Uncertain flexoffers, a scalable, uncertainty-aware model for energy flexibility

The usage of Renewable Energy Sources in electricity grids is spreading more and more, and energy flexibility has a crucial role in it. Many models for flexibility have been proposed; in particular, Flex-Offer (FO) is an approximate model with the following properties: i) it models flexibility from different device types in an unified format; ii) it is scalable with respect to optimization for long time horizons, and aggregation of many loads; iii) it captures most/all of the available flexibility. This work proposes an extension of that model, uncertain FlexOffers (UFOs), which has the same properties, plus iv) considers the uncertainty affecting flexibility over long time horizons. We show that UFOs are 5 orders of magnitude faster than exact models for optimization, can retain 88.3% of the total flexibility when imbalance penalties are high, and allow to aggregate up to 3000 loads and up to 96 time units in less than 30 minutes.