Due to the hike in fuel price and environmental awareness by the International Maritime Organization, more attention has been given in order to optimize the fuel consumption of ships. The capability to predict the fuel consumption of ships plays a significant role in the optimization process. To date, most research on predicting ship fuel consumption did not consider marine environmental factors such as wind, wave, current, and etc. Furthermore, traditional statistical methods on predicting ship fuel consumption have low accuracy. In this paper, two different sets of data showing the fuel consumption of a voyage ship with and without the influence of marine environmental factors were obtained. The Back-Propagation Neural Network (BPNN) and Gaussian Process Regression (GPR) techniques in machine learning were used to train and predict the two datasets. Thereafter, the predictive performance of these two techniques was compared and analyzed. Results showed that both techniques were able to accurately predict the ship fuel consumption, especially on the dataset with the influence of marine environmental factors. Quantitatively, the mean prediction accuracy for GPR (mean R 2 = 0.9887) is slightly higher than BPNN (mean R 2 = 0.9817). However, GPR requires longer runtime (mean T = 2236.4 s) compared to BPNN (mean T = 14.7 s). Due to the longer runtime, GPR is less preferable for online and real-time prediction of enroute ship fuel consumption. The ship real-time fuel consumption data can be accurately predicted by machine learning, which will be beneficial to achieve the goal of ship fuel consumption optimization and greenhouse gas emission reduction in the future. INDEX TERMS Machine learning, Gaussian process regression, back-propagation neural network, enroute ship, fuel consumption prediction.

No abstract

Port State Control (PSC) inspection data is used for determining the inspection pattern of PSC in Malaysia and identifying the relationship between the inspection place, flag state, number of deficiency, detention result, and ship risk profile. Based on 8,089 inspection reports from 2015 to 2019, the mining association rule is proposed as a learning approach due to its determination pattern in the information bank. The learning of association rules of PSC inspections is performed primarily on the Apriori Algorithm, in order to produce alluring rules. Inspection patterns of Malaysian ports revealed that flag state, ship risk profile, and inspection place generally lead to no detention result, as well as zero deficiency recorded on-board. The reported quantity of detention was significantly related to the high number of deficiencies raised for ships registered under blacklisted countries. Furthermore, the analysis of deficiency discovered the pattern of inspection at Malaysian ports is frequently related to zero and a low number of deficiencies raised by inspectors. Lastly, five major ports were selected for providing a useful rule to help PSC officers in organising an effective inspection plan. A similar approach can also be used for other ports beyond Malaysia for comparative analysis.

No abstract

Accurate, reliable, and real-time prediction of ship fuel consumption is the basis and premise of the development of fuel optimization; however, ship fuel consumption data mainly come from noon reports, and many current modeling methods have been based on a single model; therefore they have low accuracy and robustness. In this study, we propose a novel hybrid fuel consumption prediction model based on sensor data collected from an ocean-going container ship. First, a data processing method is proposed to clean the collected data. Secondly, the Bayesian optimization method of hyperparameters is used to reasonably set the hyperparameter values of the model. Finally, a hybrid fuel consumption prediction model is established by integrating extremely randomized tree (ET), random forest (RF), Xgboost (XGB) and multiple linear regression (MLR) methods. The experimental results show that data cleaning, the size of the dataset, marine environmental factors, and hyperparameter optimization can all affect the accuracy of the model, and the proposed hybrid model provides better predictive performance (higher accuracy) and greater robustness (smaller standard deviation) as compared with a single model. The proposed hybrid model should play a significant role in ship fuel consumption real-time monitoring, fault diagnosis, energy saving and emission reduction, etc.