基于大数据分析的船舶功率优化应用  

Ship Power Optimization with Big Data Analysis

作　　者：刘柱[1] 姚久武 李迪阳[1] 张宝清 周利江[1] LIU Zhu;YAO Jiuwu;LI Diyang;ZHANG Baoqing;ZHOU Lijiang(Qingdao Ocean Shipping Mariners College,Qingdao 266071,China;China COSCO Shipping Corporation Limited,Beijing 100031,China;Transportation College,Chongqing Jiaotong University,Chongqing 400074,China)

机构地区：[1]青岛远洋船员职业学院,山东青岛266071 [2]中国远洋海运集团有限公司,北京100031 [3]重庆交通大学交通运输学院,重庆400074

出　　处：《中国航海》2019年第2期12-16,共5页Navigation of China

基　　金：交通运输部“交通运输行业高层次技术人才培养项目”(人教人才[2017]518号);中国远洋运输(集团)总公司科研项目(2016-1-R-005);重庆市教委科学技术研究项目(KJ1600509);重庆市科委基础研究与前沿探索项目(cstc2016jcyjA0561)

摘　　要：为探明船舶主机油耗和优化方向,基于“COSCO Spain”和“COSCO Portugal”两船在一段时间内连续航行的实例数据,构建BP(BackPropagation)神经网络模型。运用大数据技术学习历史数据经验,抽象出主机功率—对水速度期望曲线L;随机改变主机功率到神经网络模型重新输出结果后,前后比较可评价耗油情况并确定主机功率的推荐调整策略。该方法与“等功率”航行做法相比更具有优势,可达到指导船舶管理和降本增效的目的,并提供一种新的基于数据的航运科学研究范式。The practical operation data acquired from the MVs "COSCO Spain" and "COSCO Portugal" in a chosen period of time are studied to make clear their fuel oil consumption situation and find the way of improvement. The shaft power-log speed curve L is defined through big data study. A BP neural network is constructed and trained with the historical data. The responses of the trained neural network to randomly adjusted shaft power inputs are studied to check the fuel consumption and to decide the main engine power management strategy. This data-based method is superior to the present used "equal power" practice in guiding the ship management for lower operational costs and higher efficiency.

关 键 词：大数据 船舶航速 功率优化 BP神经网络 

分 类 号：U676.3[交通运输工程—船舶及航道工程]