search

GMLC-1-4-2 / battery_interface

Implemenation of Device Models and their Battery Equivalent Interface
MIT License
5 stars 0 forks source link

Energy Market Service Analysis #121

Closed afernandezcanosa closed 5 years ago

afernandezcanosa commented 5 years ago

@afernandezcanosa (cc @Hayden-Reeve, @emayhorn, @DavidWiniarski-pnnl, @TomEdmunds, @can7huang ) Energy Market / EV

baseline - RefMonteCarlo = False sim_step = 5 min subfleets = 100 (Clearly, this needs to be changed - I'll run a more granular simulation to avoid strange responses with None requests)

caiso_prices

SimResults_EnergyMarket_ElectricVehiclesFleet_20190521T1338

Service metrics There are no service metrics at this time. Should I try to get service efficacy coded?

It seems to provide a reasonable response although the fleet is not able to charge more than its baseline, so the negative request is disregarded. However, there is some charging curtailment where the prices are the highest at around 8 PM.

Any feedback would be very appreciated. Thanks!

(Edited plot) SimResults_EnergyMarket_ElectricVehiclesFleet_20190522T1124

afernandezcanosa commented 5 years ago

If these results seem acceptable, we should think about integrating the service with the other two "eligible" devices (Electrolyzers - @rkadavil and Water Heaters - @jmaguire1).

Hayden-Reeve commented 5 years ago

@afernandezcanosa , These results look good to me and the EV response looks consistent with what it has provided for other services.

afernandezcanosa commented 5 years ago

Update of the results:

Energy Market / EV

baseline - RefMonteCarlo = False sim_step = 5 min subfleets = 100

Results: SimResults_EnergyMarket_ElectricVehiclesFleet_20190604T1445

Impact metrics: ImpactMetrics_EnergyMarketService_ElectricVehicles_20190604T1445.xlsx

afernandezcanosa commented 5 years ago

RESULTS

@rkadavil @Hayden-Reeve @emayhorn @DavidWiniarski-pnnl These are the results that I have obtained for the two fleets that are currently available to provide energy market service. Both cases seem to respond as expected, reducing load when the prices are higher and increasing load (in the case of the electrolyzers) when the prices are lower. @rkadavil Let me know if the results make sense to you.

Two additional comments:

  1. Should I code at least the energy impact metric in this service? I can just copy-paste the code that you have for the peak management service, @emayhorn.
  2. @rkadavil Your PR #123 does export the impact metrics in the same folder where the fleet code is located and it does not follow the naming convention. This is not critical at this time, but we should consider making this process of exporting the results standard for all the runs.

Energy Market / Electric Vehicle

baseline - RefMonteCarlo = False sim_step = 5 min subfleets = 100

Results: SimResults_EnergyMarket_ElectricVehiclesFleet_20190604T1445

Impact metrics: ImpactMetrics_EnergyMarketService_ElectricVehicles_20190604T1445.xlsx

Energy Market / Electrolyzer

sim_step = 5 min number_electrolyzers = 50

Results:

SimResults_EnergyMarket_ElectrolyzerFleet_20190606T1114

Impact metrics:

EnergyMarketService.xlsx

emayhorn commented 5 years ago
  1. service

@afernandezcanosa Have you downloaded the latest master code? The energy impact metric should be being computed by the devices and results are output to the metrics file.

See the following code in the test.py file. You should be able to copy this line for energy market service. https://github.com/GMLC-1-4-2/battery_interface/blob/63636ecae30cb061f8dc8f9df2174e13d4a416db/src/test.py#L178

Hayden-Reeve commented 5 years ago

@afernandezcanosa ,

The results make sense to me.