Initial situation
In the predecessor SFOE P&D project GIASES, the technical prerequisites for the use of flexibilities in the distribution grid were established and tested in a real grid environment. In order to exploit the full potential of these flexibilities for meaningful load shifting – and thus a reduction in the need for grid expansion – the owners must be actively involved in the use of flexibility. The SFOE P&D project EkoFlex defines and evaluates various business models that enable precisely this involvement.
Objectives
The main objective was the demonstration and implementation of a multifactorially optimised flexibility management. By aligning the interests of utilities with customer acceptance and behaviour, the need for grid reinforcement was to be reduced as far as possible. In addition, the algorithms for disaggregation, forecasting and load shifting developed in the predecessor project were to be optimised and supplemented by an MLOps workflow.
Approach
The approach combined technical development with social science research. Qualitative guided interviews and a subsequent quantitative survey examined the acceptance of price- and reward-based demand response programs. In parallel, the algorithms for the disaggregation and forecasting of heat pumps and EV charging stations were further developed and integrated into a production-ready MLOps process via MLflow. New tariff models – Flex, Flex Plus, as well as existing heat pump and boiler tariffs – were introduced at the Technische Betriebe Vilters-Wangs at the beginning of 2025 and evaluated on the basis of real consumption data.
Key findings
The surveys confirm a fundamental openness on the part of end customers towards flexibility programs, provided that the effort remains low and, in addition to financial incentives, ecological incentives as well as a contribution to regional security of supply are apparent. However, the behavioural data revealed a clear intention-behaviour gap: the newly introduced Flex tariff led to only a negligible shift in consumption away from peak tariff times. By contrast, the active, automated control of flexibilities proved to be significantly more effective. On the technical side, the forecasting algorithms for heat pumps were noticeably improved and a new probabilistic model for EV charging stations was developed. The disaggregation of heat pumps in single-family homes remains challenging due to the lack of separately metered training data.
Conclusions
Price-based instruments alone are not sufficient to effectively unlock the flexibility potential in distribution grids. One of the greatest levers lies in the direct, automated control of controllable loads by utilities, complemented by behaviour-oriented measures that make the individual and collective added value visible. The results form a robust basis for the further development of a flexibility market in Swiss distribution grids and contribute to the objectives of the Energy Strategy 2050.