By aggregating the flexibility (flex) of hundreds of consumer appliances different demand-side management (DSM) strategies can be carried out, such as demand response, energy trading or ancillary services. However, this means that users must share their available local flex with the aggregator while being fairly rewarded for their commitment. Unfortunately, not everyone is willing to expose their energy data, which makes privacy a major obstacle to broaden the use of household flex in demand control.
Nevertheless, flex will not be possible without data. Users demand is highly linked to their behavior, which introduces many uncertainties in the available flex. Moreover, local flex can also be achieved from heterogeneous sources, increasing the complexity of the problem. In this way, although data must stay private, we must at least be able to capture uncertainties.
The PrivateFlex project aims to tackle these issues by making use of cryptographic algorithms such as computation over encrypted data (COED) to keep flex data local and private, while at the same time allowing the flex to be traded at an aggregated level. In addition, the second objective is to better characterize flex by means of machine learning (ML) techniques both at a local and at an aggregate level.
Impact of the project: Cost-effective solutions that address privacy issues and deal with uncertainty are important to ensure that demand response is implemented in residential buildings. The whole value chain is represented in the consortium by leading industrial partners active in Flanders.