Elastic electricity prices in the IMPEXP region

[Pernille.S]

I would like to model elastic electricity prices in the IMPEXP regions since constant electricity prices in neighbour countries, that are independent the traded volume, is a very coarse assumption. From the documentation I interpret that this is an included option to TIMES from the note on Elastic supply cost curves in TIMES from 2010. But I am not sure how this feature works……….?

I would like to model the electricity price of the import and export to depend on the quantities that are traded. Is this possible to model with the ELAS parameter? If yes, does it exist a test model where this is illustrated? Have anybody done this? Do I need to base the reference prices on a reference run or can I set the reference prices and reference quantity outside the model?
All help is highly appreciated!

[Antti-L]

I would like to model the electricity price of the import and export to depend on the quantities that are traded. Is this possible to model with the ELAS parameter? If yes, does it exist a test model where this is illustrated? Have anybody done this? Do I need to base the reference prices on a reference run or can I set the reference prices and reference quantity outside the model?

 There is no ELAS parameter in TIMES, but you can use the damage cost facility to model elastic supply cost curves, as described in the document http://www.iea-etsap.org/docs/Elastic-Supply-Curves.pdf. The support for elastic supply cost curves was added into TIMES on Denise's request in 2010, and should be considered an experimental feature.

If you don't want to use the automatic transfer of reference prices and quantities from a Baseline run, you can also define the elastic supply costs manually, as described in the document. You only need to define the costs to be applied to the commodity production, by specifying DAM_ELAST(r,c,'N') to be either EPS or -1, depending whether you want the additional costs to be shifted of not. Otherwise the manual specification would be similar to normal damage cost specification.

A notable limitation is that you cannot define timeslice-specific elastic cost curves.  There is no test model publicly available, but the document contains a worked example (based on the automatic transfer of Base prices and quantities).

[Antti-L]

In your case, as you don't seem to want to make use of the automatic transfer of reference prices and quantities from a Baseline run, it might be more transparent to simply use several import/export processes/commodities for defining several cost steps for the exogenous imports/exports.

This is the usual way of defining a supply-cost curve in TIMES. You could model several import processes with different prices/costs, and bounds for the volumes in each step.  Alternatively, because the imports are exogenous, you could also use a single process with several commodities to define the cost steps, thereby retaining a common capacity for all of the steps. Then you would still need to aggregate the different cost steps.

[Pernille.S]

Dear Antti

Thank you for the quick reply! I will test this feature out.

I have some questions for clarification.
-     You mention that I cannot use time elasticity on a time slice level. I assume I still can use prices in the IMPEXP region that are time slice dependent. Correct?
-     For electricity export I wish the price to decrease with quantity. How do I do this? From the damage documentation on damage function it is specified that DAM_ELAS cannot be negative….. [0 , INF].

[Antti-L]

Ok, so you will be testing the elastic supply cost curve feature. Let me know how it goes.

Concerning timeslice-specific import/export prices, you can of course define such by using IRE_PRICE or FLO_COST/FLO_DELIV.  But with the elastic supply cost curve feature you cannot define elastic cost curves separately for each timeslice.  As it has been implemented now, the elastic cost curve is always based on the aggregate annual amount, with the base price at the base quantity, and elastic prices in both directions (see the formula on page 3 of the doc).  The implementation could, of course, be generalized to allow also timeslice-specific curves to be defined.

Note also that the elastic supply cost curve is always applied to the total production of the commodity in question.  Therefore, if you want it to be applied only to the amount imported or exported, you need to have a dedicated commodity assigned to the imported electricity and exported electricity.

Concerning electricity exports, for which you wish the price to decrease with quantity, that will automatically be the case, because the export price is a revenue (a negative cost) and the elastic cost will be a positive cost increasing with quantity. When the elastic cost is added to the export price, the sum will thus be a marginal revenue which is decreasing with quantity.

[Pernille.S]

Dear Antti

We have finally now some time to test this modeling approach and we have another question;

If we use several trading processes with different prices/costs, and bounds for the volumes in each step, how can we prevent that both export and import occurs in the same time slice? With using several trading technologies the model can export at a high price and import at a low price in the same time slice. Can this be solved with a user constraint? We are not interested in using integer model decisions.

Arne & Pernille

[Antti-L]

Tough question.

You can of course prevent both export and import occurring in the same time slice, by exogenously assuming for each timeslice, that only export/import can occur in that timeslice. But if you want the model to decide by itself the sign of the net imports in each timeslice, I cannot see how you can prevent the model from exporting and importing in the same time, if that happens to be profitable. If you find a way, I'd be interested to learn how.