PV Generation Peak Varies During Parametric Analysis

Hello,

I am doing some parametric analysis where I am varying the PV capacity (5-10-15 kWDC), battery bank capacity (0-3-6-12 kWh), and dispatch choice (day ahead, day behind, auto grid target). While the annual power generated (annual kWh) remains pretty constant for a particular PV capacity, it seems that the production for the 5 kWDC option varies quite significantly according to the battery bank capacity.

For example, the maximum kW produced (max over year) varies as 3.5, 4.0, 4.3, and 4.9 kW for the 5kWDC system across the 0-3-6-12 kWh of the battery bank capacities. This variation is not present for the 10 kW or 15 kW system, where the max kW is the same for the 10 kWDC (7.0 kW max) and 15 kWDC (11.3 kW max) system.

The annual average generation loadshape for each iteration also shows a variation across the daylight hours for the 5 kWDC system, shows a little variation for the 10 kWDC, and very minimal variation for the 15 kWDC system. There is also a some variation across the dispatch options.

Could you please clarify the source for this phenomena? I did not expect the PV production to change due to the battery and dispatch options. Attached is the model, cleaned up for your viewing.

Thank you very much for your help!

Sincerely,

Dimitry Burdjalov, EIT
Applied Energy Group, Inc.

Dear Dmitry,

Thank you for the clearly stated question and for providing a file for me to review.

I believe that the variation you are seeing is caused by the relative size of the load, battery bank capacity, and PV array capacity. The peak annual load (on the Electric Load page) is 6.8 kW, and the monthly peak varies from 5.12 kW to 6.8 kW. SAM decides how to dispatch the battery based on the dispatch option you choose, but it generally attempts to minimize the peak demand from the grid or to reduce grid power usage overall, so it will dispatch the battery in response to both power available from the PV array and the load.

It looks like the hour of peak "system power generated" is at 16:00 on June 4, where the peak load just follows the peak PV output at Noon. (You can see that on the Time Series tab of the Results page.) The battery increases the system's output to reduce that peak demand. For the smaller PV systems, the bigger the battery, the more the battery can increase the system power in that hour. For larger arrays, the array's output is sufficient to meet the load, so the battery does not need to shift the system's output to reduce that peak.

Does that explain the different peak output values you are seeing? I think if you tried increasing the load, you would see the same pattern for the larger battery capacities as you are seeing for the 5 kW array now.

Also, one thing to keep in mind: When you use the "Specify desired array size" on the System Design page, SAM does its best to get as close as possible to the array nameplate capacity that you specify, but is limited in its ability to do so by the size of individual modules and inverters. So, for example, when you model a system with a desired array size of 10 kW, the actual array size is 8 kW, and for a desired size of 15 kW, the actual size is 14 kW.

Best regards,
Paul.