机理与数据融合驱动的碳捕集预处理系统动力学预测研究OA

[Objective]The carbon capture pretreatment system faces challenges such as high energy consumption,unstable purification efficiency,and significant fluctuations in SO2 absorption efficiency due to variations in pH value of washing solution.This study proposes a hybrid modeling method combining mechanism models with data-driven by taking the carbon capture pretreatment system in a power plant as the research object.[Methods]By integrating chemical reaction kinetics and decision tree algorithms,the model is implemented in Python to predict key parameters accurately,including the pH value of the washing solution and the SO2 mass concentration at the system outlet.[Results]The model yields a correlation coefficient of 0.85 and 0.80 for the pH value of the washing solution and the SO2 mass concentration at the system outlet,respectively.The values fall within an acceptable error band,indicating the proposed model has good simulation performance.Moreover,a sensitivity analysis driven by baseline plant data further reveals that the model faithfully reproduces the system response to perturbations in inlet flue-gas temperature,liquid-to-gas ratio,and alkali feed rate.[Conclusion]These outcomes furnish a quantitative foundation for subsequent optimization of the CO2-capture pretreatment system,offering clear avenues for energy minimization and robust steady-state operation.