[Question] Straggling Data Calculation

[syaoran68]

Didn't really know how to actually title this properly, but I'm wondering if people had issues regarding inaccurate totals that are happening from night time.

I will see about an average of like 7 hours worth of 0.01 kwh of production. Its not as noticable when you take the entire day into account, but as you start the day off its super noticeable.

I'll see about 0.07 to 0.09 kwh of solar production from a given night. and doing even generic math it'll throw off the monthly totals by a bit.

The only reason why i started looking into it was the amount reported by my electrical company being received and what sunpower reportedly sends its off by quite a bit. Not sure why for that either. Then i noticed the tiny production in the middle of the night.

Thanks!

-Eddie

[krbaker]

I'm not entirely sure which values you are looking at but a few thoughts:

1. The inverter data for their power output may not go all the way to zero (I think they actually turn off with zero sun so the last data might be in a very small # of watts, if you integrate that overnight it would cause funny numbers). Whenever you can use the actual sum kwh from the inverters (if you have to use the inverters)
2. The inverters are not revenue grade meters, I'd expect up to 5% error or something
3. If you have production CTs and get the revenue grade ones I think you should be < 1% error
4. the consumption CTs are just 'for fun' and are not revenue grade ever on the sunpower system
5. if you are doing your own integration we are sampling quite infrequently so I would expect a decent amount of error.

Depending on where the production / consumption CTs are located you may also be doing subtraction to get one of those values (if you have a line side consumption meter it has to subtract out the solar production to figure out your home production).

[syaoran68]

for the solar production data input on HA's energy chart i'm using the sensor.sunpower_production_lifetime_power for solar production. i do understand that there will be some 1% error. i'm just wondering what would be the best way to remove that error. as i'm afraid that it'll skew the data a bit.

i mean if you think about it.. HA will errantly think theres 0.01 kwh coming in over maybe 5-7 hours a night.. so that'll lead to a minimum of 0.05kwh a day of extra power its mis-totaling. thats an additional 1.55kWh of power..

i appreciate the response!

[krbaker]

Okay, you are using a lifetime_power sensor, those are a running total directly on the monitoring system so nothing wrong that the integration could be doing.

You could look at how your CT is installed maybe its wires are run in parallel with something else. Mine for example actually runs backwards overnight because the monitoring system itself is powered from the combiner panel and the CT measures the lines out of the combiner.

[edhuang04]

I think i'm seeing the reverse of yours.. looks like throughout the night it does climb..

I'm going to try to swap to individual inverter lifetime power since it looks like the inverters don't report any usage after sundown.

looks like theres some tiny amount of production at night right now at 10:00pm 😂