Research On Fuel Consumption Modeling And Speed Optimization Based On Real Ship Monitoring Data

As an important role in global trade,the shipping industry has developed rapidly in recent years.However,a large number of greenhouse gases and harmful gases such as carbon dioxide,nitrogen oxides,and sulfur oxides are generated during the process of shipping goods and passengers,which seriously endangers climate change and human health.To alleviate severe climate and environmental problems,International Maritime Organization(IMO)and relevant organizations and governments have formulated a series of policies and regulations to promote energy conservation and emission reduction in the shipping industry.The speed optimization based on data driven method is a current research hotspot,but there is still has some promotion space in data processing,fuel consumption prediction,and optimization methods.Considering the characteristics of ocean-going vessels in large-tonnage,heavy cargo,and long voyage distance,it not only directly reduces the fuel consumption and operation cost but also actively responds to IMO’s the ship energy efficiency management for ship owners to make reasonable and effective speed decisions.This paper analyzes the main factors affecting ship’s energy consumption based on monitoring data,carries out the data collection and analysis of energy efficiency,and studies the energy efficiency improvement methods for optimizing the speed of ocean-going routes.The innovations of this paper mainly include the following aspects:(1)Aiming at the problem that the existing data cleaning methods of energy consumption mainly rely on the knowledge of the navigation field,a data processing method that considers the internal relationship of energy consumption characteristics is proposed.The method can not only identify obvious errors,and missing and repeated data,but also effectively identify potential abnormal data.Firstly,the energy consumption characteristic data of different loading states are extracted according to the range of ship drafts.Then,the visualized characteristics data is displayed and analyzed to find the distribution regulation of ship energy consumption,main engine speed,and speed over ground in the different loading states.Finally,these energy consumption data are cleaned precisely and efficiently.The experimental results show that the proposed method can increase the energy consumption prediction accuracy and improve the data quality.(2)To solve the problem that it is difficult to obtain a better combination of hyperparameter values by using empirical values or manual settings in the existing energy consumption models,a ship energy consumption prediction model based on a gradient boosting decision tree(GBDT)algorithm with adaptive parameter adjustment is proposed.Firstly,grid search,random search,and Bayesian search methods are used respectively to optimize the model’s hyperparameters.Considering the model’s performance evaluation index and running time of the computer program together,the Bayesian search algorithm is selected as the hyperparameter optimization method.Then,combined with the five-fold cross-validation method,the optimal hyperparameter group is obtained.The experimental results show that the prediction accuracy of the proposed GBDT algorithm is improved by 5.63%after hyperparameter optimization,contributing to the higher,prediction accuracy of ship energy consumption than other algorithms.The algorithm can accurately predict energy consumption and meet the requirements of real-time prediction.(3)The previous speed optimization issues are mainly based on historical navigation trajectory information,while it is difficult to achieve real-time speed optimization.To improve it,this paper establishes a speed optimization method based on real-time environmental forecast information.Firstly,an algorithm for extracting weather forecast data with sailing at any speed during a period of time via grid is designed.Then,combined with the ship navigation state,the main engine working conditions,and the information of weather and sea,the GBDT model is applied to obtain the corresponding relationship between ship speed and energy consumption under different draught,trim,weather and sea conditions.Taking the minimization of fuel consumption during the whole voyage as the optimization goal,we have ensured the safety of ship navigation and arrived at the destination port in the scheduled time in the meanwhile.The particle swarm optimization algorithm is used to solve the speed optimization problem.Finally,taking two voyages of a large bulk carrier and container ship as case studies,we have simulated the proposed method and verified its feasibility.The simulation experiment result show that the fuel consumption of the two voyages of the bulk carrier is reduced by 5.31%and 9.55%,and the container ship is reduced by 8.3%and 6.3%respectively.In addition,we analyzed the effect of speed optimization for different sailing times,which provide important reference for formulating reasonable arrival time.This series of studies have an important social significance in the energy-saving and emission reduction aspects,and also have important academic value in theoretical research in the energy managementand speed optimization,which will contribute to the implementation of "dual carbon strategy" and "green shipping".