Peak reduction via time-variable tariffs and automated load control: Results from a Swiss pilot study

Time-variable grid tariffs are a potential means to incentivize grid-friendly load shifting, thereby reducing power peaks and grid reinforcement needs. Various real-world studies have investigated the average consumption change during high- and low-price periods, but only a few have analyzed the effects on peaks. In contrast, this work focuses on daily power peaks at a transformer station level, testing a fixed time-of-use tariff and a dynamic real-time tariff in a real-world pilot. Both schemes consider automated control of electric water heaters, heat pumps, and electric vehicle charging. The analyses show that the real-time tariff led to higher injection peak reductions in summer, while neither scheme could consistently reduce consumption peaks in winter. Additionally, the work highlights challenges regarding experimenting in real-world settings with ongoing system changes and proposes a baseline load estimation approach addressing these changes. • Automated electric water heater, heat pump, and electric vehicle charging control. • Reinforcement learning-based decision-making with limited information. • Baseline load estimation approach considering ongoing system changes. • Real-time tariff yields higher injection peak reductions than Time-of-Use tariff. • Consumption peaks in winter cannot be consistently reduced.