期刊
ENERGY
卷 278, 期 -, 页码 -出版社
PERGAMON-ELSEVIER SCIENCE LTD
DOI: 10.1016/j.energy.2023.127834
关键词
Carbon capture and storage (CCS); Post -combustion carbon capture; Surrogate machine learning; Deep learning; SHapley additive exPlanations (SHAP); Explainable artificial intelligence (XAI)
The objective of this study is to develop an artificial intelligence (AI) based approach that predicts the utility consumptions of different absorption-based carbon capture (ACC) process designs. The AI models show an error between 0.4% and 3.6% for predicting the specific reboiler duty, which is crucial for the operational costs of ACC plants. Moreover, the study identifies the influence of each process parameter on energy consumption using explainable artificial intelligence.
Absorption-based carbon capture (ACC) is the most mature technology for abating CO2 among different postcombustion carbon capture technologies. ACC is, however, energy-intensive and requires large heating and cooling utility consumption, resulting in high operational costs. Reliable and fast estimation of these utility consumptions would be valuable for technical and financial feasibility assessment of different ACC process designs. The objective of this study is to develop an artificial intelligence (AI) based approach that predicts these utility consumptions of different ACC process designs. For prediction of specific reboiler duty, which is the most important energy consumption of the plant and the main source of operational costs in the ACC process, the error of the AI models is between 0.4% and 3.6% for different scenarios of utilizing either limited or extensive training data. Moreover, the contribution of each process parameter to energy consumption was identified using explainable artificial intelligence. The findings of this study facilitate the energy-efficient design of ACC plants and provide a prioritized list of parameters to adjust for reducing the energy consumption of an ACC process.
作者
我是这篇论文的作者
点击您的名字以认领此论文并将其添加到您的个人资料中。
推荐
暂无数据