Question on emission factors

[seungwoo.kang]

Dear colleagues,
 
I am seungwoo KANG of CMA (Centre of applied mathematics) in MINES Paristech, France. 
 
During using the TIAM model, I found out an unexpected comportment on emission factors.
 
The emission factors of aggregated energy sources (INDCOA, INDETH, etc) remain constant in time while their compositions change.
 
For example, INDCOA's CO2 emission factors is about 98.30 tCO2 based on reference year's share of COAHCO and COABCO for each region. And these emission factors are declared by "~COMEMI" table in IND_Emi sheet of VT_IND files.
 
Then, whatever the scenarios are imposed, the share of COAHCO and COABCO changes. Nevertheless, INDCOA's emission factors are always fixed at initial declaration (for CO2, 98.3 tCO2)
 
This problem becomes more serious in transport biofuel. TIAM model has fueltech_diesel_(TRA)_new (TRADST005), which allows biodiesel use in TRADST up to 100% by 2030. For the reference, this fuel tech process is different with biodiesel blending (TRABDL). Even in this case, the emission factors of TRADST do not change according to the actual share of BIODST in TRADST.
 
If this feature is intentionally implied in TIAM model, I'd like to get some more explications. If not, can anyone give some ideas how to implicitly update the emission factors complying with the chosen shares by model itself? Or did I miss something?
 
Thank you,
 
Seungwoo KANG
 

[Antti-L]

I have not worked much with TIAM lately, but I just want to emphasize that both TIMES and VEDA support defining emission factors on any process flows (by using FLO_EMIS), and therefore it is easy to model the emissions according to the actual fuel shares. You could just remove those offending ~COMEMI emission factors and replace them with emission factors defined on the input fuels of the fuel technologies.

[danial]

I have not worked much with TIAM lately, but I just want to emphasize that both TIMES and VEDA support defining emission factors on any process flows (by using FLO_EMIS), and therefore it is easy to model the emissions according to the actual fuel shares. You could just remove those offending ~COMEMI emission factors and replace them with emission factors defined on the input fuels of the fuel technologies.

Dear Antii-L, 
As mentioned by Kang, this problem can miscalculate the total CO2 emission. We need the COMEMI table to be time dependant. I used FLO_EMIS as you suggested, but the problem stays the same since FLO_EMI is assigning emission to a process, not to a commodity. For instance, a vehicle that bought in earlier years, when the fuel was blended with less than 100% biofuel, will have still positive emissions in the future. However, we change the composition of the fuel (biofuel) to 100%, so it should not be positive even for the same vehicle.

[Antti-L]

Sorry, I am not sure I am able to follow. Which problem are you talking about?  Kang referred to a modelling problem, where emission factors were defined for blended fuels by COMEMI, regardless of the blending shares upstream. That problem is easy to solve by using emission factors on those upstream fuel technologies.

With FLO_EMIS you can do everything what you can do with COMEMI, but also much more. COMEMI is actually converted by VEDA to the TIMES attribute VDA_EMCB, which defines emission factors on process input flows. So, FLO_EMIS can do everything COMEMI can do, and it is also year-dependent.

If you mean that emission factors of vintaged processes should be possible to define also on a non-vintaged basis, that would be a different, completely unrelated issue. Both FLO_EMIS and VDA_EMCB (and so COMEMI as well) result in vintaged emission factors for vintaged processes. But fuel technologies are in most cases non-vintaged and do not have that issue.

[demar]

I tried your approach.... 

And now the model does not recognize the Sector Fuel Technology Existing (FTE) that has input shares. Any thoughts on this?

/M

[Antti-L]

Can you elaborate what do you mean by "does not recognize"?

[demar]

Can you elaborate what do you mean by "does not recognize"?

I was missing to define the process that created the "inputs".- Now it works... Yet, I tried to build the scenario wiht a TFM_INS table as suggested by you... but it seems my model is not really recognizing it... As the results are the same for the base run (with no scneario) and the run with scenario. Am I doing something wrong?

/M

[Antti-L]

You need to specify the Attribute (Share-I), Year (e.g. 2030) and PSet_PN (e.g. FTE-TRADSBL), as well as the value, on every row in that TFM_INS table.

[demar]

It worked!! Thank you so much.

It would be good if TIMES sometimes could give us better insight on certain kind of "errors". It would be nice if it said that it saw numbers which it did not use.

Thank you again,
/M

[AKanudia]

Did you not see a sync log reporting some numbers that did not generate any records?