calculating fuel costs using cost_flo and/or eq_combalm

[iris]

Hello, 
I have a model in which fuels enter sectors through a distribution process. The distribution process is a throughput hub with no leakage. It transforms a shared commodity (say "ELC") to a sector commodity (say "INDELC") and so is used to enable tracking of the commodities in a given sector.

I am trying to calculate the total costs for these distribution processes in each sector as the sum 

• Fixed operating + maintenance costs
  


• Investment costs
  


• Fuel costs
  

To calculate fuel costs, I want to take the price per unit commodity * var_fout (since the distribution processes in each sector acts as direct throughput hubs and var_fin = var_fout)


The approach I have taken is as follows:

• For processes defined on annual timeslices  which have a cost_flo defined: cost = var_fout * cost_flo 
  


• For processes defined on non-annual timeslices, which do not have a cost_flo defined: cost = var_fout * eq_combalm (for the specified var_fout commodity)
  

My questions are

1. Does using the cost variables in this way make sense to calculate fuel costs? 
   


2. Is the differential treatment in fuel cost calculations of annual and non-annual processes appropriate?
   


3. Since EQ_combalm is zero when the commodity is non-marginal, the treatment for non-annual commodities will have zero price at certain times. Does this make sense?
   


4. How is cost_flo calculated? Does it incorporate marginal costs of the commodities? (Apologies, I couldn't find this information in the documentation)
   


5. Is there any double counting in summing the fuel costs calculated in this way with the investment + FOM costs?
   


6. Are there additional terms that need to be included in the calculations?
   

I have read through some other relevant posts, but am struggling to make sense of the answers in relation to my specific problem.
EQ_COMBAL.M (Commodity Balance- Marginal) (iea-etsap.org)
Fuel costs (iea-etsap.org)
Determination of Transmission and Distribution Charges (iea-etsap.org)
EQ_COMBALM for ELCHIG (kanors-emr.org)[/font][/size]

Many thanks
Iris

[Antti-L]

Thanks for your question that may be of help to others, too.

However, I don't quite understand why you say that you would calculate cost = var_fout * cost_flo, because the VBE attribute Cost_Flo already gives the flow costs (it is NOT a unit cost but has been already calculated by multiplying with the flow levels). More generally, none of the annual cost reporting attributes (Cost_*) are unit costs, but they report total annual costs for the process and/or commodity in question, by period (except for salvage, which is a special case and can be ignored when looking at annual costs).

Therefore, to calculate the total annualized costs for the distribution process p, with input c_in and output c_out, I would suggest the following:
 TotCost(p) = Cost_Inv(p) + Cost_Fom(p) + Cost_Act(p) +
    Sum(ts, Var_Fin(p,c_in,ts) × Eq_Combalm(c_in,ts)) + Cost_Flo(p,c_in) + Cost_Flo(p,c_out)

Now to your questions:

> Does using the cost variables in this way make sense to calculate fuel costs?
No, in my view multiplying var_fout with cost_flo would not make sense.

> Is the differential treatment in fuel cost calculations of annual and non-annual processes appropriate?
See my suggestion above, with no such difference.

> Since EQ_combalm is zero when the commodity is non-marginal, the treatment for non-annual commodities will have zero price at certain times. Does this make sense?
Zero price would imply overproduction and should thus occur rarely (e.g. excess variable renewable production in some timeslices). But yes, it makes sense in such cases.

> How is cost_flo calculated? Does it incorporate marginal costs of the commodities? (Apologies, I couldn't find this information in the documentation)
Cost_flo is calculated by multiplying the user-defined flow costs (FLO_COST, FLO_DELIV, and also IRE_PRICE for exogenous trade) by the flow levels, to obtain the total annual flow-related costs for the process in question.

> Is there any double counting in summing the fuel costs calculated in this way with the investment + FOM costs?
No, if you use my suggestion above. But the commodity prices include also the producer's surplus.

> Are there additional terms that need to be included in the calculations?
See my suggestion above.

[iris]

Dear Antti-L
Thank you so much for the clear and comprehensive reply. I should be able to implement this.

From what I can see, am I correct in thinking that your formula for TotCost(p) is a general formula for calculation of operational costs of any process, not only distribution processes?

If I am interested in the total  operational costs of all processes (p) in a sector (ie. the distribution processes as well as downstream processes) would summing over all processes work?

TotCost_sector = sum(p, TotCost(p)) ..... (1)

Or does TotCost(p) for  processes downstream of the sector distribution processes include costs propagated from TotCost for the distribution processes? (i.e. would (1) lead to double counting of costs)

[Antti-L]

The TotCost(p) formula I suggested above is valid when you want to calculate the total annual costs of any individual process (you would just need to add some more terms if you have more inputs), when you want to include the prices of the input fuels in the costs of that process, which is what I understood you wanted.  It is also valid for calculating the total annual costs of several non-chained processes, such as all distribution processes at the boundary of a certain sector.  But if you would sum these costs for multiple processes over the energy chain, such that the processes downstream have inputs from processes upstream, you would obviously be counting the same costs multiple times, because in general, the commodity prices (commodity balance marginals) cover the accumulated costs over the whole production/distribution chain up to that point.

Concerning total system costs, see e.g. the following thread: https://forum.kanors-emr.org/showthread.php?tid=578

There I explain the following:
Reg_obj / Reg_wobj gives the total present value of all costs by region, in the discounting base year (G_DYEAR).
Reg_ACost gives you the sum of the annualized costs by region and cost type.
Time_NPV gives the present value of the time (in years) in each period, in the discounting base year (G_DYEAR).

The annualized costs are by default reported for each Milestone year, and therefore you cannot reconstruct Reg_obj / Reg_wobj from Reg_ACost. However, if you use the following switches, you should, indeed, be able to reconstruct Reg_obj from Reg_ACost with high accuracy, by multiplying the sum of Reg_ACost with Time_NPV for each period, and summing together. (Exclude the IRE component if you want to verify individual regional totals).

 $SET OBLONG YES
 $SET ANNCOST LEV

As explained above, you can get the total annual costs of the energy system by region, by summing up all the cost components of Reg_Acost.  And, if you include any tax and subsidy components, you get the total annual system costs+taxes−subsidies, fully consistent with the objective function.

In the same way, you can also calculate total sectoral annual costs, by using the more detailed Cost_* attributes instead of the aggregate Reg_ACost, and assigning each process to a unique sector.  You should easily be able to see that, for example, Reg_Acost(r,t,INV) is the sum of Cost_Inv(r,v,t,p) over all processes, and so on. This is the only method I can suggest for calculating total sectoral annual costs that is guaranteed to be consistent with the total energy system costs. You are of course free to develop other methods yourself as you like, and you can always check the consistency of your totals against the total regional system costs described above (using just Reg_Acost or the Cost_* attributes excluding Cost_Salv.).

---
P.S.  You can also use the Val_Flo attribute if using the TotCost(p) formula for calculating the total costs of individual processes.  It would give you directly the values Var_Fin(p,c_in,ts) × Eq_Combalm(c_in,ts). You can have those values reported by setting RPT_OPT('FLO','3')=1.

[iris]

Thanks Antti-L. That is all clear. I need to provide total sector costs, as well as distribution process costs by sector, so I will use both approaches that you suggest.