I have setup my SolarEdge monitoring, but the figures on PVOutput are consistently overstated. The figures on the SolarEdge monitoring site and the figures from my electricity supplier match, but PVOutput consistently has about 3-4kw too much for all my figures. Any ideas what I might be doing wrong?
Consumption or generation…
If you look carefully at the data coming back from the Solaredge API on a 3-phase Inverter you see that L1 and L2 power is always the same. That made me think and dive into the data…
It looks like:
- Total active power is plainly L1 + L2 + L3 , from the reported Active powers, Ok, but…
- The reported Reactive power from L2 and L3 are interchanged…
- The reported L3 Active power is actually the one from L2…
- You can then calculate L3 Active power from SQRT(AppwrApPwr-RePwrRePwr) (there will be rounding errors, as the powers used to calculate with are already rounded).
The error made from this is not more than 0.5% on the few days i calculated, but maybe the error can be higher for some situations or installations ?
Consumption or generation…
Well it appears everything is - import export always reads high, but consumption and generation can be high or low.
Some figures from this week
Date Source Generation consumption Export Import
25/3 portal 5.211 21.759 0.269 16.817
25/3 retailer 0.28 16.87
25/3 pvoutput 7.018 22.936 1.451 17.369
pvoutput difference +1.807 +1.177 +1.182 +0.552
26/3 portal 41.598 19.213 26.6 4.215
26/3 retailer 26.6 4.24
26/3 pvoutput 40.651 18.238 28.228 5.815
pvoutput differences -0.947 -0.975 +1.628 +1.6
Solaredge portal seems to be using totalEnergy while PVOutput is calculating energy from totalActivePower.
There will be an option in the future to use totalEnergy instead of totalActivePower.
Total energy reported from solaredge API on 26/03 was 41.700
1 Like
Excellent news…thanks
Any idea when the future might be 
No ETA as yet since when this was initially looked at the totalEnergy provided by solaredge did not match up to the totalActivePower.
Hi, total Active power reported by Solaredge is instantaneous/sampled power, so it’s not very reliable to compute Energy from. I have 1 system with 2 PVOutput systems online.
- Using modbus tcp to read power,energy and more every 1 second and computes Min, Max and Avg power of 5 mins. Max is sent as instantaneous power(v2) and lifetime energy as energy (v1, avg power on main pages of PVOutput come from energy):
https://pvoutput.org/intraday.jsp?id=69932&sid=62143&dt=20190702&gs=3&m=0 (See extended data, disable solar radiation etc to see more ) - Using Live Uploader of PVOutput from solaredge which uses Total active power:
https://pvoutput.org/intraday.jsp?id=69932&sid=64530&dt=20190702&gs=3&m=0
There are some interesting things i found also:
- Modbus reports DC and AC power, but AC power is ALWAYS 98.5% of DC power… So they might be calculating AC power from efficiency measurements in the lab and -easier to measure- DC power ?
- AVG power values taken from lifetime energy are sometimes higher than MAX power reported from modbus. So which one is right ? Energy too high or power too low ?
At the moment for a new system i’d use totalenergy as power, as with instantaneous power you end up with cloud cover measurements at the 5 min mark…
Measurements are usually 2% accuracy anyway , but having the distribution of cloud cover added to the equation is worse.
But historically, for comparison sake it may be best to leave everything as it is, which manufacturers may think also
Theoretically AC power will always be less than DC power as the efficiency of the inverter will be less than 100%
I would also expect that the efficiency of the inverter will be affected by its temperature. I.E. as temperature increases efficiency decreases. 98.5% is pretty good.
Calculating the electrical energy produced by the inverter based upon the instantaneous power measured at a fixed interval will always be fraught. As the sampling rate increases the accuracy of the actual energy output should converge on the actual value. The limitation though is how often the electronics in the inverter sample the data.
Changing to Won’t Implement as further review of sample sites still show large discrepancies between SolarEdge reported totalEnergy and energy calculated from totalActivePower.