Battery Merchant Module - Optimal Dispatch

Hi,

I am testing the battery merchant module.  I've tried to test a simple case for a stand-alone project with only revenue from Energy Market (hourly price data). I've changed the price curve for the first 24 hours. However, the results I have do not seem optimal :

1. The battery does not charge when the price is the lowest... In my example, the battery should charge during hour 5 & 6 and not during the previous hours. Any idea how to configure the model parameters to get a more optimal dispatch?
2. The battery does not fully discharge when the price is the highest. In my results, it is more financially interesting to discharge more energy during hour 10 than 9. I guess this is because the model does not have the option to partially discharge during an hour when the battery power is not constrained. Am I right?

Thank you for the help!

My results and the SAM file in attachment

Hello,

The battery charges up to its maximum state of charge (SOC) and discharges to the minimum SOC defined at the top of the Battery Dispatch page.

In your file, the limits are set to 15% and 95%:



You can show the "Battery state of charge" variable in the results to see the effect of these limits:



In addition to the power price, the dispatch model considers the cost of cycling the battery in an effort to minimize the number of replacements over the life of the project defined by the analysis period on the Financial Parameters page.

For more discussion on limitations of the merchant plant, see sam.nrel.gov/forum/forum-general/3068 .

Best regards,
Paul.

Hi Paul

thank you for your answer.

however, I still do not understand why the model does not charge from the grid during hours 5 & 6 (instead of 1&2) since prices are the lowest. Do you have any idea why we have this sub optimal dispatching?

Hello,

The charging behavior is a bug. The algorithm should be charging during those low cost hours. We will fix that for the Fall 2022 version of SAM: github.com/NREL/ssc/issues/865 .
 
Changing the discharging behavior would be an enhancement to the current model. SAM's battery dispatch algorithm uses booleans (yes/no) for dispatch decisions, so cannot control the magnitude of discharges. It only controls whether or not to discharge. We will consider making that enhancement for the Fall 2022 version, depending on available resources: github.com/NREL/ssc/issues/866 .

Best regards,
Paul.