为精确预估船舶油耗，推动航运业向绿色低碳转型，提出一种基于改进BP神经网络的船舶油耗预测方法。通过对原始航行数据进行预处理，去除噪声、偏差和异常值；利用核主成分分析法将数据中的10个原始变量降维为5个主成分，减少数据维度；采用遗传算法优化BP神经网络，建立高精度的船舶油耗模型。以1艘液化石油天然气运输船为研究对象，实验结果表明，优化后的BP神经网络油耗模型在预测性能方面获得较大提升，训练集和验证集的均方根误差分别降低了0.1122和0.1068，决定系数提高1.58%。该研究成果能够为船舶节能减排提供可靠的决策支持。
In order to accurately predict the fuel consumption of ships and promote the transformation of the shipping industry to green and low-carbon, a ship fuel consumption prediction method based on improved BP neural network is proposed. By preprocessing the original voyage data to remove noise, bias and outliers; using kernel principal component analysis to reduce the dimensionality of the data by downgrading the 10 original variables in the data to five principal components; and adopting genetic algorithm to optimize the BP neural network to establish a high-precision model of ship fuel consumption. Taking a liquefied petroleum gas carrier as the research object, the experimental results show that the optimized BP neural network fuel consumption model obtains a large improvement in the prediction performance, and the root mean square error of the training set and validation set is reduced by 0.1122 and 0.1068 respectively, and the coefficient of determination is improved by 1.58%. The research results can provide reliable decision support for energy saving and emission reduction of ships.
2025,47(11): 149-154 收稿日期：2024-8-17
DOI：10.3404/j.issn.1672-7649.2025.11.026
分类号：U677.2
基金项目：厦门市自然科学基金资助项目（502Z202372019）
作者简介：吴泽颖(2000-),男,硕士,研究方向为交通运输规划与管理
参考文献：
[1] MEPC, RESOLUTION. 2022 Guidelines for the development of a ship energy efficiency management plan (SEEMP)[Z]. 2022.
[2] YUAN Z, LIU J X, ZHANG Q, et al. Prediction and optimisation of fuel consumption for inland ships considering real-time status and environmental factors[J]. Ocean Engineering, 2020, 221(12), 108530.
[3] RAN Y, WANG S A, DU Y Q. Development of a two-stage ship fuel consumption prediction and reduction model for a dry bulk ship[J]. Transportation Research Part E: Logistics and Transportation Review, 2020, 138: 101930.
[4] YAN X, SUN X, YIN, Q. Multiparameter sensitivity analysis of operational energy efficiency for inland river ships based on backpropagation neural network method[J]. Marine Technology Society Journal, 2015, 49(1): 148-153.
[5] THANH L L, GUNWOO L, PARKK S, et al. Neural network-based fuel consumption estimation for container ships in Korea[J]. Maritime Policy & Management 2020, 47(5): 615-632.
[6] HU Z H, JIN Y X, HU Q Y, et al. Prediction of fuel consumption for enroute ship based on machine learning[J]. IEEE Access, 2019(7): 119497-119505.
[7] NGUYEN V, RAJAMOHAN S. Using artificial neural networks for predicting ship fuel consumption[J]. Polish Maritime Research, 2023, 23(2): 39-60.
[8] 孟忍, 董学平, 甘敏. 基于KPCA和BiLSTM的分解炉出口温度预测[J]. 合肥工业大学学报(自然科学版), 2023, 46(2): 169-174.
MENG R, DONG X P, GAN M. Temperature prediction of decomposition furnace outlet based on KPCA and BiLSTM[J]. Journal of Hefei University of Technology (Natural Science Edition), 2023, 46(2): 169-174.
[9] 张博峰. 基于KPCA与径向基神经网络的滚动轴承故障诊断[J]. 内燃机与配件, 2021(13): 133-135.
ZHANG B F. Rolling bearing fault diagnosis based on KPCA and radial basis neural network[J]. Internal Combustion Engines and Accessories, 2021(13): 133-135.
[10] 刘璐, 杨丹, 陈睿杰, 等. 基于KPCA-GA-BP神经网络的POI质量预测研究[J]. 电信科学, 2023, 39(1): 108-116.
LIU L, YANG D, CHEN R J, et al. Research on POI quality prediction based on KPCA-GA-BP neural network[J]. Telecommunication Science, 2023, 39(1): 108-116.
[11] 王雨涵, 张亚萌, 魏国亮. 基于BP神经网络和高阶模糊认知图的股票价格预测[J]. 智能计算机与应用, 2023, 13(8): 100-106+113.
WANG Y H, ZHANG Y M, WEI G L. Stock price prediction based on BP neural network and high-order fuzzy cognitive map[J]. Intelligent Computer and Applications, 2023, 13(8): 100-106+113.
[12] 陈亮, 沈良朵, 陆凡, 等. 基于主成分的BP神经网络共享单车骑行预测[J]. 网络安全技术与应用, 2023(7): 51-55.
CHEN L, SHEN L D, LU F, et al. Principal component-based BP neural network shared bicycle riding prediction[J]. Network Security Technology and Application, 2023(7): 51-55.
[13] 徐森雨, 田利军. 基于GA-BP神经网络的航空货运量预测[J]. 综合运输, 2023, 45(9): 75-79.
XU S Y, TIAN L J. Air cargo volume forecasting based on GA-BP neural network[J]. Comprehensive Transportation, 2023, 45(9): 75-79.
[14] 熊豪杰, 魏怡. 基于改进CNN-BP的多波束声纳高程数据预测研究[J]. 计算机科学, 2023, 50(S1): 575-578.
XIONG H J, WEI Y. Research on multibeam sonar elevation data prediction based on improved CNN-BP[J]. Computer Science, 2023, 50(S1): 575-578.
[15] 周兰喜, 王楠, 周旭. 基于营运数据的船舶能效分析及预测[J]. 船舶与海洋工程, 2024, 40(2): 34-38.
ZHOU L X, WANG N, ZHOU X. Analysis and prediction of ship energy efficiency based on operation data[J]. Ship and Ocean Engineering, 2024, 40(2): 34-38.
[16] 王凯, 徐浩, 黄连忠, 等. 基于机器学习的船舶能耗智能预测方法分析[J]. 船舶工程, 2020, 42(11): 87-93.
WANG K, XU H, HUANG L Z, et al. Analysis of intelligent prediction method of ship energy consumption based on machine learning[J]. Ship Engineering, 2020, 42(11): 87-93.
[17] 陈伟南, 黄连忠, 张勇, 等. 基于BP神经网络的船舶主机能效状态评估[J]. 中国舰船研究, 2018, 13(4): 127-133+160.
CHEN W N, HUANG L Z, ZHANG Y, et al. Energy efficiency state assessment of ship main engine based on BP neural network[J]. China Ship Research, 2018, 13(4): 127-133+160.
[18] 郑齐清, 陈佳鑫, 杨何伍. 基于PCA分析法的船舶主机能效评价[J]. 广州航海学院学报, 2023, 31(2): 41-46.
ZHENG Q Q, CHEN J X, YANG H W. Energy efficiency evaluation of ship main engine based on PCA analysis[J]. Journal of Guangzhou Nautical College, 2023, 31(2): 41-46.
