Power output constraint - EV battery

[Kristina.Haaskjold]

Hi, 

We are trying to model battery electric vehicles with different charging and discharging outputs, depending on whether the charger is slow/fast charger and the size of the battery capacity of the vehicle. 

We have tried to formulate a constraint to limit fast and slow charger power outputs (e.g., UC_FLO) based on the predefined ratio between vehicle battery power output and wall charger power output. In this case, NCAP_AFC is used to limit the vehicle battery power output, and as we understand battery modelling in TIMES, VAR_CAP would then represent the battery capacity in terms of power (e.g., GW). 

Here is an example of the constraint we would like to implement in our TIMES model: y_slow - x_bat*7/80 <= 0 and y_fast - x_bat*50/80 <=0, where y_slow and y_fast are the output power flow from slow and fast chargers and x_bat is the vehicle battery capacity. 7 and 50 kW are the respective power output of a slow and fast charger, while 80 kW is the assumed battery output capacity. 

Thanks in advance.

[Antti-L]

> NCAP_AFC is used to limit the vehicle battery power output, and as we understand battery modelling in TIMES, VAR_CAP would then represent the battery capacity in terms of power (e.g., GW).

Yes.  But what does the storage topology look like?  Do you have two charged inputs (e.g. Y_SLOW and Y_FAST) and one discharged output commodity (e.g. EVELC)?  So, assuming the discharged commodity is EVELC, have you just defined NCAP_AFC(EVELC,DAYNITE)?

> Here is an example of the constraint we would like to implement in our TIMES model: y_slow - x_bat*7/80 <= 0 and y_fast - x_bat*50/80 <=0, where y_slow and y_fast are the output power flow from slow and fast chargers and x_bat is the vehicle battery capacity.

Do you mean that Y_SLOW and Y_FAST are two input (charging) commodities to the battery storage?  If so, to me that constraint looks similar to a commodity-dependent availability factor, which could be defined with NCAP_AFC.  But maybe I am missing something?

Or perhaps you mean that Y_SLOW and Y_FAST are not commodities that appear in the battery topology, but are only output flows from the chargers, perhaps both being outputting EVELC?  If that is the case, is it possible to consider that they would be outputting Y_SLOW and Y_FAST, which would then be inputs to the battery, or does your modeling approach require or prefer the battery not having Y_SLOW and Y_FAST as inputs?

[Antti-L]

As there was no response, a small follow-up answer:

> We have tried to formulate a constraint to limit fast and slow charger power outputs (e.g., UC_FLO) based on the predefined ratio between vehicle battery power output and wall charger power output.
> Here is an example of the constraint we would like to implement in our TIMES model: y_slow - x_bat*7/80 <= 0 and y_fast - x_bat*50/80 <=0, where y_slow and y_fast are the output power flow from slow and fast chargers and x_bat is the vehicle battery capacity.

As I mentioned in my first post, knowing the topology would be helpful to understand what the issue is.  But anyway, what you described above seems very easy to do:  Just define UC_FLO for the VAR_FLO of y_slow / y_fast, and UC_CAP for the battery capacity, and employ a timeslice-dynamic UC for referring to power flows. In other words, I am not able to see any problem about defining these constraints (but maybe it was related to power flows?). However, as I suggested in my first post, one might also be able to avoid defining any UC constraints for this purpose, depending on the modeling approach.

[Kristina.Haaskjold]

Dear Antti,

Sorry for the late response. We changed the topology as you suggested to have two charging inputs and used the NCAP_AFC to define the charging power. Seems to work as attended!
However, we would also like to restrict the availability of the batteries to represent that the cars are only used for a small fraction of the time. As we model the batteries as one big battery, it is far to flexible and we therefore get too low capacity investments compared to the actual battery vehicle fleet. We tried to do this by adding NCAP_AFAC to the output commodity of the battery, however the model does not seem to respect the share as we attended.

We added NCAP_AFAC of 0.025 for the output commodity ELC_CARU. In the results, the total charging output over the year (sum(VAR_FOut)) of the battery is 2633 while the VAR_CAP = 12026, giving a fraction of 0.22. Are we misunderstanding the use of NCAP_AFAC?

[Antti-L]

Thanks for the follow-up.

> We added NCAP_AFAC of 0.025 for the output commodity ELC_CARU. In the results, the total charging output over the year (sum(VAR_FOut)) of the battery is 2633 while the VAR_CAP = 12026, giving a fraction of 0.22.

Recall that when using NCAP_AFC/AFAC, the capacity represents the nominal maximum output flow level. It seems to me that in your case the capacity is in MW and the flow is in GWh? 
If you multiply the capacity by the annual hours, you would get the nominal maximum annual discharge output:
  12026 × 8.76 = 105348 GWh.
Now, comparing that to the actual output, we get:
  2633 / 105348 = 0.025  = NCAP_AFAC.

Based on that, I don't see any misunderstanding. Can you still explain further why you think you might be misunderstanding the use of NCAP_AFAC?

[Kristina.Haaskjold]

Thanks for the fast response!
Forgot to multiply by the annual hours. Agree that the NCAP_AFAC is then working as attended.