Scenario Extraction and Complementarity Analysis of Cascade Hydropower-Photovoltaic Hybrid System using UMAP and DPC

Combining photovoltaic (PV) power with a cascade hydropower (CHP) system is an effective way to promote the consumption of PV. However, the PV uncertainties create challenges to the traditional operation of pure CHP system and thereby cause difficulties in the analysis of complementarity. Typical scenarios generation method is a choice to solve this problem. However, the conventional clustering-based scenario reduction methods have low accuracy and struggle to address high-dimensional feature data. Therefore, this paper proposed a typical scenario extraction method based on the unified manifold approximation and projection (UMAP) algorithm and density peaks clustering (DPC) algorithm. The global and local structure information of feature data can be captured accurately, so the spatial and temporal correlation characteristics between the uncertain CHP and PV variables can be reserved. Thus, optimal typical scenarios of complementary operation can be extracted. The proposed method is applied on a practical hybrid system, the effectiveness is verified and the complementarity is analyzed based on several indexes.