SHAPE Parameters for TRACAR AF and ActFLO

[LuisPDias]

Hi all, we are trying unsuccessfully something in the model that can benefit similar work.

Objective: calibrate 2020 vehicle stock (CAR processes) by region. 

Context: We have included the COVID-19 impact in the p.km demand trajectory, and the model responds by reducing total vehicle stock, as the system endogenously adjusts capacity to meet the lower service demand. However, this approach does not accurately capture the real-world situation during 2020 when vehicle ownership remained largely unchanged while utilization intensity declined.
Therefore, the correct way to represent, is to also adjust the availability factor (AF) (mileage) and activity–flow (ACTFLO) (occupancy rate) parameters of the processes. These attributes are defined for each process and vintage and vary across both time and region. 

We assumed the temporal modification (to represent a short-term reduction followed by a restoration to pre-pandemic conditions) of AF and ACTFLO can be efficiently implemented using SHAPE functionality which allows defining year-specific scaling factors for attributes without re-specifying the underlying process data. 

 Other supp info:
- Model base year: 2010; time frame: 2011; 2015; 2020; 2025 and 5 yr periods till 2070; 31 regions;
- vintage is active for all CAR processes.

Problem: 
We have followed other examples in the forum for activity and emissions adapted to our case, i.e. using AF and AFX and Act_FLO and FLO_FUNCx
We want also to apply the SHAPE to all Cars, i.e. base year and new vehicles, that are implemented between 2010 and 2025.

The method we followed was:
For base year technologies (in the BY_TRANS file) and similar approach to new techs (in the trans file) but with the difference of the start year 2011):
1) Defined 31 different Shape curves representing specific AF and ACTFLO for each region.


2) then we associate the SHAPE curve index to each of the regions (figure 2) (blank values are due to absence of the process in that region). This example is for AF but we have a similar approach to ACTFLO using FLO_FUNCx.

  - VEDA2 is reading this shape definition and links to the technology




I would expect to see some activity variations and a increase of stock (via the necessity of additional stock in 2020 due to lower AF and ACTFLO) to match the demand. But this doesn’t happen.
 

Do you have some advice or see any error in the SHAPE definition or our approach is wrong.

 
Thank you in advance!

[Antti-L]

I have also been using NCAP_AFX for shaping car mileages by age, and it has been working very well for me, for many years. And likewise, NCAP_CPX appears to work well for defining "capacity survival" by age. However, "ACTFLO" (an alias for PRC_ACTFLO under VEDA) does not support FLO_FUNCX, and so that well explains that it does not work.

Of course, any shaping by age cannot really work for the existing capacity stock, because it does not have any well-defined age.  It can therefore be mainly recommended for new capacities and NCAP_PASTI capacity vintages only.  Recall also that shaping of flow parameters works for vintaged processes only.

[Antti-L]

If you happen to have a reproducible test example of some case where NCAP_AFX does not work as expected, could you please post such a test model here, and I will be happy to look at it, and find out why it does not work as expected.  Surely, if you see any such issue in some bigger production model, it should be easy for you to reproduce the issue in a small test model as well.

> Do you have some advice or see any error in the SHAPE definition or our approach is wrong.

Well, concerning seeing any error, I do see that you define NCAP_AFX for 2010 and use IE option 5, meaning that apparently there is no NCAP_AFX defined for any capacity vintages before 2010. So that might be a mistake (if you tried to define it for PRC_RESID capacities). I also see that your technology names look like those in the JRC EU model, but that model does not have any new technology named TRA_Car_Dis_Exe (it is an existing technology only), and so it seems that might also be a mistake. But this is only guessing, as I cannot really see what all you have in the model.