Discrepancy w/ provided vs. derived irradiance

We're using a SAM script (thanks for making that available!) to get hourly distributions for the TMY2 locations at 40 azimuth/tilt combinations. Using the hourly data, we're deriving monthly distributions and total annual kWh/m2/day.

To derive kWh/m2/day for each month, I divide the total production estimate for that month by the system's AC Watts times the days in that month. The derived irradiance data is about 9% off from provided irradiance data for my test location (CA Sacramento).

FYI, I get AC Watts with: STC Watts * (1 - inverter derate) * (1 - total user losses)

What am I doing wrong? Thanks!

Hi Michael,

I'm not sure I completely understand what you are trying to do. Are you trying to calculate the average daily solar energy from the system's electrical output?

Dividing the total kWh electricity production by the AC power rating in Watts would give you a kWh/W metric rather than a measure of the available solar energy.

Can you share a snippet of code so that I can see exactly what you are trying to do?

You can email it to me at sam dot support at nrel dot gov if you prefer not to post it here.

Best regards,
Paul.

Thanks Paul, I'm just trying to reverse calculate to get kWh/m2/day. I want to make sure I understand the SAM model's output.

I assume that if:
AC System Size * Avg. Daily Irradiance * Days in Month = Production in Month

Algebra requires that:
Avg. Daily Irradiance = Production in Month / (AC System Size * Days in Month)


Here are my SAM inputs:
Weather: screencast.com/t/MNcyG26wo
System Design: screencast.com/t/SiuqO2HUVaeF

Here are my applicable SAM ouputs:
screencast.com/t/KBlHj7KZFkR

My logic above doesn't quite get me 1.869kWh/m2/day. I'd like to understand what I'm missing.

Thanks,
Michael

Hi Michael,

You might want to refer to the PVWatts technical reference manual to see the equations it uses. It is available for download under "PVWatts" on the Performance Documentation page of this website:

sam.nrel.gov/performance

As you know, PVWatts and the other performance models in SAM perform hour-by-hour calculations (or subhourly calculations), so your monthly calculation will be an approximation, and you should expect some difference between monthly totals of SAM results and your estimates.

The basic algorithm for PVWatts for each time step looks like this:

1. Calculate the incident irradiance (W/m2) for the hour (see reference manual for details)

2. Calculate irradiance transmitted (W/m2) to cell to account for the effect of the module cover on irradiance (see manual)

2. Calculate the array's output: Pdc_gross = Pdc_nameplate (W) * I_transmitted (W/m2) / 1000 (W/m2)

3. Adjust gross array output to account for the effect of cell temperature (see manual)

4. Apply losses: Pdc_net (W) = Pdc_gross (W) * ( 1 - dc_loss (%) / 100%) )

5. Calculate inverter efficiency, eta_inverter (see manual)

6. Calculate inverter output: Pac (W) = Pdc_net (W) * eta_inverter

7. Limit inverter output to minimum of zero and maximum of inverter rated AC power (Pac_nameplate = Pdc_nameplate / dc_to_ac_ratio)

Best regards,
Paul.

Paul, thanks for the write up and reference.

...I look forward to getting my head around them both!