Enhanced time series aggregation for long-term investment planning models of energy supply infrastructure in production plants

Utilizing inherent flexibility from production plants can contribute to successfully incorporate volatile renewable sources into power grids. Integrated modeling approaches for electrical and thermal demands can capture flexibility options from multi-modal coupling of energy supply systems. Models with full annual time series are computationally highly complex. Computation time can be significantly reduced by aggregation methods. This paper proposes a novel approach for time series aggregation based on integer programming. The effectiveness of the proposed method is illustrated for three production plants in Southern Germany. In contrast to a heuristic seasonal selection, the novel aggregation approach systematically achieves similar results as the full annual time series. Results suggest that the relevant statistical features for the examined sites can be represented by an optimized selection of 10 days.