关键词: 碳捕集, 电厂, 优化调度, 代理模型, 人工神经网络

Abstract:

Post-combustion carbon capture technology is widely regarded as a highly effective approach to mitigating carbon emissions from coal-fired power plants. Nevertheless, the associated chemical carbon capture devices exhibit considerable energy consumption and impose a substantial load on power generation. It is essential to couple these devices with the power generation process for peak shifting to reduce operational costs. This study presented an integrated model of the carbon capture process and power generation, featuring the addition of two auxiliary devices: a flue gas bypass and solvent storage tanks. Due to the complexity of the carbon capture process mechanism model, this paper built a surrogate model through process simulation and neural network training. Considering the revenues from both the electricity and carbon markets, the optimization goal was to maximize the daily income of the power plant. A 600MW power plant was studied to achieve its optimal scheduling scheme and operational parameters. The results had shown the regularity and characteristics of the cooperative scheduling between power plants and carbon capture devices, and the effectiveness of the integration method was verified.

Key words: carbon capture, power plant, optimized scheduling, surrogate model, artificial neural network