 years. And the if your Powerwall use is less than this then worrying about excess wear due to a net surplus becomes less important. Of course, there a lot of assumptions about Powerwall life here but it is a starting point. And it ignores the impact of storage capacity degradation at the end of its life but if anyone has better data to base Powerwall life on I'd like to see it.

Given that nobody has 20 year data on Powerwalls, this seems like a very reasonable analysis.

All the best,

BG