基于混合预测模型的船舶海水冷却系统状态参数预测
|
摘要
为了提高船舶海水冷却系统状态参数的预测精度,提出了自回归滑动平均线性子模型和BP神经网络非线性子模型相结合的海水冷却系统状态参数混合预测模型。介绍混合预测模型的建模过程以及预测流程,选取"育鲲轮"海水冷却系统5天状态参数作为训练样本输入到混合预测模型,对第6天状态参数进行预测,将混合预测模型的预测值分别与自回归滑动平均预测模型预测值、实际观测值对比并运用平均绝对百分比误差法进行验证,结果表明搭建的混合预测模型可以提高海水冷却系统状态参数的预测精度,具有良好的预测能力。
In order to improve the prediction accuracy of the ship seawater cooling system state parameters, the seawater cooling system state parameters hybrid prediction model combining auto regression moving average linear model and BP neural network nonlinear model. The modeling process and prediction process of hybrid prediction model are introduced. Selecting 5 day's state parameters of MV "YUKUN" ship seawater cooling system as training sample and inputting the training sample into the hybrid prediction model, getting the sixth day's prediction data. The predicted values of the hybrid prediction model are compared with the predicted values of the ARMA liner model and the actual observation value, and using the MAPE to verify the prediction model, the result shows the hybrid prediction model can improve the prediction accuracy of the seawater cooling system and have good prediction ability.
In order to improve the prediction accuracy of the ship seawater cooling system state parameters, the seawater cooling system state parameters hybrid prediction model combining auto regression moving average linear model and BP neural network nonlinear model. The modeling process and prediction process of hybrid prediction model are introduced. Selecting 5 day's state parameters of MV "YUKUN" ship seawater cooling system as training sample and inputting the training sample into the hybrid prediction model, getting the sixth day's prediction data. The predicted values of the hybrid prediction model are compared with the predicted values of the ARMA liner model and the actual observation value, and using the MAPE to verify the prediction model, the result shows the hybrid prediction model can improve the prediction accuracy of the seawater cooling system and have good prediction ability.
引文
[1]周振星.船舶中央冷却系统节能优化设计研究[J].船舶与海洋工程,2012(4):37-40.
[2]王在忠,张跃文.基于MCGS组态软件的船舶冷却水系统故障诊断设计[J].中国水运(下半月),2014,14(7):144-146.
[3]郭克余.船舶海水冷却系统状态感知技术研究[D].大连:大连海事大学,2017.
[4]王新全,孙培廷,邹永久,等.基于GA-BP模型的船舶柴油机排气温度趋势预测[J].大连海事大学学报,2015,41(3):73-76.
[5]茹斌,张天伟,王宇欣.基于小波去噪及ARMA模型的故障率预测方法研究[J].测控技术,2014,33(10):43-46+50.
[6]王江,袁宝远,任丽芳,等.新型智能算法在基坑变形监测中的应用[J].中外公路,2007(01):23-26.
[7]温国强,文妍,谭继文.基于EMD与神经网络的滚珠丝杠故障诊断[J].机床与液压,2013,41(21):164-167.
[8]武美先,张学良,温淑花,等.BP神经网络的双学习率自适应学习算法[J].现代制造工程,2005(10):29-32.
[9]杨皓,黄东胜.AR模型应用于振动信号趋势预测的研究[J].南京工程学院学报(自然科学版),2007(2):45-49.
[10]吕军,李星.自适应网络流量线性预测算法及应用[J].计算机应用研究,2005(12):243-246.
[2]王在忠,张跃文.基于MCGS组态软件的船舶冷却水系统故障诊断设计[J].中国水运(下半月),2014,14(7):144-146.
[3]郭克余.船舶海水冷却系统状态感知技术研究[D].大连:大连海事大学,2017.
[4]王新全,孙培廷,邹永久,等.基于GA-BP模型的船舶柴油机排气温度趋势预测[J].大连海事大学学报,2015,41(3):73-76.
[5]茹斌,张天伟,王宇欣.基于小波去噪及ARMA模型的故障率预测方法研究[J].测控技术,2014,33(10):43-46+50.
[6]王江,袁宝远,任丽芳,等.新型智能算法在基坑变形监测中的应用[J].中外公路,2007(01):23-26.
[7]温国强,文妍,谭继文.基于EMD与神经网络的滚珠丝杠故障诊断[J].机床与液压,2013,41(21):164-167.
[8]武美先,张学良,温淑花,等.BP神经网络的双学习率自适应学习算法[J].现代制造工程,2005(10):29-32.
[9]杨皓,黄东胜.AR模型应用于振动信号趋势预测的研究[J].南京工程学院学报(自然科学版),2007(2):45-49.
[10]吕军,李星.自适应网络流量线性预测算法及应用[J].计算机应用研究,2005(12):243-246.