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(Back Propagation)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.
引文
[1] NORSTAD I,FAGERHOLT K,LAPORTE G.Tramp Ship Routing and Scheduling with Speed Optimization[J].Transportation Research Part C:Emerging Technologies,2011,19(5):853-865.
[2] SHENG XM,CHEN EP,LEE LH.Policy Modal for Liner Shipping Refueling and Sailing Speed Optimization Problem[J].Transportation Research Part E:Logistics and Transportation Review,2015,76:76-92.
[3] 邢玉伟,杨华龙,张燕.考虑碳税成本的班轮航线配船与航速优化[J].上海海事大学学报,2017,38(4):1-5.
[4] 俞姗姗,汪传旭.不同碳排放调控政策下的船舶航速优化[J].大连海事大学学报,2015,41(3):45-50.
[5] 王冀.动态规划和模糊算法在船舶航速最优调度中的应用[J].舰船科学技术,2016,38(18):136-138.
[6] 黄连忠,万晓跃,孙永刚,等.基于模拟退火算法的船舶航速优化研究[J].船舶,2018(S1):8-17.
[7] 李铮,潘晓萌.船舶航速优化算法研究[J].舰船科学技术,2016,38(12):7-9.
[8] 张进峰,马伟皓,刘永森,等.考虑营运成本和排放的船舶航速多目标优化模型[J].中国航海,2017,40(1):129-134.
[9] 王胜正,申心泉,赵建森,等.基于ASAE深度学习预测海洋气象对船舶航速的影响[J].交通运输工程学报,2018,18(2):143-151.