search

pvlib / pvlib-python

A set of documented functions for simulating the performance of photovoltaic energy systems.
https://pvlib-python.readthedocs.io
BSD 3-Clause "New" or "Revised" License
1.19k stars 998 forks source link

Night time power issue for some microinverters #140

Closed ejmiller2 closed 8 years ago

ejmiller2 commented 8 years ago

I am trying to use pvlib to compare some systems with microinverters versus string inverters. I have noticed that for many microinverters I am getting a positive AC power output when the DC power is 0. For example, using the inverter: Enphase_Energy__M250_60_2LL_S2x___ZC____NA__208V_208V__CEC_2013_

With 0V and 0W: pvsystem.snlinverter(inverters.Enphase_Energy__M250_60_2LL_S2x___ZC____NA__208V_208V__CEC_2013_, 0, 0)

I get: 1.7860471814922816

I would expect to see a negative result with a magnitude of Pnt, which is what I see for most of the other string/central inverters I have tried. This is causing issues for my comparisons since my night time values for the microinverter are adding to the total power production, while the string and central inverters are being reduced. This can add up to quite a difference when accumulating over a year.

If I try the same module and same inverter in SAM, the "Power generated by system" is showing the negative Pnt value I expect for the inverter when the DC values are 0.

The code in pvsystem.py for this situation is ac_power[ac_power < Pso] = - 1.0 * abs(Pnt)

Shouldn't the comparison be between the DC power and Pso (p_dc < Pso)?

wholmgren commented 8 years ago

I think your understanding is correct and the code should be changed. However, PVLIB Matlab's pvl_snlinverter.m specifies:

ACPower(ACPower<Ps0)= -1.*abs(Pnt); % Inverter night tare losses

so I would like to get some clarification/confirmation from @cwhanse or @jsstein. This is in at least PVLIB Matlab 1.2, 1.3 and 1.32dev (I was surprised to find that this is the current download available from PVPMC).

cwhanse commented 8 years ago

That code should be ACPower(DCPower<Ps0) rather than ACPower(ACPower<Ps0).

Thanks for the note.

Cliff

wholmgren commented 8 years ago

Thanks, Cliff.

@ejmiller2 would you like to make a pull request with the fix? Either way is fine.

Let's plan on including this in a 0.3.1 release in a week or two.

ejmiller2 commented 8 years ago

@wholmgren,

Thanks for the quick turnaround. I'm currently running pvlib that I grabbed as a conda package. Once I figure out how to run from source, I may want to do a pull request. For now I'm ok at least knowing why there's a discrepancy in my results.

ejmiller2 commented 8 years ago

@wholmgren Hopefully I have done that correctly...