| 摘要: |
| 目的 在碳捕集封存技术发展背景下，为能更高效准确地计算CO2在水中的溶解度，分别探究基于机理分析与基于数据驱动的CO2溶解度计算方法。方法 首先，根据公开文献收集、整理了2 029组不同温度、压力条件下CO2在水中的溶解度数据并建立数据库；随后，利用该数据库对经典的亨利定律进行了修正，并且在修正时考虑压力在水合物生成区域内可能生成水合物的影响；最后，基于BP神经网络建立了数据驱动模型来计算CO2在水中的溶解度。数据驱动模型建立的过程中，以相对误差作为模型评价指标，优选温度、压力和矿化度作为输入变量。结果 亨利定律修正后压力低于15 MPa时计算结果的平均相对误差为25.29%，与修正前相比下降了27.11个百分点；压力高于15 MPa时计算结果的平均相对误差为29.49%，与修正前相比下降了91.86个百分点。数据驱动模型的平均相对误差降低至20.64%，计算结果比修正后的亨利定律更为准确。结论 提出了简便有效、精度更高的CO2溶解度计算方法。 |
| 关键词: 碳排放 CO2 溶解度 亨利定律 BP神经网络 |
| DOI：10.3969/j.issn.1007-3426.2025.02.008 |
| 分类号: |
| 基金项目:国家自然科学基金“裹挟粉砂的水基流动体系中CO2-CH4水合物分解动力学机制”（52104069）；北京市自然科学基金“深水油气混输管道中赋存蜡的水合物生成传热传质机制研究”（3192027）；中国石油大学（北京）科学基金“天然气水合物开发过程中水合物分解动力学机理研究”（2462023BJRC018）；中国石油大学（北京）科学基金“基于大数据的天然气管网智能运行与控制研究”（2462020YXZZ045） |
|
| Calculation method of CO2 solubility based on mechanism analysis and data driven |
|
Fengyuan YAN1,2, Haishen LEI3,4, Rui WANG5, Junyao GUO2, Ju LIN2, Jinxu LI1, Lei ZHANG1, Bohui SHI2, Shangfei SONG2
|
|
1.Institute for Advanced Materials and Technology, University of Science and Technology Beijing, Beijing, China;2.National Engineering Research Center of Oil and Gas Pipeline Transportation Safety, College of Mechanical and Transportation Engineering, China University of Petroleum, Beijing, China;3.The Fifth Electronic Research Institute of Ministry of Industry and Information Technology, Guangzhou, Guangdong, China;4.Ministry of Industry and Information Technology Key Laboratory of Industrial Software Engineering Application Technology, Guangzhou, Guangdong, China;5.Pipe China Digital Co., Ltd, Beijing, China
|
| Abstract: |
| Objective In the context of the development of carbon capture and storage technology, in order to more efficiently and accurately calculate the solubility of carbon dioxide (CO2) in water, both mechanism analysis-based and data-driven methods for calculating the solubility of CO2 in water are explored, respectively. Method Firstly, based on public literature, 2 029 sets of solubility data of CO2 in water under different temperature and pressure conditions were collected and organized, and a relevant database was established. Subsequently, the classic Henry's law was revised using the database, and the influence of pressure on the formation of hydrates in the hydrate formation region was considered during the revision. Finally, a data-driven model was established based on BP neural network to calculate the solubility of CO2 in water. In the process of establishing the data-driven model, relative error was used as the model evaluation index, and temperature, pressure, and mineralization were selected as input variables. Result The average relative error of the calculated results when the pressure is below 15 MPa after the correction of Henry's Law is 25.29%, a decrease of 27.11 percentage points compared to before the correction; the average relative error of the calculated results when the pressure is higher than 15 MPa is 29.49%, which has decreased by 91.86 percentage points compared to before correction. The average relative error of the data-driven model has been reduced to 20.64%, and the calculated results are more accurate than the modified Henry's Law. Conclusion A simple, effective, and more accurate calculation method for the solubility of CO2 in water has been proposed, which can provide an important reference for calculating the solubility of CO2 in water in engineering. |
| Key words: carbon emission carbon dioxide(CO2) solubility Henry's law BP neural network |
