# IEA-ETSAP Forum

Full Version: Curtailment costs
You're currently viewing a stripped down version of our content. View the full version with proper formatting.
Dear Antti, Is there a way/best practice to model Curtailment? to calculate the amount of energy "not sold", "not fed into the grid" (per technology, i.e. Solar, Wind power plants,..), and translate this into costs? (per instance, as a variable cost per each kWh that is "not consumed"). My aim is to reflect in the model the cost implications of curtailment per technology. Best regards, Diana.
Interesting question, but I have myself no deeper insights into the curtailment issues. As far as I can see, in TIMES curtailment basically occurs only when the price of electricity falls to zero. The marginal of the commodity balance is zero when consumption + storage charge (=the energy sold) is falling below production + storage discharge (=supply). Thus, the value of the lost generation due to curtailment would in general be zero. Negative prices can only occur in TIMES if you have modelled the electricity balances with a fixed equation type (FX), and if the curtailment of variable generation is being limited (for example, by using fixed or lower-bounded AFs).  However, I think that most variable generation is already today being implemented with active control systems allowing optimized curtailments, and so I would think negative prices should probably no longer occur in the future, unless I am missing something? Currently, apart from the residual load facility (see https://iea-etsap.org/docs/TIMES-RLDC-Documentation.pdf), there are no other features in TIMES dedicated specifically to the modelling of curtailments. Therefore, the modeller is mostly left with the generic features of TIMES (including the dispatching features), with respect to curtailment modeling. For taking into account the operational constraints of thermal plants, you can use the TIMES dispatching features (minimum stable operation levels, start-up / shut-down costs, minimum on-line / off-line times, ramping constraints, and partial load efficiencies), for better modelling of the system under potential curtailment situations. If you can formulate what you want in mathematical terms, I see no reason why it could not be modelled in TIMES. If you have your (linearized) mathematical formulation ready, but don't know how to implement it in TIMES, I am sure I can try to help with it.
(20-05-2019, 06:09 AM)Dianaprz7 Wrote: [ -> ]Dear Antti, Is there a way/best practice to model Curtailment? to calculate the amount of energy "not sold", "not fed into the grid" (per technology, i.e. Solar, Wind power plants,..), and translate this into costs? (per instance, as a variable cost per each kWh that is "not consumed"). My aim is to reflect in the model the cost implications of curtailment per technology. Best regards, Diana.
Dear Diana, Can you please elaborate why you think curtailment has a cost in the optimization? If the model finds it optimal to curtail I assume that it is rather a cost to not curtail. Agree? This is given that you have included sufficient investment options in flexibility measures such as batteries, new grid investments, fuel switch and so on... Pernille
Dear Antti and Pernille.S, Thanks a lot for your help. I think I can work with the TIMES dispatching features Antti mentioned. Dear Pernille, I see it this way: The model chooses to curtail because curtailment does not have a cost for certain technologies or considers only the cost of generation. In real life instead, curtailment does have a cost. So what would be good is that the model chooses to invest in grid or storage, etc, rather than curtail and, as you said, if the model finds it optimal to curtail, it should be taking into consideration that curtailment does has a cost (such as paying for the consumption of the extra energy or stopping/starting the power plant). But is just an idea, still in the process to decide if it will be taken into consideration and how. Thanks again