基于数据挖掘的船舶能耗分析与预测

Analysis and Prediction of Ship Energy Consumption Based on Data Mining

船舶能耗影响因素众多,涉及大气环境、水文条件、航速工况、地理位置、任务编排以及有效载重等多种特征因素,若采集所有特征会造成信息采集系统、数据存储系统和数据分析系统压力过大,且可能采而不用。因此,以船舶能耗分析和建模为目标,基于随机森林挖掘不同方面的数据与船舶能耗之间的关联关系,合理规划能量管理的采集范围。同时,基于关联性强的因子进行能耗预测和里程预测,可为合理安排航行任务提供重要的决策依据。

There are many factors affecting ship energy consumption,including atmospheric environment,hydrological conditions,sailing speed,geographical location,task arrangement,effective load and other characteristic factors.If all characteristics are collected,the pressure of information acquisition system,data storage system and data analysis system will be caused.And maybe the data was collected but not used.Therefore,this paper takes ship energy consumption analysis and modeling as the goal.Firstly,based on the correlation between data from different aspects with random forest mining,the collection range of energy management is reasonably planned.Then,energy consumption prediction and mileage prediction are made based on factors with strong correlation,which provides important decision-making basis for reasonable arrangement of voyage tasks.