A source-load collaborative stochastic optimization method considering the electricity price uncertainty and industrial load peak regulation compensation benefit

Energy-intensive industrial load offers substantial capacity and rapid adjustment capabilities, which can be effectively coordinated with deep peak regulation (DPR) methods of thermal power to optimize the peak regulation state of the system. The uncertainty of electricity prices and the current peak regulation compensation mechanism significantly affect the economic viability of industrial load regulation. In this study, electrolytic aluminum load (EAL) is used as a representative industrial load. This paper combines the complete ensemble empirical mode decomposition adaptive noise (CEEMDAN), whale optimization algorithm (WOA), and long short-term memory network (LSTM) to propose a CEEMDAN-WOA-LSTM prediction model for electricity price scenarios. Subsequently, comprehensive cost and fine adjustment models for electrolytic aluminum load (EAL) are developed, incorporating the current peak regulation compensation mechanism. Finally, a source-load collaborative stochastic optimization method is proposed, integrating the scenario method and chance constraints. The effectiveness of the proposed scheme is verified using a real regional system, demonstrating significant reductions in total social peak regulation costs, a substantial decrease in renewable energy (RE) abandonment rates, reduced frequency of thermal power DPR, and improved economic efficiency of thermal power. Additionally, the current peak regulation compensation mechanism effectively guarantees the benefits of EAL and encourages its adjustment willingness.