Multilevel Coordinated Strategy for Electrolytic Aluminium System to Enhance Renewable Energy Consumption

ABSTRACT

Because of the spatial mismatch between renewable energy sources and load centres, China continues to experience substantial curtailment of wind and solar power. To mitigate this issue, several regions have adopted electrolytic aluminium loads to absorb local renewable energy. Nevertheless, the absence of effective scheduling strategies for the electrolytic aluminium system has posed increasing challenges to ensuring its safe and cost‐efficient operation. This paper proposes a multilevel coordinated strategy for the electrolytic aluminium system, which accounts for the flexible interaction of source, grid, load and storage. To quantitatively characterise the coupling relationship between the electrical power and thermal dynamics of the electrolytic aluminium load, an electric–thermal coupling model is established. Based on this model, a control and scheduling framework for the electrolytic aluminium system is developed. Furthermore, a long short‐term memory (LSTM)‐extreme gradient boosting (XGBoost) hybrid probabilistic interval prediction method is proposed to enhance the forecasting accuracy of wind and solar power generation. Building on these foundations, a multilevel coordinated optimisation model is constructed that incorporates multiple influencing factors. The proposed strategy aims to maximise renewable energy utilisation while achieving dual objectives in terms of both economic performance and carbon reduction. Case studies conducted on an actual electrolytic aluminium system located in Inner Mongolia province confirm the effectiveness of the proposed approach.