An Assessment Method for High-Proportion Renewable Energy Accommodation Capacity Based on Improved Typical-Day Extraction

With the advancement of the “dual carbon” goals, the assessment of the accommodation capacity of high-proportion renewable energy power systems is facing challenges. This paper, taking the Ningxia power grid as the background, proposes a calculation model for the scale of renewable energy that can be accessed based on an improved typical-day analysis. In response to the shortcomings of traditional clustering algorithms, which rely on subjective experience to determine the number of clusters and have random initial centers that are prone to falling into local optima, this paper constructs a two-step improved K-means strategy: first, an improved DBI (Davies-Bouldin Index) that integrates multiple features such as load, wind and solar power output, and remaining accommodation space is introduced to objectively determine the optimal number of clusters; second, the idea of hierarchical clustering is combined to optimize the initial clustering centers. On this basis, the source–load characteristics of typical days are mapped to the future planning boundary. Simulation comparisons based on the Ningxia power grid show that this algorithm significantly outperforms traditional methods in core indicators such as the error of reconstructed energy storage, and can accurately reflect the real energy storage requirements and operational characteristics of the system, providing scientific quantitative support for the development of renewable energy and the optimal allocation of regulation resources in regional power grids.