基于实船数据的船舶航速与油耗优化建模

为获得水路运输中船舶的最佳航速,实现燃油效率的最大化,以长江航线货运船的航速、燃油等真实监测数据为基础,提出一种新的船舶航速与油耗优化模型。模型包括白箱模型(White-Box Model,WBM)和黑箱模型(Black-Box Model,BBM)两个部分:WBM由改进后的物理方程和数学公式组成;BBM是BP(Back Propagation)神经网络。两者结合成新的灰箱模型(Grey-Box Model,GBM)。通过分析船舶航行过程中影响船舶燃油消耗的多种因素,确定WBM参数,分别用串联和并联的方式将白箱与BP网络结合构建船舶航速与油耗优化模型,采用牛顿-拉普森迭代算法进行求解。计算结果表明:优化模型可在选定的航段内找到最佳的指导航速,并且模型的R2达到了0.945。此外,将建立的模型与单一的WBM进行对比验证,所建立的船舶航速与油耗优化模型得到的最佳航速更加精确,整体误差也从0.130减小到0.071。

Modeling of Fuel Consumption Versus Sailing Speed Based on Ship Monitoring Data

The field data based model is built to optimize speed of a vessel for the maximal fuel efficiency in waterway transportation.The data the model based on are from monitoring the operation of bulk carriers on the Yangtze River route.The model consists of a WBM(White-Box Model)and a BBM(Black-Box Model).The former is the improved physical and mathematical formulas with parameters based on theoretical analysis and the latter is a BP(Back propagation)neural network.The resultant models of GBM(Gray Box Model)type are constructed through integrating the two parts in series or parallelly.The models are solved by means of Newton-Raphson iterative algorithm.Calculation shows that the model is effective with the coefficient of determination up to 0.945.The introduction of the neural network makes noticeable improvement of ship speed proposal of the model,the error of which is 0.071 versus 0.130 from the model without the neural network.