Wildly wrong 'average power' as well as the calculated 'net' power-used

Problems
#1

As seen here: https://www.pvoutput.org/intraday.jsp?id=64598&sid=57421&dt=20180312 :

  • I’m getting spots of 0W for average power, when I’ve uploaded actual power values
  • net power-used on the next output then self-calculates into insane net power-used (73 kWh, 84 kWh, etc) on reality of about 2 kWh consumption.
  • I’m using an Eagle-200 to push “raw-xml” to PVoutput’s eaglepost.jsp as “net” values at the meter.
  • I’ve tried Wattvision importing as net, but net-power-used gets miscalculated in the same way.

This is data I’m pulling from SunPower’s monitoring API which has energy summaries at hourly granularity, and instantaneous power values every 5 minutes. I push them to PVoutput, via addstatus.jsp, first with the energy values, and then the latest power value.

I’ve also tried the live loader to push a day’s dump of energy, or power values.

Should I just not push the energy values at all to stop confusing PVoutput (as in Average Power Generation way off)?

One of the reasons I push energy is to try to correct the final day’s energy generated. For instance, the SunPower monitor says the system generated 37.2 kWh for the day, but PVoutput insists on calculating
44,199 Wh.

Another example https://www.pvoutput.org/intraday.jsp?id=64598&sid=57421&dt=20180312&gs=0&m=0. Consumption values look okay until 13:30, and then it’s off the rails again. All the 0W average power when the actual power-values are populated isn’t helping.

Thanks for help and ideas.

#2

Well, something happened because after this post, the content looks corrected, after being off for hours.

Now it’s just the previous day, https://www.pvoutput.org/intraday.jsp?id=64598&sid=57421&dt=20180312&gs=0&m=0 , that’s all over the place.

It would be nice to have an option to live edit the values on the same page, or strip only consumption/net values, etc so I don’t have to delete the entire page, and reenter upload again.

#3

Average power is calculated based on the energy provided.

  • If there is a large jump in energy from one interval to the next then there will be a spike in the power.
  • If energy does not change then average power will be 0W.
  • If energy is not sent, then it is calculated from power.
#4

Thanks, so then for 3/11,
13:20 E 13.911kWh P 3.890kW = 0kW net
13:30 E 14.358kWh P 3.410kW = 4.992kW net
13:35 E 14.479kWh P 1.950kW = 7.938kW net

Those huge jumps in net power used must be bad consumption/net values from Wattvision?

#5

If Wattvision is reporting Net data, then this will be combined with generation to determine gross consumption.

Consumption = Generation + Net
  • Positive Net = Import
  • Negative Net = Export

When wattvision reports negative net (export) greater than generation then consumption will be negative and reset to 0W by PVOutput.

You also need to ensure the timestamps are in-sync between wattvision and the device that is reporting solar, otherwise the net calculation will be off.

#6

My issue is I know my consumption is around 2 kW. I’m watching it live via RainForest Cloud, but PVoutput’s calculated consumption value is at least 3.9 times off.

The later calculated net consumptions are in the 14+ kWh, even wilder off.

I don’t see how it could have arrived at that value even with the described math, or mis-synced timestamps:
13:55: 0.5 kW (gen) + net [ -5 … 2 kW ] = 12 kW consumption ?
20:10: 0 kW (gen) + net [ -5 … 2 kW ] = 15 kW consumption ?

I do appreciate there’s no way PVoutput should be wrong, but the math isn’t there.

If there’s a column showing the raw Wattvision/Eagle “net” value (or to dump it via csv) it would help to check the calculated consumption. Thanks.

#7

If you look at that 13:55 value on the 11/3/2018
11/03/18 1:55PM 14.743kWh 1.463kWh/kW 560W 564W 0.056kW/kW 21.3C - 27.145kWh 12,531W
11/03/18 1:50PM 14.696kWh 1.458kWh/kW 630W 636W 0.063kW/kW 20.8C - 26.101kWh 11,887W

The difference in “Energy Used” was 27.145 kWh - 26.101 kWh = 1.044 kWh over 5 minutes.
If you average that energy use out over an hour you get 1.044 kWh * 60 minutes / 5 minutes = 12.528 kW which is the average power consumed over that 5 minute period.

If you’re not actually using that much power over the 5 minute period, then you will have to try and figure out why your system is reporting those figures (26.101 kWh and 27.145 kWh). The 12 kW is just a result of the dodgy “Energy Used” values.

1 Like
#8

Thanks! I pulled the raw numbers from Wattvision, and the “consumption” calculations via raw “net” values are definitely wildly off.

I’m manually, and writing some routines, to rescrub all of the power-in numbers against the raw Rainforest/Wattvision numbers myself to correct the somehow broken “consumption” calculations.

#9

Another live example:
2018-03-14 07:10 0.000kWh 0.000kWh/kW - - - 13.9C - 12.811kWh 774W
2018-03-14 07:05 0.000kWh 0.000kWh/kW - - - 13.9C - 12.746kWh 4,316W
2018-03-14 07:00 0.000kWh 0.000kWh/kW 0W 0W 0.000kW/kW 13.9C - 12.386kWh 740W

Even the 07:00 number seems a little off. This is data pushed via Eagle’s XML:RAW push, so it’s “always” net consumption.

Meanwhile Wattvision’s copy of the net-consumption in watts. Nothing weird there:
3/14/2018 6:55 592
3/14/2018 6:55 526
3/14/2018 6:55 528
3/14/2018 6:55 527
3/14/2018 6:55 530
3/14/2018 6:55 528
3/14/2018 6:55 527
3/14/2018 6:56 526
3/14/2018 6:56 597
3/14/2018 6:56 596
3/14/2018 6:56 531
3/14/2018 6:56 529
3/14/2018 6:56 527
3/14/2018 6:56 527
3/14/2018 6:56 529
3/14/2018 6:57 527
3/14/2018 6:57 527
3/14/2018 6:57 562
3/14/2018 6:57 562
3/14/2018 6:57 497
3/14/2018 6:57 496
3/14/2018 6:57 496
3/14/2018 6:58 498
3/14/2018 6:58 496
3/14/2018 6:58 595
3/14/2018 6:58 634
3/14/2018 6:58 693
3/14/2018 6:58 682
3/14/2018 6:58 724
3/14/2018 6:58 714
3/14/2018 6:59 691
3/14/2018 6:59 692
3/14/2018 6:59 639
3/14/2018 6:59 677
3/14/2018 6:59 695
3/14/2018 6:59 682
3/14/2018 6:59 776
3/14/2018 7:00 790
3/14/2018 7:00 720
3/14/2018 7:00 707
3/14/2018 7:00 720
3/14/2018 7:00 714
3/14/2018 7:00 753
3/14/2018 7:00 706
3/14/2018 7:00 745
3/14/2018 7:01 726
3/14/2018 7:01 734
3/14/2018 7:01 736
3/14/2018 7:01 740
3/14/2018 7:01 774
3/14/2018 7:01 790
3/14/2018 7:01 729
3/14/2018 7:02 730
3/14/2018 7:02 725
3/14/2018 7:02 721
3/14/2018 7:02 725
3/14/2018 7:02 725
3/14/2018 7:02 721
3/14/2018 7:02 799
3/14/2018 7:02 785
3/14/2018 7:03 763
3/14/2018 7:03 717
3/14/2018 7:03 731
3/14/2018 7:03 705
3/14/2018 7:03 712
3/14/2018 7:03 699
3/14/2018 7:03 791
3/14/2018 7:04 790
3/14/2018 7:04 797
3/14/2018 7:04 740
3/14/2018 7:04 708
3/14/2018 7:04 708
3/14/2018 7:04 782
3/14/2018 7:04 775

#10

Is this the system? - https://pvoutput.org/intraday.jsp?id=64598&sid=57421&dt=20180314

It seems the direction configured for wattvision keeps on changing between Net and Consumption on the same day.

This will generate two copies of the data from wattvision to be treated as Net and Consumption and uploaded - the result will be unpredictable.

#11

Yes, that’s the system.

I was trying to figure out which was the best way to set WattVision import. When I had left it as Net for a day (per 3/11, 3/12), it was all over the place. So I thought I’ll switch back to consumption-only, but that ends up not very useful either.

It looked like PVoutput would recalculate all the values of the day, every time I’d switched. I apologize if that’s fuzzing the system.

Meantime, I’ve also ‘auto’-rescrubbed the power-consumed values (after WattVision) with a custom app according to some formulas I’m playing with to figure out the latency between the Sunpower Monitor values and Eagle’s (via WattVision).

Thanks.

#12

It is my understanding that Wattvision was devised to be used to monitor consumption without the presence of a PV system by counting pulses generated by a smart meter or rotations of a disk for the dumb meters. Wattvision detector does not put a sign on the pulses so you cannot say whether it is consumption or net while your panels are producing power. Unless you can measure the production at the exact same period that the Wattvision is making its measurement the results will be inaccurate during the time your PV system is producing less power than you are using. You would have to pair the Wattvision measurement with the production measurement at the same intervals to determine the consumption.

#13

lwsmiser,

I believe that you are correct that Wattvision started out as you indicated, but they do also support getting data from a Rainforest Eagle device, in which case Wattvision can track net power and does indeed report net generated power as a negative value. I believe that this is what the OP is doing.

1 Like