How do I get accurate usage data to pick SDG&E rate plan for existing system?

[so_cal_burbs]

My system has hit its five year anniversary which means SDG&E will be kicking me off of tiered use and onto a TOU plan. I have read some threads/posts about building a spread sheet but haven't seen how to pull apart the data in a way that allows me to see Generation/Usage on an hourly basis which is what I would need to compare things like EV-TOU2 to EV-TOU5. I'm kinda surprised since I am signed up to multiple ways to monitor my energy usage that I can't find any that offer that level of detail w/o hundreds of clicks. My goal is to create a revenue generation style sheet that separates usage and generation and let me plug in different rate plans (w/transmission costs, etc...). But I'm stuck at the very first step.

Places I've looked for this data:
SDG&E - Green Button shows Consumption/Generation in 15 minute increments but they only track Net so doesn't appear to be useful.
PVOutput.org - Even as a donor they don't have a web interface to allow download of detailed data other than one day at a time. They do allow daily/monthly summaries. So clicking through 365 days would work. They do support an API that would allow it if anyone knows of a product that leverages it.
Emporia Energy - I installed their system nine months ago and they come closest to giving me what I need. I had to combine their core product with a Vue wireless product and haven't been able to combine the info. I have a support ticket open to see if there is a way to do it.
SolarEdge - Don't see a way to get the data from them.

Trying to get this done to help determine whether I should get more production installed before NEM 3.0 kicks in.

[RichardCullip]

I've done something similar with my SDG&E usage data and SolarEdge production data. I downloaded the GreenButton data from SDG&E's website for a year's worth of data. I then parsed that data into TOU periods with some clever Excel magic (Crosstab Query). I then used SolarEdge's Monitoring API to download solar production data for the same year's worth of data. Once again with some Excel magic I parsed the solar production data into the same TOU periods and put the two together to model what my bills would be under certain rate plans. It's not a trivial task as SDG&E has multiple rate changes in a given year and the TOU periods vary by week day and weekends/holidays and some months of the year. Throw in SDG&E's billing period (mine is near mid-month) and the fun is just getting started. It can be done.

Good luck.

[so_cal_burbs]

The only reports I see available from monitoring.solaredge.com are: 1)Site Commisioning, 2) Periodic AC Energy, 3) Site Status and 4) Modules Mismatch Analysis. I've run each of them and they at best show a roll up summary, nothing that would let me tease out the hourly production. Can you tell me where you found that info?

[RichardCullip]

The only reports I see available from monitoring.solaredge.com are: 1)Site Commisioning, 2) Periodic AC Energy, 3) Site Status and 4) Modules Mismatch Analysis. I've run each of them and they at best show a roll up summary, nothing that would let me tease out the hourly production. Can you tell me where you found that info?

I went to monitoring.solaredge.com, logged into my system and clicked on Admin. This gave me access to generating the API key and Site ID used to query the production data. You need to get familiar with the data available using the SolarEdge API queries. I use Excel to query the data using the API and the details on how to do this are a bit too complex for trying to educate you thru this forum.

If you can't get to the Admin page that generates the API Key and Site Id parameters you need to talk to your installer into granting you more access. Good luck with that.

[OCJ]

SDG&E - Green Button shows Consumption/Generation in 15 minute increments but they only track Net so doesn't appear to be useful.

What do you mean by "only track Net"? You could roll it up to 1-hour but why bother?

Grab your Green Button data for 2022, paste in into Excel, then use something like this to get your TOU periods in a column:

=IFS(OR(WEEKDAY(B15)=1,WEEKDAY(B15)=7,ISNUMBER(MAT CH(B15-DATE(YEAR(B15),1,0),{1,52,150,185,248,315,328,360} ,0))),IFS(C15<TIME(14,0,0),"Super Off-Peak",C15<TIME(16,0,0),"Off-Peak",C15<TIME(21,0,0),"On-Peak",C15<TIME(23,59,59), "Off-Peak"),OR(MONTH(B15)=3,MONTH(B15)=4),IFS(C15<TIME( 6,0,0),"Super Off-Peak",C15<TIME(10,0,0),"Off-Peak",C15<TIME(14,0,0),"Super Off-Peak",C15<TIME(16,0,0),"Off-Peak",C15<TIME(21,0,0),"On-Peak",C15<TIME(23,59,59), "Off-Peak"),TRUE(),IFS(C15<TIME(6,0,0),"Super Off-Peak",C15<TIME(16,0,0),"Off-Peak",C15<TIME(21,0,0),"On-Peak",C15<TIME(23,59,59), "Off-Peak"))

Now you can total up your generation that the utility received, consumption, and net, and you can multiply out by the different rates to see what the differences are between plans. This formula works for 2022 and includes holidays and March/April mid-morning Super Off-Peak.

[so_cal_burbs]

Thank you for the IF statement code, that makes things a bit simpler. I was trying to do it using the revenue model JPM has described where PVWatts is used to estimate production and SDG&E is used to estimate usage then the two blended to determine Revenue/Expense. Your method also works but it is limited to one year of actual weather rather than mathmatically modeled average weather.

But at the end of the day the Net method using only Green Button data got me close enough. As an EV owner I found what others have often posted that EV-TOU5 will be a bit cheaper than EV-TOU2.

[OCJ]

Thank you for the IF statement code, that makes things a bit simpler. I was trying to do it using the revenue model JPM has described where PVWatts is used to estimate production and SDG&E is used to estimate usage then the two blended to determine Revenue/Expense. Your method also works but it is limited to one year of actual weather rather than mathmatically modeled average weather.

But at the end of the day the Net method using only Green Button data got me close enough. As an EV owner I found what others have often posted that EV-TOU5 will be a bit cheaper than EV-TOU2.

I'm actually writing something better in Python, because I do want to merge other data with the Green Button data, namely PVWatts and the system output. I want to model charging under solar vs that 12am to 6am window (or 2pm on weekends/holidays) to skip the NBCs, as well as modeling when to use the battery and when not to for credits vs. battery degradation.

For EV-TOU5, it used to be worse than EV-TOU2 and TOU-DR1 in certain ways because NBCs counted against the MDC, but not against the fixed charge, but since May from what I understand that's no longer the case. Now also, the minimum charge is $.380, so $11.56 per months vs. $16 per month for EV-TOU5, PLUS NBCs. Where before you wouldn't get charged NBCs unless it exceeded your MDC, somewhere around the 465 kWh mark.

Now, TOU-DR1 does have the baseline credit, but I have a battery, so I never use at On-Peak and also send my On-Peak generation for credits. Perhaps I can get my hands on some non-battery Green Button data some day.

I can already see the NEM 1.0 people shrugging about those NBCs...

[CharlieEscCA]

I've got a Windows program that I can share with anyone interested that calculates the exact revenue value of your SolarEdge system for the EV-TOU2 table for your selected range of dates between Jan 1, 2020 and the current rates in place. If your TOU plan is different, you only need to edit a text file that contains the monthly peak / off peak / super off peak total electric rates.

You do need to have a SolarEdge API key to execute.

I can share the program with anyone who is interested, or if you are comfortable giving me your API key (by message), I can run my program and send you back the results.

SolarEdge Jan 2020 to current.png

There are three lines for everyday which I use in my separate spreadsheet which allows my to use with the SDGE day usage output to track my entire usage picture day by day, and where I stand in terms of NEM balance throughout the true up cycle. Combining SDGE usage data with my generation data, I am also able to calculate what my bill would have been if I did not have solar. In fact, it is this "what my bill would have been" which puts me at full ROI payback in the next two months.

[J.P.M.]

So-Cal-Burbs: For those on NEM 2.0, Some thoughts:

1.) On total usage, I've got 15 minute data from my Sunpower monitor (which is nowhere as extensive or useful and it once was) and 15 minute green button data so all that works well with a spread sheet.

2.) I'm on NEM 1.0 and have about 9 1/2 years to go so on tiered rates so for my own use I don't need to fiddle with T.O.U. rates and such for figuring out a bill. I just read my SDG & E meter 1X/day and go on collecting green button data..
A bit of an aside: On the green button data, look for dropouts. They're pretty common to my experience and will quickly screw up a spreadsheet.

3.) I've tried to stay informed about SDG & E rates and tariffs for my own info and also make an effort to try to help neighbors understand their bills.
To those ends I started out several years ago thinking about the same things you now describe.

4.) After several attempts at using the green button data and my (or any of my neighbors') actual data, I got familiar enough with it and wound up with something that gets about as close as Richard gets when I'm explaining a neighbors' bill to them (mostly DR-SES tariff) bills but, as you seem to be finding, accuracy is dependent on the array monitoring system. I believe most of my remaining difference with actual billing comes from some loss of NBC data when I combine the 15 minute data into 1 hour data as described below.

5.) Because the $$ rates for any 4 15 minute intervals within a clock hour are the same, I condense the data and can use an 8,760 row spreadsheet.
I can then use that modified format green button data (checked and corrected/adjusted for dropouts) with a PVWatts hourly output option on the same spreadsheet.

6.) For rate comparison - that is, which tariff results in the lowest bill - since all SDG & E T.O.U. rate tariffs all use the same hour breakdowns for rates over 365 days, plugging in the various rates by simply changing the assignment cell data allows a quick comparison of which rate results in the lowest annual bill. I've found the most revealing, easiest and probably the most accurate way to do this is to use the PVWatts hourly output option for the array output rather than actual data. Doing so avoids some possible/likely problems with formatting monitoring data as well as accounting for weather vagaries which can distort the results by making decisions based on only one or two year's data. Also, one drawback is that the PVWatts data must be adjusted for daylight savings time. Not doing so will screw everything up.


7.) As for calculating (or backing into) a prior bill given green button data, I found it to be a tiresome chore for an 8,760 row spreadsheet, but maybe not as onerous as it sounds.
The task was made quite a bit easier, for me anyway, because there are basically only 2 schedules of 7 days * 24 hour/day = 168 hour hour weeks over 153 days for the summer schedule (06/01 to 10/31) and over 212 days for the winter schedule (11/01 to 05/31). That means once 2 168 hour weekly schedules are done, most of the 8,760, hours is a copy job with care for what day of the week a year starts on (making Jan.1 a Sunday helps), and adjusting for the additional super off peak hours of 1000 to 1400 for March and April as well as making 5 holidays to be weekend rate days even if they fall on a weekday.

8.) I originally did this work to find, given a specific SDG & E T.O.U. rate schedule, how the $$ value of 1 STC kW of PV array's generation varied as f(array location, orientation).
FWIW, in north county San Diego, under NEM 2.0, using the PVWatts model with 10% system losses, inverter(s) sized for as close to 0 clipping as possible and as close to zero array over genration as possible (but under generation being OK), Miramar MCAS irradiance and other weather data and the latest DR-SES rate schedule, the optimum array orientation is at an azimuth of ~ 205 degrees with an array tilt of about 30 degrees. The rolloff in revenue for other orientations is slight for small variations off optimum with the rate of decrease of revenue generation accelerating as the orientations get further away from the optimum.
Currently, all the other SDG & E T.O.U. rate schedules have similar but slightly different optimum (maximum $$ generation per installed STC kW) array orientations due to slightly different per kWh rates, but overall, and at this time anyway, the differences in optimum array orientation as f(different SDG & E T.O.U. rate schedules) are so slight as to be not worth worrying about.

I'm not in San Diego county at this time, but as I seem to recall, the current offset value of 1 STC kW of installed PV in north San Diego county (that is, how much a SDG & E electric bill will be reduced) at optimum orientation is about $712/yr. So, a 5 kW array will offset ~ $712*5 = $3,560/yr.
If anyone wants to know about other orientations, I'll be back in town ~ 02/04 or so.

Lots of interesting data (to me anyway) came out of this exercise. Like the average current value of a residential PV generated kWh over the course of a year vs. the average value of a consumed kWh. Or, how much the value of a residential PV generated kWh has increased since 2019 vs. how much the average cost of a purchased kWh has increased over the same period.

Take what you want of the above. Scrap the rest.

[so_cal_burbs]

Thank you for sharing your thoughts.
Charlie, I'd love to get a copy of your program to play with-how?. JPM, enjoy your trip-thank you for taking time to reply.

I did end up building a spread sheet with my Green Button data and then put in all the likely TOU plans. As described it isn't difficult but it is tedious and took a while. So I can compare total bill from 2022 based on actual net usage/production. But SDG&E has no way to break out separately the generated versus consumed.

Sounds like combining a tool like Charlie has made and put the actual solar production into the Green Button data spreadsheet I can do some basic math to tease out the correct usage/generation for each hour. From there I can substitute PVWatts solar estimation for production and get a solid prediction for the future. I'm going to email the guy (folks?) that run pvoutput.org and request a feature to pull out more detailed data from the website. They have all the info but currently they don't have a way to get it other than through the API.

For future researchers that find this thread my work did allow me to come up with how much additional production/year I'd need to move my bill down to $n but is specific to last year's weather and usage patterns. The commonly accepted wisdom about SDG&E plans proved correct in my case. I have two EVs and the EV-TOU-5 is about 10% less than the TOU-2 and roughly the same that I was paying under Std-DR. I do recommend setting your EV to charge starting at midnight in the months before you switch to a TOU to help make these calculations more accurate. Hmmm, now that I've typed that I think I'll add some new fields to my spreadsheet to let me add/subtract usage from each TOU period to simulate moving usage.

[J.P.M.]

Thank you for sharing your thoughts.
Charlie, I'd love to get a copy of your program to play with-how?. JPM, enjoy your trip-thank you for taking time to reply.

I did end up building a spread sheet with my Green Button data and then put in all the likely TOU plans. As described it isn't difficult but it is tedious and took a while. So I can compare total bill from 2022 based on actual net usage/production. But SDG&amp;E has no way to break out separately the generated versus consumed.

Sounds like combining a tool like Charlie has made and put the actual solar production into the Green Button data spreadsheet I can do some basic math to tease out the correct usage/generation for each hour. From there I can substitute PVWatts solar estimation for production and get a solid prediction for the future. I'm going to email the guy (folks?) that run pvoutput.org and request a feature to pull out more detailed data from the website. They have all the info but currently they don't have a way to get it other than through the API.

For future researchers that find this thread my work did allow me to come up with how much additional production/year I'd need to move my bill down to $n but is specific to last year's weather and usage patterns. The commonly accepted wisdom about SDG&amp;E plans proved correct in my case. I have two EVs and the EV-TOU-5 is about 10% less than the TOU-2 and roughly the same that I was paying under Std-DR. I do recommend setting your EV to charge starting at midnight in the months before you switch to a TOU to help make these calculations more accurate. Hmmm, now that I've typed that I think I'll add some new fields to my spreadsheet to let me add/subtract usage from each TOU period to simulate moving usage.

For what I might have contributed that you found useful, you're most welcome.

As I wrote, I've done stuff quite similar to what you describe with my (and other's) green button data while using the 15 minute periods combined to 1 hour and used the PVWatts hourly output and came to similar conclusions to yours.
I also have all my generation in 15 minute increments and have used that with the green button data to get hourly (or 15 minute if I choose to keep that format) for total usage by hour. As stated previously, the slight bit I'm off on total $$'s I attribute to the smearing of NBC's either up or down a bit when I combine the 15 minute data to 1 hour data. Anyway, by (mostly) confirming prior billing for myself and other usage by neighbors (who are on NEM 2.0 and various T.O.U. schedules) I've satisfied myself that the POCO is probably right almost all the time.
If/until I get an EV, I'm staying on tiered rate schedule DR as long as possible. Better, cheaper, easier to understand. If/When I get an EV, I'll need to reevaluate, but since my billing usage is never likely to get out of tier one until I use > ~ 5,000 kWh/yr. by charging EV's, I'm probably not going to switch rate schedules unless or until things change with CPUC or SDG & E policies.