基于ARMA模型的船舶海水冷却系统参数预测

Ship Seawater Cooling System Parameter Prediction Based on ARMA Model

摘要:

船舶海水冷却系统与船外海水直接接触,工作环境较为恶劣,而基于小波理论、灰色理论等参数预测方法受环境影响较大,为了实现对船舶海水冷却系统状态参数的准确预测,提出了根据平稳时间序列建立自回归移动平均模型(ARMA)的方法；介绍了ARMA模型原理及建模过程；选取“育鲲轮”海水冷却系统6天的状态参数作为训练样本,输入到ARMA预测模型中进行训练；在MATLAB环境下,获得预测数据；运用平均绝对百分比误差对预测模型的准确性进行验证并对误差进行分析,结果表明所建立的船舶海水冷却系统状态参数预测模型具有良好的预测能力,能有效地反应未来一段时间海水冷却系统的工作状态的变化,提示系统是否存在异常,为早期故障诊断提供有效手段,进而为船舶的稳定运营提供了条件。

Abstract:

The ship seawater cooling system contacts with seawater, so the working environments and conditions are bad, and some parameter prediction methods are greatly influenced by the environment such as wavelet theory, gray theory and so on. In order to realize the state parameters prediction of ship seawater cooling system correctly, ARMA prediction model method of stationary time series is proposed. Then the principle and modeling process of ARMA model is introduced, selecting 6 days’ state parameters of MV “YUKUN” ship seawater cooling system as training sample, and inputting the training sample into the ARMA model, getting the prediction data by MATLAB. Then using the MAPE to verify the prediction model and analyzing the error, the result shows the model has good prediction ability.And the model can effectively response the changes of seawater cooling system’s working state in the period ahead and suggest weather the system is abnormal, and provide effective ways for the early fault diagnosis. Furthermore, the model provides advantages for the stable operation of the ships.