Multi-timescale rolling optimization strategy for integrated energy systems considering renewable energy consumption and source–load uncertainty

Integrated energy systems enhance renewable energy consumption and supply efficiency through optimized scheduling of energy production, conversion, and storage. This paper introduces a day-ahead stochastic-intraday robust two-phase multi-timescale rolling optimal scheduling model that addresses source–load uncertainty, improving renewable energy utilization and reducing operational costs. In the day-ahead phase, the model builds on traditional Monte Carlo simulation by updating the discrete probability density matrix based on the temporal characteristics of source–load data. It employs adaptive Monte Carlo simulation for generating source and load scenarios. Additionally, an improved k-means algorithm, utilizing the elbow method and considering the triangular inequality of Euclidean distances, enhances scene analysis efficiency and accuracy. In the intraday phase, the model incorporates slack variables and constraint sets, integrating them with a cardinality uncertain scenario set based on the traditional light robust optimization model. This approach reduces the conservatism typically associated with traditional robust optimization (RO) models. The proposed intraday multi-timescale rolling RO model facilitates efficient renewable energy consumption and balanced energy supply across various timescales. Overall, the findings demonstrate that the proposed model leverages multi-energy complementarities to optimize renewable energy consumption and ensure a balanced energy supply in a dynamic environment.