Development of adaptive time patterns for multi-dimensional power system simulations

The changes in the European power system come with the necessity of modeling the power system in high detail. Especially, when applying stochastic simulation approaches this leads to increasing problem sizes. In this work, we introduce a methodology to reduce the size of optimization problems in the temporal dimension to achieve lower computation times. The method is based on a mixed-integer optimization reducing the modeled time intervals that can be applied to a wide variety of optimization problems. The potential of the approach is proven by a linear unit dispatch problem for the European power system in the year 2024. The comparison of an equidistant and predefined time pattern with the preceding optimization of an adaptive time pattern shows improvements in accuracy regarding the deviation in yearly power generation, which range between 20 % and 25 %, without increasing computational requirements regarding time or hardware.