Big discrepancy between Powerwall data with SCE usage data

[JWM3]

I have a 10KW solar system and installed 4 Powerwalls to pair with the solar about one month ago.
On my Tesla app, I can see the Net Grid use is negative on most days. However, in my SCE account, the daily usage chart shows I'm using a lot of electricity on most of the days. Especially on peak hours, many days show I used few kWh electricity while the Tesla app shows I use all the electricity from the batteries during peak hours and there is some solar output feed to the grid as well.
Does anyone have an idea why this is happening?

 

[BGbreeder]

Is it possible that your SCE account numbers are actually credits?

 

[Vines]

I have a 10KW solar system and installed 4 Powerwalls to pair with the solar about one month ago.
On my Tesla app, I can see the Net Grid use is negative on most days. However, in my SCE account, the daily usage chart shows I'm using a lot of electricity on most of the days. Especially on peak hours, many days show I used few kWh electricity while the Tesla app shows I use all the electricity from the batteries during peak hours and there is some solar output feed to the grid as well.
Does anyone have an idea why this is happening?

Do you have partial or whole home backup?

If partial backup, are there Neurio CT meters capturing the non backup loads?

 

[JWM3]

Do you have partial or whole home backup?

If partial backup, are there Neurio CT meters capturing the non backup loads?

I have the partial home backup, but the non-backup load only has a NEMA 14-50 plug which is not in use. All of the rest of my home loads are behind the Tesla gateway. Just looked at my meter a few moments ago, and it is showing -2.47 kw feeding to the grid and Tesla app shows solar is feeding 5 kw.

 

[power.saver]

I have the partial home backup, but the non-backup load only has a NEMA 14-50 plug which is not in use. All of the rest of my home loads are behind the Tesla gateway. Just looked at my meter a few moments ago, and it is showing -2.47 kw feeding to the grid and Tesla app shows solar is feeding 5 kw.

Do you have PTO from SCE yet?

 

[tomuo]

It's always the CTs.
The sensors on the main feed are not configured correctly, call your installer.

 

[jjrandorin]

I have a 10KW solar system and installed 4 Powerwalls to pair with the solar about one month ago.
On my Tesla app, I can see the Net Grid use is negative on most days. However, in my SCE account, the daily usage chart shows I'm using a lot of electricity on most of the days. Especially on peak hours, many days show I used few kWh electricity while the Tesla app shows I use all the electricity from the batteries during peak hours and there is some solar output feed to the grid as well.
Does anyone have an idea why this is happening?

You say you had "installed" a month ago, but you dont mention if you have permission to operate (PTO). The reason @power.saver asked this (and its important) is because if you dont have a compatible meter, you can actually be charged for energy you are generating instead of credited for it.

The other thing that tends to happen here is the CTs get installed incorrectly, but in this case it sounds like the tesla app is correctly capturing what is happening but SCE isnt.

So, do you have official PTO from the utility yet?

 

[Vines]

Looks like the CT's for the solar are not configured correctly. Call your installer back out.

 

[JWM3]

I don't have PTO yet, will call Tesla to ask them to send someone check the system. What is “CT”?

 

[Vines]

I don't have PTO yet, will call Tesla to ask them to send someone check the system. What is “CT”?

CT is a Current Transformer, basically an electronic device that reads the power through the wires in your electrical panels, then reports this to your Tesla system.

 

[yblaser]

Looks like the CT's for the solar are not configured correctly. Call your installer back out.

The factor of two difference certainly makes it look like a CT configuration error. I would assume though it's an issue with whatever CT is measuring the grid. If it was a solar CT issue then it would just look like the home was consuming more than it actually was.

 

[Antonc]

I noticed major discrepancies as well, so I downloaded the data from SCE, which is reported in 15 minute increments as kWh, with one data blob being for usage followed by another data blob for the same day reporting kWh received from our rooftop generation (minus what we use for our home & the recharge the Powerwalls). The Tesla app data is in 5 minute increments sampled as kW, so I average the 3 readings (should I have used sum? didn't seem right for sampled kW's) and divide by 4 to get kWh from the Tesla app data. Then I compared the two. The glaring problems included that 1) on days where Tesla says we used ZERO peak-hour grid power, SCE sometimes says we used lots 2) during times when the Tesla app says we are delivering tons of power to the grid, SCE sometimes says we deliver none and 3) the discrepancies are ALWAYS in SCE's financial favor.
I called SCE and they were very defensive and seemed to not understand how batteries work. I was told I'd get a "letter" in 30-40 days (and who said customer service is dead?).
At one point, the "customer service" rep said, "I know what customers like to believe", in reference to me pointing out that the discrepancies were always in SCE's favor. Apparently others on this thread have noticed discrepancies, but I didn't find a satisfactory explanation for why it's happening to such a significant degree. Certainly small differences could be explained as seen above, but this is over 100 kWh difference during the billing period!

 

[Antonc]

I noticed major discrepancies as well, so I downloaded the data from SCE, which is reported in 15 minute increments as kWh, with one data blob being for usage followed by another data blob for the same day reporting kWh received from our rooftop generation (minus what we use for our home & the recharge the Powerwalls). The Tesla app data is in 5 minute increments sampled as kW, so I average the 3 readings (should I have used sum? didn't seem right for sampled kW's) and divide by 4 to get kWh from the Tesla app data. Then I compared the two. The glaring problems included that 1) on days where Tesla says we used ZERO peak-hour grid power, SCE sometimes says we used lots 2) during times when the Tesla app says we are delivering tons of power to the grid, SCE sometimes says we deliver none and 3) the discrepancies are ALWAYS in SCE's financial favor.
I called SCE and they were very defensive and seemed to not understand how batteries work. I was told I'd get a "letter" in 30-40 days (and who said customer service is dead?).
At one point, the "customer service" rep said, "I know what customers like to believe", in reference to me pointing out that the discrepancies were always in SCE's favor. Apparently others on this thread have noticed discrepancies, but I didn't find a satisfactory explanation for why it's happening to such a significant degree. Certainly small differences could be explained as seen above, but this is over 100 kWh difference during the billing period!

Upon completing the analysis of this billing cycle, the discrepancy is over 800 kWh for this bill alone!!! While on many days, the SCE & Tesla data align perfectly, on other days there are significant discrepancies. If this is happening systemically, it would add up to MILLIONS of $$$ being wrongly taken from SCE solar generators and landing in the pockets of ???

 

[yblaser]

Upon completing the analysis of this billing cycle, the discrepancy is over 800 kWh for this bill alone!!! While on many days, the SCE & Tesla data align perfectly, on other days there are significant discrepancies. If this is happening systemically, it would add up to MILLIONS of $$$ being wrongly taken from SCE solar generators and landing in the pockets of ???

I think this is more of an issue with your particular setup than a vast conspiracy by SCE. Personally when I have compared the data between the Tesla app's grid usage for my system and SCE's data it has agreed to better than 99% limited mostly by the significant figures I used to compare.

Do you have any non-backed up loads? If not configured properly these loads could be missing from the Tesla data. There could also be an issue with one of your CTs being bad, misconfigured, or misplaced. It could also be an issue with the SCE meter itself but I think that is pretty rare.

You can see the instantaneous power draw on the SCE meter. Monitor that and the Tesla app while you turn on and off some of you larger loads to try to understand the discrepancy better.

 

[Redhill_qik]

Upon completing the analysis of this billing cycle, the discrepancy is over 800 kWh for this bill alone!!! While on many days, the SCE & Tesla data align perfectly, on other days there are significant discrepancies. If this is happening systemically, it would add up to MILLIONS of $$$ being wrongly taken from SCE solar generators and landing in the pockets of ???

I have a partial home backup which means that some of the loads are in the main panel and are monitored through CTs. The CTs weren't configured correctly at installation causing significant discrepancies in the app with the real data. I was able to get them to mostly fix this, but there are still issues with some loads (I think it is a 120V vs 240V issue).

Start by looking the reported kW from the SCE smart meter display and verify that it is close +/-0.050kW with the Tesla App. If it isn't then you have a configuration issue and need to get Tesla out to fix it.

The Tesla app data is in 5 minute increments sampled as kW, so I average the 3 readings (should I have used sum? didn't seem right for sampled kW's) and divide by 4 to get kWh from the Tesla app data.

This is the wrong way to do it. You should have multiplied each Tesla data point by 5 mins and then divided the sum by 60mins to get kWh.

 

[Antonc]

I have a partial home backup which means that some of the loads are in the main panel and are monitored through CTs. The CTs weren't configured correctly at installation causing significant discrepancies in the app with the real data. I was able to get them to mostly fix this, but there are still issues with some loads (I think it is a 120V vs 240V issue).

Start by looking the reported kW from the SCE smart meter display and verify that it is close +/-0.050kW with the Tesla App. If it isn't then you have a configuration issue and need to get Tesla out to fix it.


This is the wrong way to do it. You should have multiplied each Tesla data point by 5 mins and then divided the sum by 60mins to get kWh.

Thanks for your reply. Maybe I'm missing something, but it seems odd that most days the data from SCE & that from Tesla is spot on in agreement, if my methodology is wrong. 5 minutes * 3 = 15 minutes. 15 minutes = 0.25 hours. kW * hours = kWh. Multiplication by 0.25 = division by 4. Also, because kW is measured as a sample, not as an accumulation of energy, an average of the 3 samples taken during that 15 minute interval seems like the most-logical way to estimate kWh for each 15 minute interval. Summing them would cause a 3X overestimate of the energy produced. But, if someone can prove me wrong there, then it means that SCE has screwed us over 3X worse than I am suggesting.

 

[Redhill_qik]

Thanks for your reply. Maybe I'm missing something, but it seems odd that most days the data from SCE & that from Tesla is spot on in agreement, if my methodology is wrong. 5 minutes * 3 = 15 minutes. 15 minutes = 0.25 hours. kW * hours = kWh. Multiplication by 0.25 = division by 4. Also, because kW is measured as a sample, not as an accumulation of energy, an average of the 3 samples taken during that 15 minute interval seems like the most-logical way to estimate kWh for each 15 minute interval. Summing them would cause a 3X overestimate of the energy produced. But, if someone can prove me wrong there, then it means that SCE has screwed us over 3X worse than I am suggesting.

Data from the Tesla App is in 5 minute increments when in Day mode and I guess the SCE data that you have is in 15 minutes. If you want to compare the results on daily basis in kWh then you need to do the following:

• For each Tesla grid data point multiply by 5 minutes, then sum the total for the day and then divide by 60 mins to get total kWh for the day
• For each SCE data point multiply by 15 minutes, then sum the total for the day and then divide by 60 mins to get total kWh for the day

It isn't clear what you are looking at, so posting CSV files for your Tesla data, SCE data and a table of calculations would be valuable.

Again, the Tesla monitoring of the grid is almost surely the problem especially if you have unbacked up loads.

 

[yblaser]

Thanks for your reply. Maybe I'm missing something, but it seems odd that most days the data from SCE & that from Tesla is spot on in agreement, if my methodology is wrong. 5 minutes * 3 = 15 minutes. 15 minutes = 0.25 hours. kW * hours = kWh. Multiplication by 0.25 = division by 4. Also, because kW is measured as a sample, not as an accumulation of energy, an average of the 3 samples taken during that 15 minute interval seems like the most-logical way to estimate kWh for each 15 minute interval. Summing them would cause a 3X overestimate of the energy produced. But, if someone can prove me wrong there, then it means that SCE has screwed us over 3X worse than I am suggesting.

Sorry I'm not quite following your calculation. The easiest way is to multiply each 5 min data point in the Tesla data export by 5/60 to convert it from average kW to kWh. Then you can just some up whatever interval you want to get how many kWh you used in that period. The SCE data is already in kWh so you do not need to do any manipulation except summing the interval of interest.

I just did the above for a random day this month on my account and the discrepancy was 0.2 kWh is favor of SCE out of 12.9 kWh total net usage

 

[Antonc]

Thanks for the thoughtful responses. The assumptions underlying the "5/60" method and my method are different, but both are built on assumptions. The "5/60" method assumes that the power is constant within each 5 minute interval. My method assumes that the power varies within the range defined by each group of 5 minute intervals within each 15 minute cluster. I am skeptical about any method that relies on a single randomly selected day. Remember, that most of the time, my calculations match SCE's numbers pretty closely, so by chance, I'd also be likely to conclude that SCE was accurate, if I just randomly picked one day. I did try the "5/60" method on 3 days (side-by-side summaries start at line 258 on "4matted" tab), one being a day that differed widely from SCE & the two following, one of which is essentially in agreement with SCE and our 2 methods give practically the same results. My spreadsheet is attached for your review. Despite the fact that we all KNOW that large corporations are run by people that care only for the well-being of their customers and would >never< consider their bonuses over the good of the public and the "safeguards" of oversight commissions are absolutely perfect with no corruption >ever< clouding their judgement, I still see no evidence to discredit my calculations. Have a go at my spreadsheet and see if you can find something major that I'm doing wrong....Unfortunately, I don't see how to upload a spreadsheet with multiple tabs...so I changed the extension to ".txt" just to get it uploaded. Change it back to ".ods" and it will open... created in LibreOffice, but I think Excel will open it.
Oh and BTW, I asked SCE to come out and check the meter. I didn't figure that there was a problem with the meter, but it's a part of their standard deflection procedure. No surprise... the meter is fine. The problem... and yes, I still contend that there is one... is downstream of the meter.

 

[Antonc]

Redquill_qik"It isn't clear what you are looking at, so posting CSV files for your Tesla data, SCE data and a table of calculations would be valuable.

Again, the Tesla monitoring of the grid is almost surely the problem especially if you have unbacked up loads."
File is attached in previous post & no, I don't have any unbacked up loads... the whole house is on the batteries.