bkp_duke
Well-Known Member
I received a call to schedule today . . . for March 2018. They said given things in hurricane-affected areas, this is the first available install date.
-
Want to remove ads? Register an account and login to see fewer ads, and become a Supporting Member to remove almost all ads.
Powerwall 2: Installation
- Thread starter JeffreyY
- Start date
-
- Tags
- Powerwall Tesla Energy
This is a pair of AC powerwalls, and I’m getting my data from the tesla mobile app. Two powerwalls are configured for 90% reserve in case of power outage, and the software appears to make every effort to flatline the utility energy draw (+/- 0.1 kw read on Utility meter) from the moment the Solar system energizes until the powerwal is filled to 100%, then it sends excess power to utility. As sun sets or solar production drops below house load, powerwall discharges to keep utility flatline until the powerwall hits 90%. Once the powerwall hits 90%, then house power = utility power draw.
My maximum load is 7.7 kw overnight when the tesla charges. The house/ fridge draws 300-500watts when everything else is off. Max load during the day is typically 2000-3500watts, but often 900watts.
Over the month of November, the two powerwalls charged 90 kwh, and discharged 60 kWh. In October, it was discharge 125 kWh, and charge 160kwh. Both months show 2ac powerwalls consume over 1000 watts per day.
Powerwall software ver started as 1.6, and is now 1.9.1
There is no way to have powerwall charge during off peek and discharge during peek times. (Known feature request) There is no way for it to not discharge 100% into the EV car battery if the utility failed after I went to sleep, and thus leave me dark the next morning when the sun comes up, and the ac input side of the Solar inverter doesn’t see power, and panels remain off.
In other words, I’m disappointed I was not sold a dc powerwall for configuration where 100% of all house loads are configured to run off the powerwall. The perceived concept was: if utility power is gone for 3-7days, can I run around unplugging things and “get by”comfortably on 7-29 kWh per day of Solar power (rainy Winter to sunny summer production),and if absolutely necessary, add 15-20 kWh of range to my EV per day, which is 50+ mikes and should get me to a charger location which has power in the event of some regional catastrophie.
Last edited:
That is correct, powerwall gets 3 new exterior electrical panels
1) backup power breaker box, wired in parallel with backed up loads
2) new backup load center —where all your existing loads are relocated to
3) tesla powerwall brain box, which has a mechanical relay to isolate the home from the utility. If you don’t isolate, and attempt to backfeed power from your house into the grid, you power the neighborhood, and your power supply fails
" There is no way to have powerwall charge during off peek and discharge during peek times. "
No way to charge powerwall during off peak hours? I thought that was the whole point of CA TOU rate arbitrage and for the SGIP program to pay out a fee for you to get payback from the program in order to "perform peak load support". Did you get SGIP money? They won't pay it out until they can validate the install and verify that it can dump 50x a year at least most of the store energy in a two hour period for the rated KW output of the units to qualify for SGIP.
For me, if overnight charging the car, I would avoid using stored battery power from the wall to charge your car. Each kWh from the grid has a price but a kWh pulled from a powerwall costs you more due to charge/discharge cycling. I'd rather see you bank the battery over night, charge the car from the grid using the "grid lull" period (grid runs more efficiently at night and offers TOU rates in CA) - and then your battery is more full in the morning and can Net-meter out more kWh to the grid during peak periods. The theory is - night time grid costs per kWh are say .10 each. Store them in a battery and sell to the grid tomorrow during peak periods of Net Metering at .30 each, for example.
Folks should remember that using their wall batteries at night - when they may have a TOU rate program as well as a more efficient grid isn't a greatest way to utilize powerwalls. Grids are heated during the day and transformers and lines run "slower" due to heat-effects of electric circuits. The best time to charge an EV is at night between 2am and 6am. Cool lines, cooler transformers, less demand all around, no peaker plants running (most likely).
To verify your number:
"Both months show 2ac powerwalls consume over 1000 watts per day."
I see you mean Watt Hours (1kWh). You are saying you are seeing more than 1kWh loss per day - if it is 160kWh in, 125kWh out, that is 1.16kWh lost during charge cycles and you are not doing "full sweep" cycles where the entire State of Charge is used up. Effectively, you are now using more than 1kWh more as a household after the PW were installed than before. Now, the value added is that you may be able to withstand quick power outs and also longer-term power failures nicely. How often did that happen before the install? Can you island the house off Solar PV and never really need the grid to come back for days - meaning do you have enough powerall and solar to run your house in a blackout for weeks situation? If not, then you could up-size things (more Solar PV and another PowerWall). The DC solution is better with the SolarEdge inverters, where someone can basically go off-grid.
Last edited:
Good to know. I will make sure my car chargers are NOT pulled onto the Powerwall circuits when I do my install.
Tesla just told me that, per my request, they've updated my design to keep the existing EV charging circuit in the main panel where it won't be backed up. In the event we actually need to do a bit of EV charging during an extended outage, we'll just use a standard 120V outlet.
In the same email, they clarified that once the "TOU load shifting" feature becomes available, it'll work with every load in the home, not just the backed up loads. Therefore, if we need to charge an EV during TOU peak hours, the Powerwalls will kick in and provide power (provided the grid is up).
The house auto islands off grid when power goes out. Then there is say 20-25 kWh in the battery and 4-35kwh of daily solar production depending upon time of year and daily weather, a nasty dark rainy day is 4 kwh, and cloudless summer solstice is 35 kWh.
As per the current software and configuration options for the ac powerwall in the tesla mobile App, the system is functioning in a manner to maximize the possibility that the powerwall will never pay for itself. They seem to have assumed the goal was to generate an “off grid” idealized home.
Yes approxImately 1.16 kWh of energy loss from a pair of AC powerwalls discharging only 10% of their total energy capacity per day. Ambient temperatures are 55-75 degrees F, units are wall mounted a foot off the ground, and have at least 12-18” horizontal clearance on both sides.
No response on the SGIP rebate, and I’ve been on E-6 tou net metering for 7 years.
I used about 2.3 megawatt hours more than my south facing Solar array generated last year, mostly due to EV charging. 2-4 megawatt hours of power to drive 10-12,000 miles in a tesla is a reasonable annual number... figure 300-360 kWh/mile depending upon driving style and car model/options.
From the mobile app.
Green Is powerwall, energy into powerwall is below the line (negative scale) above the line is energy from the powerwall
Orange is Solar generation
Blue above the line, before 8 am and after 5 pm is house usage and utility supplied power overvalued as they are identical
Grey is grid net metering, and I’ve shown an isolated image for clarification.
There was no EV charging in 24 hours shown.
Attachments
GenSao
Member
I am surprised you are still on the E-6 TOU Rate with an E-vehicle. It makes fiscal sense to switch to the EV-A Rate to take advantage of the low overnight rate (12.2¢ EV-A vs. 16.7¢ E-6 Tier 1). The batteries can be used to mitigate the higher peak and off peak costs from 7 AM to 11 PM. Excess stored energy would be sold at the higher Peak summer EV-A rate of 45.4¢ vs E-6 Tier 1 rate of 35.9¢.
Attachments
I agree. Anybody that charges an EV off-peak will save significantly on EV vs. E-6. I found that the better solar collection on E-6 during the mid-day is more than offset by the cheaper overnight charging. I studied my specific situation using my actual SmartMeter data and found my annual energy cost would have been $1,129 on E-6 and $837 on EV. This was an analysis using my 2013 actual usage and the rates in effect at that time. The tiered rates in the Winter when your solar generates less make EV charging very expensive in comparison to EV.
Could you go off-grid entirely if you wanted to or would it make no sense given the ability to sell to the grid? If Net-metering went away, I can see some homeowners just cutting the cord to the grid in CA given that the net-metering window is at peak-rates now. This hurts the grid and puts undue demand back into the grid. So, I assume Net Metering is to stay for CA. They need to come up with a better TOU rate plan especially for those with hybrid homes like yours. Maybe a super TOU rate where your daytime grid price is .45/kWh and night-time is .09/kWh and you re-charge the batteries during the grid-lull of 4am-6am and produce to the grid all the way through 8 or 9pm when the grid demand is highest in CA (peak is roughly 6pm-9pm). Why are you only cycling 10% of the batteries now? It looks like you take from the grid after 5:30pm for a couple hours and that would be nicely-smoothed if you use the powerwall for an extra hour or two. This would then allow capture of more sun into the batteries in the morning. Was 10% a requirement of the install?
Last years true up bill was $235 after paying the minimum meter fee. This year I’m concerned about powerwall sucking up Solar energy during part peak and peak, and taking its daily 1.x kWh parasitic recharge losses from solar generation.
As to the 10% value, in September that seemed to balance powerwall charge / discharge with roughly the noon or 1 pm rate change. now that it’s cloudy, all bets are off with static capacity percentage setting. Also software version 1.6 refused to adjust that setting once made.
As to the 10% value, in September that seemed to balance powerwall charge / discharge with roughly the noon or 1 pm rate change. now that it’s cloudy, all bets are off with static capacity percentage setting. Also software version 1.6 refused to adjust that setting once made.
Attachments
For those who have completed installations in PG&E land, did you give Tesla power of attorney to register the units with PG&E? I'm finally getting around to reviewing the documents they want me to sign and it includes allowing Tesla to change my rate plan. I don't think they have that plan, but I find it an odd thing to initial.
Any input on what others did here would be appreciated.
Any input on what others did here would be appreciated.
Hi,
Am I understanding correctly that this situation would be due to the EVs draining the PW overnight?
It would be slightly wasteful, but I think an electrician could install a contactor inline with the EV feeds that is controlled by the utility side of the gateway. That would shut off the EV power when the utility goes down. The power waste would come from leaving the contactor energized all the time. This could be reduced by adding an ON/OFF/AUTO switch to disable the contactor when the EVs are unplugged. This takes an extra step, but with a little work, the EV plug holder could also be the disable switch. Hang the cord up, contactor is turned off.
Lotsa, Applicable to your scenario and others, Lazard is changing their views on energy storage based on "parasitic losses" (which should include round-trip efficiency losses during inverter activity and heat-losses.
The Grim Details Lurking in Lazard’s Energy Storage Analysis
The value of keeping the lights-on through a grid outage cannot be quantified, however, their review seems to be simply based on economics of constant-flow systems. Constant-flow systems utilizing Li-Ion will also have to compute annualized slimming of capacity due to daily charge/discharge cycles (ie. peak load shaving and peak-grid support per the SGIP). Each year, the overall system capacity drops a bit through year 10 or so where 3600+ charge cycles may require full battery-set replacement, depending on capacity remaining. A 1MWh storage system today will be roughly 700kWh system in 10 years if using 3000+ cycle NMC cells or LiFEPO4, which offer about 8000 cycles at 1C with 79% system capacity at the end of 8000 (per old A123 Systems testing cycles).
Last edited:
Mostly an issue of incomplete software control over the system. There is a marketing trade off between configuration capability, consumer education, and how many menus and option switches can exist in a mobile app. Maybe the crew here can hack the API, and achieve this...
Ideally, I would like the following: during sunny forecast summer days, charge battery in early am before dawn, and then let Solar go to grid leaving battery at 100% until peek power arrives, then discharge say 10-20% of the battery during peek power rates, more if tomorrow is predicted to be sunny, and storms / High winds are not in the weather forecast. This is what the utility wants.
Then assuming a car knocks out power at 11:00 pm, keep the house and the car backed up, and deplete battery to say 20% total capacity (yes, charge the car using expensive battery power) reserving enough to keep house powered up based upon historical usage, and expected sun rise next morning. Then charge batteries from sun until utility power comes back. Require me to manually charge the car if power is out the next night, again keeping some in reserve. Use the phone and tesla touch screen to notify me that indeed utility power is out because I was clueless with 100% full house ups.
Thus I’ve purchased an intelligent whole house ups, my food in the fridge doesn’t spoil, and also I’m able to consume Solar power during the day which would be lost to a full battery and low daytime unoccupied home use. (Assume typical working family). Furthermore, should an outage last a week.... major earthquake / hurricane, etc, I have ability to add range to my electric vehicle and go on with life like nothing happened.
Thus if the battery had a hypothetical worst case 14 year payback, and 10 yr life, I’ve bought the utility and convenience of whole house UPS, and the price difference was my energy security insurance premium.
Obviously if it’s winter, storms are forecast, and day ahead solar generation is likely to be minimal, then use my cell phone and wake me up at the 11:00 pm power failure to tell me to shut down my unnecessary power loads as I’m In for a week of camping. Keep the fridge and minimal lights / sattellite /broadcast tv going (cable will be out after 2-4 hours if you are lucky). After a week, I’ll have a working fridge, gas heat, microwave, tv, some lights, and lots of new friends,
Ideally, I would like the following: during sunny forecast summer days, charge battery in early am before dawn, and then let Solar go to grid leaving battery at 100% until peek power arrives, then discharge say 10-20% of the battery during peek power rates, more if tomorrow is predicted to be sunny, and storms / High winds are not in the weather forecast. This is what the utility wants.
Then assuming a car knocks out power at 11:00 pm, keep the house and the car backed up, and deplete battery to say 20% total capacity (yes, charge the car using expensive battery power) reserving enough to keep house powered up based upon historical usage, and expected sun rise next morning. Then charge batteries from sun until utility power comes back. Require me to manually charge the car if power is out the next night, again keeping some in reserve. Use the phone and tesla touch screen to notify me that indeed utility power is out because I was clueless with 100% full house ups.
Thus I’ve purchased an intelligent whole house ups, my food in the fridge doesn’t spoil, and also I’m able to consume Solar power during the day which would be lost to a full battery and low daytime unoccupied home use. (Assume typical working family). Furthermore, should an outage last a week.... major earthquake / hurricane, etc, I have ability to add range to my electric vehicle and go on with life like nothing happened.
Thus if the battery had a hypothetical worst case 14 year payback, and 10 yr life, I’ve bought the utility and convenience of whole house UPS, and the price difference was my energy security insurance premium.
Obviously if it’s winter, storms are forecast, and day ahead solar generation is likely to be minimal, then use my cell phone and wake me up at the 11:00 pm power failure to tell me to shut down my unnecessary power loads as I’m In for a week of camping. Keep the fridge and minimal lights / sattellite /broadcast tv going (cable will be out after 2-4 hours if you are lucky). After a week, I’ll have a working fridge, gas heat, microwave, tv, some lights, and lots of new friends,
Assuming the PW can perform the functions mentioned, a separate control system does seem like a logical add-on, with appropriate additional control hardware for loads. Tech support could be a pain though...
Power exchanges like the east coast's PJM and western coast independe nt operator should be able to offer "simple" data elements for predicting your tomorrow-output. Pull up the PJM site here and see that they do predict day-ahead forcasting on peak network load and some hours ahead at least on wind and solar. PJM - Markets & Operations
The grid should be "smart" - and that means predictability. Adaptive, if you will. A home system that has a shortage of batteries (ie. needs hand-holding if things change on a daily basis) will need to have predictions built into how they get things going if a "fault" occurs in the general system. That system includes weather forecasting to a simple degree. We can get data like "Zip-Code-Date-general forecast" such as "99933-12/02-Sunny" from weather aggregators. Someone's home hybrid system could calculate when to recharge batteries base on that. If tomorrow is cloudy, charge batteries from grid at 3-5am to full. If tomorrow sunny, charge batteries from 8a-10a and net-sell the excess. If grid is lost, plan to charge batteries and consume during day but do not allow AC unit or EV charging until grid back up.
The more batteries you have, as off-grid homeowners had planned in the past, such as 5-days of normal operations worth of kWh in the battery set - then all the solar PV is doing is keeping it topped up and selling to grid any excess. Should the grid fail, the system islands and you run "forever" - until it comes back again. The Tesla powerwall sales story does not include planning on full home usage for up to 5 cloudy days (storm front, rainy season, etc.) and to island a typical home, you need from 3-9 powerwalls for 5-days of full home power usage. I know some people using up to 2000 kWh per month so that is a lot of powerwalls for 5-day usage. Critical circuit isolation is always in play and leaving-out AC, EV charging and more can help a home "live longer" in the event of outages. But like my spouse says - if you are going to get a generator, get a "whole house" generator and make it easy. Of course, that raises the price of the system for the convenience. I honestly think that Tesla should offer home-units of the PowerPack (the 100kWh unit or the new 200kWh) for some homeowners. The 5000+ sq foot homes who tend to buy diesel generators for standby power would utilize them well.
The grid should be "smart" - and that means predictability. Adaptive, if you will. A home system that has a shortage of batteries (ie. needs hand-holding if things change on a daily basis) will need to have predictions built into how they get things going if a "fault" occurs in the general system. That system includes weather forecasting to a simple degree. We can get data like "Zip-Code-Date-general forecast" such as "99933-12/02-Sunny" from weather aggregators. Someone's home hybrid system could calculate when to recharge batteries base on that. If tomorrow is cloudy, charge batteries from grid at 3-5am to full. If tomorrow sunny, charge batteries from 8a-10a and net-sell the excess. If grid is lost, plan to charge batteries and consume during day but do not allow AC unit or EV charging until grid back up.
The more batteries you have, as off-grid homeowners had planned in the past, such as 5-days of normal operations worth of kWh in the battery set - then all the solar PV is doing is keeping it topped up and selling to grid any excess. Should the grid fail, the system islands and you run "forever" - until it comes back again. The Tesla powerwall sales story does not include planning on full home usage for up to 5 cloudy days (storm front, rainy season, etc.) and to island a typical home, you need from 3-9 powerwalls for 5-days of full home power usage. I know some people using up to 2000 kWh per month so that is a lot of powerwalls for 5-day usage. Critical circuit isolation is always in play and leaving-out AC, EV charging and more can help a home "live longer" in the event of outages. But like my spouse says - if you are going to get a generator, get a "whole house" generator and make it easy. Of course, that raises the price of the system for the convenience. I honestly think that Tesla should offer home-units of the PowerPack (the 100kWh unit or the new 200kWh) for some homeowners. The 5000+ sq foot homes who tend to buy diesel generators for standby power would utilize them well.
As I see it, a key requirement for a "smart" grid would be market-based electric rates that vary in real time depending on supply and demand, and which allow everyone connected to the grid to participate in the market as both a generator and a consumer. Of course, carbon-based energy sources would need to be taxed appropriately in order to achieve a market that reflects true costs. Higher rates should be paid to generators in close proximity to consumers, thus favoring distributed generation. During periods of high demand, those with generating capacity (batteries, solar, wind, and all other sources) would be incentivized to export to the grid. When rates are low or even negative, those with energy storage systems and EVs would be motivated to charge them. Things are skewed today with fixed TOU rates, the federal ITC, and California's SGIP, but I think a true market-based system is the future.
Given existing, expensive battery backup systems such as Sonnen and even the current Powerwall 2 ($5500 / 13.5 kWh = $407/kWh), it probably makes sense to install only enough batteries for a couple of cloudy days, and/or the minimum needed to achieve whole home backup, and then for those who require extended off-grid cloudy-day operation, back up the batteries with a diesel generator.
However, with battery prices dropping as evidenced by planned Tesla Semi pricing, I'm thinking that Tesla may eventually be able to sell a 100 kWh PowerPack-equivalent (not including installation) to homeowners for something like $15,000 or even less. At that price, it would be virtually impossible to justify the added expense and fuel storage requirements associated with integrating a diesel generator. Also, with cheap, large battery packs, the problem of load shedding becomes less important.
One thing grid operators won't like is inability to regulate. New circuits in solar pv inverters now allow a grid distribution company to have your pv dip its voltage to help overvoltage conditions. Some may need to require regulations of larger batteries at commercial sites as Tesla has found with larger powerpack installs in Ca where the folks at SGnE and SCE are the ones who want to signal for powerpacks to start dumping to the grid upon their command and not the end users choice. Some will also want to regulate EV charging rates such that they can do demand response at sites like superchargers, etc.
My Tesla rep just contacted me confirming my rescheduled installation is on for 12/18-12/19. Looks like it is finally going to happen to me here in the San Bernardino mountains. There were multiple power outages here over Thanksgiving due to SCE "equipment failures".
