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

摘要：

燃烧后碳捕集技术被认为是目前减少燃煤电厂碳排放最可行的技术之一。但化学法碳捕集装置能耗较高，占用发电负荷，需要将其与发电过程耦合调度，通过错峰捕集降低系统运行成本。本文将碳捕集过程与发电过程集成，通过引入烟气旁路和溶剂储罐两种辅助设备，基于以小时为单位的动态电价和用电量，对耦合系统进行优化调度。针对因碳捕集过程机理模型复杂而无法与系统调度模型协同求解的难题，本文基于流程模拟和神经网络训练构建碳捕集过程代理模型，考虑电力市场和碳市场收益，以电厂日收益最大为优化目标，采用数学规划法优化设计调度方案。最后，以一个600MW的发电厂为算例，进行了系统的最优调度方案和碳捕集过程操作参数的优化。基于优化结果分析了电厂与碳捕集协同调度的规律特性，验证了集成方法的有效性，展示了其在优化碳捕集与电厂集成调度方面的应用前景。

关键词: 碳捕集, 电厂, 优化调度, 代理模型, 人工神经网络