基于机器学习的有机胺吸收CO2性能预测
艾孜买提江·艾尔肯 , 代琴 , 王亚囡 , 张天琦 , 焦燕麟 , 张稚杰 , 汪黎东 , 于广飞
现代化工 ›› 2025, Vol. 45 ›› Issue (S2) : 439 -443.
基于机器学习的有机胺吸收CO2性能预测
Machine learning based prediction on performance of organic amine in absorbing CO2
针对传统开发方式周期长、成本高、效率低的问题,从22种有机胺CO2吸收体系参数中筛选并采用了10种关键特征变量,通过7种不同的机器学习方法构建了常见有机胺溶剂吸收CO2性能的预测模型。研究结果表明,GBDT模型具有最佳的预测表现。基于SHAP、PI及MDI方法的特征重要性分析结果显示,二氧化碳分压、温度、胺基百分比是影响CO2吸收效果的关键特征,并以此揭示了其优势区间。
In order to solve the long cycle,high cost and low efficiency problems of traditional development methods,10 key characteristic variables are selected from 22 kinds of parameters in organic amine-route CO2 absorption system,and a prediction model for CO2 absorption performance of common organic amine solvent is constructed through 7 kinds of machine learning methods.It is found that GBDT model presents the best prediction performance.The characteristic importance analysis results based on SHAP,PI and MDI methods show that the partial pressure of carbon dioxide,temperature,and amino group percentage are the key characteristics that affect the CO2 absorption effect,and thus their advantage ranges are revealed.
CO2捕集 / 液相吸收法 / 有机胺 / 预测模型 / 机器学习
CO2 capture / liquid phase absorption method / organic amine / prediction model / machine learning
| [1] |
赵志强, 张贺, 焦畅, |
| [2] |
盛依依. 二氧化碳捕集技术研究及工业化进展[J]. 现代化工, 2025, 45(3):1-6. |
| [3] |
|
| [4] |
|
| [5] |
|
| [6] |
|
| [7] |
侯克鹏, 包广拓, 孙华芬. SSA-MLP模型在岩质边坡稳定性预测中的应用[J]. 安全与环境学报, 2024, 24(5):1795-1803. |
| [8] |
|
| [9] |
唐菲菲, 胡佳赢, 马英, |
| [10] |
|
| [11] |
|
| [12] |
|
| [13] |
|
| [14] |
|
| [15] |
|
| [16] |
|
| [17] |
|
| [18] |
|
北京市自然科学基金青年基金项目(8254059)
国家自然科学基金青年科学基金项目(52200126)
国家自然科学基金国家杰出青年科学基金项目(52325004)
/
| 〈 |
|
〉 |