Enphase recently posted a study entitled "Performance of Enphase Microinverter Systems v. PVWatts Estimates" on the Enphase website. One of the technical considerations addressed in the report is the discrepancy between the measurement of power production by the Envoy and the power that reaches the utility meter. There seem to be two aspects to the discrepancy. The first is described in the report as follows: “To increase the precision of power measurement, the meter accuracy for each microinverter was factored into the study’s calculations (see “Site Bias” calculation below). The meter accuracy is determined by comparing the microinverter’s internal power measurement to its actual power output, as tested during the manufacturing process of the microinverter.” Does this mean that each individual microinverter has a measured bias that was recorded during the manufacturing process? If so, how can we find out what the bias is for each of the microinverters in our system, as measured during the manufacturing process?
A second loss of power is described in the report as follows: “The energy losses due to resistance in system wiring could not be accurately accounted for in this study. Because Enphase Microinverters are collecting energy production information at the module, the reporting by Enphase Microinverters could be 1-3% higher than the energy production registered at the meter, which is what the PVWatts calculator is designed to predict. These losses vary depending on wire length, conductor size and other system design factors.”
My electric utility installed its production meter in my system on 5 August 2011. In the month and a half since that date, the electricity production as measured by the utility meter has been one percent lower than the electricity production reported by the Envoy as presented in the data in the Enlighten system. This one percent discrepancy presumably accounts for the average bias of my particular set of microinverters and the wiring loss between my microinverters and the utility meter.
I think it would be interesting for system owners to report in this forum a comparison of the electricity produced as measured by their utility production meter and the Envoy data presented in the Enlighten system. It would also be interesting for system owners to report a comparison of the electricity produced to the estimate of production by PVWatts.
I used a DC to AC derate factor of .87 in my PVWatts calculations, based on claims by installers in my area. I also multiplied the PVWatts estimates by monthly shading factors for my system. The electricity production, as reported in the Enlighten system, was 1 percent higher during the second half of June than predicted by PVWatts (average per day compared to average per day for June predicted by PVWatts). The production during July was 16 percent higher than predicted, and for August it was 17 percent higher. So far, production for September is about 4 percent below predicted. Compared to predictions based on the default PVWatts DC to AC derate factor of .77, the production of my system was 14 percent higher in June, 31 percent higher in July, 33 percent higher in August, and so far 8 percent higher in September than PVWatts estimates using the default DC to AC derate factor and including monthly shading factors. (The predicted shading during these months is small, 1 percent in June-August and 2 percent in September.)
Great idea! We really love this stuff and hope to introduce more features and products to help people understand their solar systems. Have you already uploaded your information into our PVWatts integration?
If not, here's more information about the integration:
For those wondering about the study, here's a link:
Regarding the "unit bias" data, we don't currently have a way to upload this data into Enlighten, but it's a great idea and something worth looking into.
Thanks for providing the link to the Enphase study, "Performance of Enphase Microinverter Systems v. PVWatts Estimates."
Yes, I have uploaded my system array details in Enphase's PVWatts integration window. The resulting PVWatts estimates for my system are low by about one percent because the implementation of the PVWatts integration window rounds the array size to one decimal. I think the Enphase programmers should reprogram this window to keep and use more decimals in the array size or to express the array size in Watts instead of kWatts to fix this problem. (If the size of someone's array rounded up instead of down, the estimates would be correspondingly high. This is only a meaningful problem for small arrays, but still I think this problem should be fixed so that the results will be accurate and comparable for all array sizes. For example, if the array size of a system is 3.349 kW, the PVWatts integration window will round the array size to 3.3 kW. The resulting PVWatts estimates will be low by about 1.5 percent since they will be calculated for an array size of 3.300 kW instead of an array size of 3.349 kW.)
In addition, the PVWatts integration window does not allow one to input monthly shading estimates. So, to get accurate PVWatts estimates, I temporarily put in an array size 1000 times the size of my array. I take the resulting monthly energy generation estimates and divide each estimate by 1000. Then I multiply these monthly estimates by the estimated shading factors for each month.
According to the data in Enlighten, one of my inverter-module combinations always (every day) produces more electricity than the others. The daily standard deviation among the inverter-module combinations is about 1-2 percent. The most productive inverter-module combination produces about 3 percent more electricity than average and the least productive about 1-2 percent less than average. This is with no known variation in shading of the modules. I have tried to correlate production with temperature as I noticed that the modules in the bottom row of my array are on average slightly more productive than those in the upper row, presumably because they and their inverters are cooler. One of the inverter-module combinations in the lower row seems to be an exception to this pattern. Now I wonder if part of the production variation is explained by the biases of the internal power measurement of the inverters. That's why I’d like to know the biases for my particular inverters if estimates of these biases were calculated during the inverter manufacturing process. I don’t need to have this inverter unit bias data uploaded into Enlighten, I’d just like to have the data for my own interest.
Sorry to say ... it seems the study was flawed. To truly "test" the inverters you would have to meter the actual production. It's my understanding that some installations have another meter that reads the actual "production" closer to the electrical panel....those sort of installations should be looked at and compared with standard inverter installations that use a similar metering set up.
Yes, I hope that other Enphase owners who have a utility meter will report here the difference between the Enlighten measurement of electricity production and the utility meter measurement of electricity production for their systems.
The validity of the Enphase study could be improved by adjusting the results to reflect the wiring-related losses that typically occur between the inverters and the interface to the utility grid. The Enphase report says that these losses are typically 1-3 percent and depend on factors such as the length and gauge of the wires used for the connection between the utility meter (and service panel) and the inverters.
Conductor (wire) size is measured by gauge. A smaller gauge means a larger diameter wire. (For example, 10-gauge wire has a larger diameter than 12-gauge wire.) Using a smaller gauge (that is a larger diameter) wire results in less loss of power between the inverters and the utility grid (and service panel). The loss also depends on the length of the wires which depends on the geometry of the installation (how far away the PV modules and inverters are from the connection to the utility grid and service panel). Longer wires result in a larger power loss.
Smaller gauge (that is larger diameter) wire costs more. Electrical code may allow a larger gauge (that is smaller diameter) wire than Enphase suggests for keeping the loss of power in the 1-3 percent range. This is something that owners and installers should think about in planning the installation. I think that the extra initial cost of the larger size (that is smaller gauge) wire is relatively small compared to the cost of the other system components and the work of installing the system.