Part Load Curve inverter + DC-connected battery: AC output exceeds the inverter

Hi all,

I’m seeing the inverter AC output go above the inverter’s maximum AC power in a PV-Battery case that uses the Inverter Part Load Curve model with a DC-connected battery. Has anyone run into this, or is it a known issue?

Setup (SAM 2025.4.16, PV-Battery, Commercial)

• Inverter: Inverter Part Load Curve, 250 kW max AC power, 14-point efficiency curve from the manufacturer’s measured data.
• Array: ~660 kWp, DC/AC ratio ~2.7.
• Battery: LFP, DC-connected, DC-to-DC conversion efficiency 100%.
• Dispatch: manual; the battery charges only from clipped energy and discharges at night.

What happens

In the year-1 hourly results:

• “Inverter AC output power” exceeds 250 kW in 1,486 hours (max 515.8 kW).
• In every one of those hours, the excess (AC output − 250 kW) equals “Inverter clipping loss AC power limit” exactly (correlation 1.0, difference 0).
• In 1,485 of those hours the battery is charging (“Electricity to battery from system DC” > 0).
• Once the battery reaches 98% SOC and stops charging, the output drops back to exactly 250 kW and clipping is applied correctly.

So the clipping is calculated, but while the battery is charging it isn’t subtracted from the AC output.Example from day 1 (Time stamp values in the attached file):

Time stampArray DCInverter ACReported clippingTo batterySOC10476 kW276.4 kW26.4 kW160.4 kW41%11559 kW292.2 kW42.2 kW221.0 kW72%13598 kW497.7 kW247.7 kW0.4 kW98%14555 kW250.0 kW218.3 kW098%

Year-1 impact: 912,515 kWh inverter AC output, of which 49,191 kWh (5.4%) is above 250 kW.

Reported clipping loss is 65,355 kWh.

What I’ve tried (PySAM 7.1.1, same case, synthetic weather)

• Part Load Curve + DC-connected battery: 1,352 hours above 250 kW.
• Part Load Curve with a flat (constant-efficiency) table: still exceeds the limit (796 hours), so it isn’t caused by my curve.
• Part Load Curve without a battery: never exceeds 250 kW.
• Inverter Datasheet or CEC coefficient model + DC-connected battery: never exceeds 250 kW.

It looks specific to Part Load Curve + DC-connected battery when the battery charges from clipped energy.

AttachmentData (exported from SAM):

• “Lifetime Hourly Data”: 219,000 rows (25 years × 8,760 h). Columns: Time stamp, Inverter AC output power, Subarray 1 and 2 DC power gross, Inverter clipping loss AC power limit, Battery state of charge, Electricity to battery from system DC, Electricity to inverter from battery DC.
• “Single Values”: AC clipping loss 6.66%, year-1 AC energy 885,300 kWh, gross DC energy 1,036,170 kWh/yr, energy yield 1,343.93 kWh/kW.

I can share the .sam file if useful.

Questions

1. Is this a known issue in 2025.4.16, and is it fixed in a later release?
2. Is there a workaround that keeps the Part Load Curve (we want the measured efficiency curve) with a DC-connected battery while respecting the inverter’s max power?
3. If not, which inverter model would you recommend for a hybrid inverter with a DC-coupled battery?

Thanks in advance.

Gio

Hi Gio,

In general, you the "Inverter CEC Database" or "Inverter Datasheet" model are the best options. You can use the "Inverter Part Load Curve" or "Inverter CEC Coefficent Generator" options if you have a reason to. 

For a DC-connected battery, with an inverter with larger power capacity than the array, "excess" AC power from the inverter does not go to the battery.

Please do attach a copy of your .sam file so I can answer your questions more specifically.

Best regards,
Paul.

Hi Paul,Thanks for the reply. I'm attaching the .sam file (DESPACHO 250).A quick clarification: in this case the inverter is smaller than the array, not larger. The inverter is 250 kW AC and the array is ~660 kWp (DC/AC ~2.7), so there is a lot of clipped energy available for the DC-connected battery.The issue is that while the battery is charging from that clipped energy, "Inverter AC output power" goes above 250 kW (up to 515.8 kW in year 1). In those hours the excess is exactly equal to "Inverter clipping loss AC power limit". Once the battery is full, the output is limited to 250 kW correctly. You can see this in the Data file I attached earlier, for example at Time stamps 10 to 14.We chose the Part Load Curve because we have the manufacturer's measured efficiency curve for this hybrid inverter and wanted to keep its shape. If that combination isn't supported with a DC-connected battery, we'll switch to the Inverter Datasheet model.Thanks again,
Giovanni

Hi Giovanni,

Thank you for sending your .sam file.

Sorry for my mistake below, and thanks for pointing it out: The problem here is that the PV array maximum DC output is higher than the inverter's maximum power limit, so the AC output of the inverter should not exceed its rated AC power. In your case, the reported AC output pwer does exceed the inverter maximum power.

It appears that there is a bug with the "Inverter Part Load" inverter option not correctly handling inverter power limiting. See  github.com/NatLabRockies/SAM/issues/2272 .

A workaround is to use either the "Inverter CEC Database" or "Inverter Datasheet" option for the inverter model.

A couple of notes:

• SAM reports the "Inverter clipping loss AC power limit" in the time series results so you can see when inverter power limiting happens.
• For the "Inverter Part Load" option, SAM does report inverter power limiting, but the inverter AC output and system power generated are not limited to the inverter's AC power rating.

Best regards,
Paul.