Starting a thread to document behavior I'm seeing using the new multiple peaks configurations (really multiple of any type). Maybe this information exists elsewhere but I couldn't find it when I first had my system go live a couple years ago, and I'm not really seeing if for the multiple peak configuration capabilities either. Please let me know if this is duplicative. And I know this is totally TLDR, but I wish I would have found this type information, spelled out in an example like this when I was considering my system.
For this first example I'm looking at how Mid-Peak behaves when placed between two true PEAK utility windows that are not consecutive. My utility has no mid-peak window (the red Peak windows are actual utility cost windows though with large demand charges for ANY usage). I'm configuring the mid-peak because I have more solar and PW capacity than I need in the winter, and I want to test prioritizing of running the house with extra solar.
Question being answered: Will the system determine when l have enough PW juice (SOC) to cover my peaks, including covering my configured backup reserve level (all shown configured below), and if so will it then power my house from Solar, from Grid, from the PW or what - even if my PW SOC is not 100%?
Rates shown as configured in the snapshot below are roughly in alignment with my utility plan, but not exact (not enough decimal places), but I don't think it matters as long as they are close and correct relative to each other across peak/non-peak types - since I'm more worried about system behavior than I am in tracking actual costs at this point.
My System: 4PWs, 4.1kW PV - whole house backup (no sub circuits for backup)
My Multiple-Peaks and Off-Peaks Configuration is as Follows and includes a 35% reserve I don't change during this example on this day.
Sending/selling Solar Power (not Powerwall Power) back to the grid (yellow downward peak below): Before my true utility Peak period ended at 9am (with high demand charges), the system began selling back power to the grid, only any additional solar above and beyond what the house needed for house loads. Which was perfect - since it never pulled from the grid during this time and only sold back extra. I believe if the sell-price was set to zero (instead of $0.10 as shown in the configuration above) it would have simply sent the extra to the PW. I have more testing to do in this regard, but this is the behavior I believe I have seen when inadvertently testing with mid-peak. Mid-peak would sell back to the grid if I configured a sell value, and stop selling if I made the sell value zero - but I have more testing to do in a more planned and controlled manner to confirm this.
After my morning peak (red in the configuration snapshot at the top) that ended at 9am, my SOC was about 50%, and the Mid-Peak window (yellow bar in the top config snapshot) became active. At that point the system sent all Solar power to the PW and all house loads were handled by the grid. This makes sense, it needs to get enough SOC to cover upcoming Peaks during my much cheaper mid-peak, so the real-time usage looked something like this (below):
It really didn't matter if my home usage went up or down all solar went to the PW and the house was powered by the Grid for most of the morning, thus the two snap shots above with slightly different house usage but system behaving the same no matter. I grabbed these particular snaps fairly close together, but this was what it was doing all morning during my configured mid-peak as usage went up and down.
Around 12:47 (below, and still during my Mid-Peak window) my capacity must have hit the point where the system decided I now have enough SOC to cover my evening Peak-Window (that will start at 5pm) and the following day's morning peak window, over and above my set reserve. On normal days during the winter my home, during the two daily peaks seems to use between 20 and 25% of my PW capacity. So what the systems seems to be doing here is charging until I have about about 30% SOC over and above my reserve plus the two peak period requirements - so it includes some buffer 5-10%.
So now (around 12:47) the system has started running the home off of solar, with any extra being used to continue to top off the PW. After another 1% or so was added to the PW SOC, I started my car charging very slowly (5amps, maybe ~2.5 kWs), at which point it's clearly prioritizing running the house & car charging loads off of Solar (snapshot 2). Eventually it started pulling also from the PW (3rd snap below), but only for a little while. By 1pm (4th snapshot) it must have decided it was going to impact SOC needs for my Peaks and Reserve if usage continued at this level and it went back to powering the home off of only Solar and Grid.
After slowly increasing my car charging rate up to about 7.4kW, and seeing no change in behavior, I stopped the car charging and the system went back to topping off the PW with any extra solar.
In a nut shell, this is how I expected the system to work with configured Peaks (now including multiple daily peaks) and Mid-Peaks ever since my home system went live, but this first time I seem to be able to reliably configure this behavior in a repeatable manner in the winter when we have two non-consecutive peak periods in the same day.
Additionally, and I think this is a very important capability, until now I've never been able to configure the system to send extra solar back to the grid during the peak unless my SOC was over 100% which rarely occurs in the summer during peak periods for my particular setup. And when it did occur it would be during off-peak which is not when I want to send power back to the grid. It's more valuable to send it back during peak, and now I can pretty much assure that's when extra solar will go back to the grid. Also in the winter the PW for my system will tend to sit at 100% during a large part of my daytime off-peak period (and on weekends), but I'd prefer the PW not sit at 100% for battery longevity/health reasons, and for round trip PW loss/efficiency reasons. In the winter I WANT to send extra solar back to the grid before I have to max out the PW SOC. It's just wasted efficiency and wear/tear for the system to sit at 100% constantly sipping down a few percents when my load jumps up, and then recharging as house and solar loads bounce around. As shown above, the bouncing around for my system now happens around SOC of 66%, which is almost the IDEAL SOC to do that bouncing for all of these reasons. So the way it's working now is pretty much ideal for my situation using these new configuration options. In the summer months (about 6 months of the year) I do need the full PW capacity, but for the winter 6 months this optimization seems to work well.
Hopefully this example is informative, with more to come if so. But also having said and written all of this, it seems we all have different PW capacities, different amounts of installed and actual PV generation, different reserves set, different home loads, and different timing of our home loads, different rate plans, etc - so all of these differences will result in different behaviors, and in particular different timing of these observed behavior from one system to the next.
My next testing, unless somebody beats me to it will be to test this same scenario above, but by using only off-peak instead of mid-peak to see how the system behaves compared to this test I've done the last few days, but with all other values being the same. And then further in the future I'll play with these scenarios, but with different buy-sell values to see if that forces some different behavior across Peak/Mid/Off-Peak types.
For this first example I'm looking at how Mid-Peak behaves when placed between two true PEAK utility windows that are not consecutive. My utility has no mid-peak window (the red Peak windows are actual utility cost windows though with large demand charges for ANY usage). I'm configuring the mid-peak because I have more solar and PW capacity than I need in the winter, and I want to test prioritizing of running the house with extra solar.
Question being answered: Will the system determine when l have enough PW juice (SOC) to cover my peaks, including covering my configured backup reserve level (all shown configured below), and if so will it then power my house from Solar, from Grid, from the PW or what - even if my PW SOC is not 100%?
Rates shown as configured in the snapshot below are roughly in alignment with my utility plan, but not exact (not enough decimal places), but I don't think it matters as long as they are close and correct relative to each other across peak/non-peak types - since I'm more worried about system behavior than I am in tracking actual costs at this point.
My System: 4PWs, 4.1kW PV - whole house backup (no sub circuits for backup)
My Multiple-Peaks and Off-Peaks Configuration is as Follows and includes a 35% reserve I don't change during this example on this day.
Sending/selling Solar Power (not Powerwall Power) back to the grid (yellow downward peak below): Before my true utility Peak period ended at 9am (with high demand charges), the system began selling back power to the grid, only any additional solar above and beyond what the house needed for house loads. Which was perfect - since it never pulled from the grid during this time and only sold back extra. I believe if the sell-price was set to zero (instead of $0.10 as shown in the configuration above) it would have simply sent the extra to the PW. I have more testing to do in this regard, but this is the behavior I believe I have seen when inadvertently testing with mid-peak. Mid-peak would sell back to the grid if I configured a sell value, and stop selling if I made the sell value zero - but I have more testing to do in a more planned and controlled manner to confirm this.
After my morning peak (red in the configuration snapshot at the top) that ended at 9am, my SOC was about 50%, and the Mid-Peak window (yellow bar in the top config snapshot) became active. At that point the system sent all Solar power to the PW and all house loads were handled by the grid. This makes sense, it needs to get enough SOC to cover upcoming Peaks during my much cheaper mid-peak, so the real-time usage looked something like this (below):
It really didn't matter if my home usage went up or down all solar went to the PW and the house was powered by the Grid for most of the morning, thus the two snap shots above with slightly different house usage but system behaving the same no matter. I grabbed these particular snaps fairly close together, but this was what it was doing all morning during my configured mid-peak as usage went up and down.
Around 12:47 (below, and still during my Mid-Peak window) my capacity must have hit the point where the system decided I now have enough SOC to cover my evening Peak-Window (that will start at 5pm) and the following day's morning peak window, over and above my set reserve. On normal days during the winter my home, during the two daily peaks seems to use between 20 and 25% of my PW capacity. So what the systems seems to be doing here is charging until I have about about 30% SOC over and above my reserve plus the two peak period requirements - so it includes some buffer 5-10%.
So now (around 12:47) the system has started running the home off of solar, with any extra being used to continue to top off the PW. After another 1% or so was added to the PW SOC, I started my car charging very slowly (5amps, maybe ~2.5 kWs), at which point it's clearly prioritizing running the house & car charging loads off of Solar (snapshot 2). Eventually it started pulling also from the PW (3rd snap below), but only for a little while. By 1pm (4th snapshot) it must have decided it was going to impact SOC needs for my Peaks and Reserve if usage continued at this level and it went back to powering the home off of only Solar and Grid.
After slowly increasing my car charging rate up to about 7.4kW, and seeing no change in behavior, I stopped the car charging and the system went back to topping off the PW with any extra solar.
In a nut shell, this is how I expected the system to work with configured Peaks (now including multiple daily peaks) and Mid-Peaks ever since my home system went live, but this first time I seem to be able to reliably configure this behavior in a repeatable manner in the winter when we have two non-consecutive peak periods in the same day.
Additionally, and I think this is a very important capability, until now I've never been able to configure the system to send extra solar back to the grid during the peak unless my SOC was over 100% which rarely occurs in the summer during peak periods for my particular setup. And when it did occur it would be during off-peak which is not when I want to send power back to the grid. It's more valuable to send it back during peak, and now I can pretty much assure that's when extra solar will go back to the grid. Also in the winter the PW for my system will tend to sit at 100% during a large part of my daytime off-peak period (and on weekends), but I'd prefer the PW not sit at 100% for battery longevity/health reasons, and for round trip PW loss/efficiency reasons. In the winter I WANT to send extra solar back to the grid before I have to max out the PW SOC. It's just wasted efficiency and wear/tear for the system to sit at 100% constantly sipping down a few percents when my load jumps up, and then recharging as house and solar loads bounce around. As shown above, the bouncing around for my system now happens around SOC of 66%, which is almost the IDEAL SOC to do that bouncing for all of these reasons. So the way it's working now is pretty much ideal for my situation using these new configuration options. In the summer months (about 6 months of the year) I do need the full PW capacity, but for the winter 6 months this optimization seems to work well.
Hopefully this example is informative, with more to come if so. But also having said and written all of this, it seems we all have different PW capacities, different amounts of installed and actual PV generation, different reserves set, different home loads, and different timing of our home loads, different rate plans, etc - so all of these differences will result in different behaviors, and in particular different timing of these observed behavior from one system to the next.
My next testing, unless somebody beats me to it will be to test this same scenario above, but by using only off-peak instead of mid-peak to see how the system behaves compared to this test I've done the last few days, but with all other values being the same. And then further in the future I'll play with these scenarios, but with different buy-sell values to see if that forces some different behavior across Peak/Mid/Off-Peak types.
Last edited: