在绿色航运发展背景下准确预测并优化船舶碳排放是实现航运减排目标的关键路径。基于全球航运排放量多源数据集，融合船型、船龄、时间特征与多污染物因子，构建了由随机森林与极端梯度提升树（XGBoost）组成的集成回归模型，并引入堆叠策略（Stacking）优化预测效果。实验结果表明融合污染物特征后模型性能显著提升，Stacking模型的决定系数接近1，RMSE降至232.48，明显优于单一模型。残差分析显示集成模型具有更高的稳定性与泛化能力。本研究可为船舶碳排放建模提供有效路径，也为绿色航运政策制定与动力系统优化提供理论与数据支撑。
Against the backdrop of green shipping development, accurately predicting and optimizing ship carbon emissions is a critical pathway to achieving emission reduction goals in the maritime industry. Based on a global multi-source dataset of maritime emissions, this study integrates ship type, ship age, temporal features, and various pollutant factors to construct an ensemble regression model composed of Random Forest and Extreme Gradient Boosting (XGBoost). Furthermore, a Stacking strategy is introduced to enhance predictive performance. Experimental results indicate that the inclusion of pollutant features significantly improves model accuracy, with the Stacking model achieving a coefficient of determination (R2) close to 1 and a root mean square error (RMSE) reduced to 232.48, outperforming individual models. Residual analysis demonstrates that the ensemble model possesses greater stability and generalization ability. This research provides an effective approach for ship carbon emission modeling and offers theoretical and data support for green shipping policy formulation and propulsion system optimization.
2026,48(5): 164-169 收稿日期：2025-12-25
DOI：10.3404/j.issn.1672-7649.2026.05.026
分类号：U664.81；TP311.13
基金项目：第二批中国高校产学研创新基金项目（2020ITA03033）；福建船政交通职业学院科教发展基金项目（Z202311033）
作者简介：胡晶(1981-),女,硕士,副教授,研究方向为大数据分析挖掘
参考文献：
[1] 蔡欣宇, 陈焱明, 陈奕恒, 等. 基于AIS数据的长三角沿海地区船舶碳排放估算与分析[J]. 环境科学学报, 2024, 44(8): 83-97
CAI X Y, CHEN Y M, CHEN Y H, et al. Estimation and analysis of ship carbon emissions in the coastal areas of the Yangtze River Delta based on AIS data[J]. Acta Scientiae Circumstantiae, 2024, 44(8): 83-97
[2] 栾建霖, 冯胤伟, 李海江, 等. 基于深度学习模型的船舶碳排放时空预测研究[J]. 科研管理, 2023, 44(3): 75-85
LUAN J L, FENG Y W, LI H J, et al. Research on spatiotemporal prediction of ship carbon emissions based on deep learning models[J]. Science Research Management, 2023, 44(3): 75-85
[3] 章强, 陈万响. 船舶排放控制区政策对城市空气质量影响的异质性分析[J]. 上海海事大学学报, 2022, 43(3): 69-74+82
ZHANG Q, CHEN W X. Heterogeneity analysis of the impact of ship emission control area policies on urban air quality[J]. Journal of Shanghai Maritime University, 2022, 43(3): 69-74+82
[4] 施旭航, 薛青青, 陈荣昌. 船舶垃圾污染船岸协同治理问题与对策研究[J]. 交通节能与环保, 2021, 17(4): 69-72
SHI X H, XUE Q Q, CHEN R C. Research on the problems and countermeasures of ship-ship collaborative governance of ship garbage pollution[J]. Energy Conservation & Environmental Protection in Transportation, 2021, 17(4): 69-72
[5] 顾建康, 谭效时, 向可祺, 等. 船舶制造业碳排放因子获取方式研究[J]. 船舶物资与市场, 2025, 33(4): 47-49
[6] 林博闻, 马晓凤, 庄义彬. 港口船舶尾气排放量及泊位调度优化研究[J]. 中国航海, 2024, 47(1): 88-96
LIN B W, MA X F, ZHUANG Y B. Study on port ship emission volume and berth allocation optimization[J]. Navigation of China, 2024, 47(1): 88-96
[7] 郭文俊, 杨泽民, 张叶娥, 等. 基于AIS无线网络的舰船排污智能监测系统[J]. 舰船科学技术, 2025, 47(7): 169-173
GUO W J, YANG Z M, ZHANG Y E, et al. Intelligent monitoring system for ship sewage discharge based on AIS wireless network[J]. Ship Science and Technology, 2025, 47(7): 169-173
[8] 崔秀芳, 赖炜祺, 林浩涛, 等. 基于AIS和波浪数据的船舶碳排放计算[J]. 环境科学与技术, 2025, 48(1): 192-200
CUI X F, LAI W Q, LIN H T, et al. Calculation of ship carbon emissions based on AIS and wave data[J]. Environmental Science & Technology, 2025, 48(1): 192-200
[9] 郭瑞芳, 王莹. 绿色能源背景下的舰船动力系统能耗优化研究[J]. 舰船科学技术, 2024, 46(20): 129-132
GUO R F, WANG Y. Research on energy consumption optimization of ship power systems under the background of green energy[J]. Ship Science and Technology, 2024, 46(20): 129-132
[10] 梅冠军. 改造船舶碳捕捉及存储技术应用[J]. 舰船科学技术, 2025, 47(6): 123-127
MEI G J. Application of retrofitted ship carbon capture and storage technology[J]. Ship Science and Technology, 2025, 47(6): 123-127
