Standalone Battery Peak Shaving Dispatch

I have created a model to understand the benefits of installing a standalone battery system at a location with existing solar PV. I modeled the PV generation by uploading a net load profile (electric load - solar PV generation) in order to understand just the economic benefits of the battery system.

The rate structure has a ratcheting demand component that I think might be affecting the battery dispatch. I have looked at the data tables and the month of July seems to be setting and annual billing demand (attached screenshot). I have look into the time series data to try and understand why the billing demand can not be lowered any more by battery dispatch. However, I can not find information specifically about the battery dispatch (I get lumped values of both the "net load" and battery dispatch) and I can't find anything on state of charge. Can you please direct me to where I can find this time series information and potentially explain how the peak shaving battery dispatch considers ratcheting demand charges? Thank you!

Hi James,

Another approach to this analysis would be to use the Generic System - Battery configuration, and use your PV generation profile data as input on the Power Plant page (the "Import hourly or subhourly generation profile from file" option.)

To model a ratcheting demand, expand the Billing Demand panel on the Electricity Rates page and check Enable billing demand lookback. See the Electricity Rates Help topic for more details on those inputs.

Please try that and let me know if the results better match your expectations.

Best regards,
Paul.

Hi Paul,

That is how I originally modeled the system (Generic System - Battery configuration) but the annual value of the electricity savings displayed in the cash flow output seemed to be coming from both the generic power plant and the new battery system. This is why I used the other modeling approach because I want to know the potential economic savings from only the battery system.

I feel confident that I originally applied the ratcheting demand correctly. However, I did looked into the additional documentation about the Peak shaving battery dispatch algorithm and it does not seem to consider ratcheting demand or overall monthly demand, only a 24 hour demand. I am still trying to find time series data on battery SOC/capacity to better understand the dispatch but I only see the attached battery outputs available in the 15 Minute Data time series results section. Can I find more information about year 1 hourly battery operations somewhere else?

I tried to use the Price signal forecast but the software keeps terminating when I try to simulate. Do you know what could be causing this problem?

Thank you!

Hi James,

The "Battery state of charge (%)" variable is under "Lifetime 15 Minute Data" in the time series results. The lifetime results span the entire analysis period, but you can just focus on the Year 1 values.

You are correct that the peak shaving battery dispatch algorithm dispatches the battery once over the look-ahead period in its attempt to find the monthly peak.

Thank you for letting us know about the issue with the price signal dispatch option. We will investigate that, and I will follow up when I know more.

Best regards,
Paul.

Hi again,

We will fix the bug with the price signal forecast dispatch option for the next update to SAM 2021.12.02: github.com/NREL/ssc/issues/561

And, to clarify my response above, the peak shaving dispatch option does consider monthly peaks. The algorithm is described in DiOrio, N. (2017). An Overview of the Automated Dispatch Controller Algorithms in SAM. NREL/TP-6A20-68614. ( PDF 770 KB ). Specifically, see the last paragraph of Section 3.2 on p. 7. You can find other battery-related publications on the SAM website at sam.nrel.gov/battery-storage/battery-publications.html .

The peak shaving algorithm is not aware of the electricity bill calculation, so it is not possible for it to respond to ratcheting demand rates. The battery responds to the available power from a PV or other system if there is one, and to the load.

Best regards,
Paul.

Based on how the peak shaving dispatch algorithm is set to reduce demand charges, is it possible for scenarios with different Desired bank power values at the same Desired bank capacity (4 hrs for example) to yield different battery dispatch profiles (seen in "Data tables" from a completed simulation) for the same electric load hourly load profile? Is it possible for the dispatch to do any of the following:

• Shift the load peak to a different time of day
• Add to the existing daily peak load by charging, but discharge for most of the other parts of the day
• Reduce existing daily peak load by discharging, but charge during most other parts of the day
• Moderately charge/discharge for longer periods of time in the day
• Aggressively charge/discharge for shorter periods of time in the day

 Could it in fact be a better grid-operational decision for the peak-shaving algorithm to add to peak to reduce load in other parts of the day to effectively reduce daily demand charges but while making sure a daily peak is not reduced below the current monthly peak load?