Research On Ship’s Cooling Water System State Parameter Forecast Based On Time Series

Under the background of big data,intelligent ship has become an inevitable trend for the development of shipping industry. Keeping the stability of powerplant is one of the most important parts of ship’s safe operation.Diesel engine releases a great amount of heat during work, The economy, dynamic performance and reliability of powerplant will sharply decrease if the heat can not dissipate in time. Therefore,the proper function of ship cooling water system is crucial to normal operation of the vessels. State forecasting of cooling water system is the nessential condition of realizing shipping intellectualization and significative for condition based maintenance.Firstly,the paper introduces the central cooling water system of "YuKun", elaborating the combination of three subsystems and the relevant state parameters.After analyzing the mode and influence of failure of main sea water pump for example, we come to the conclusion that it is practicable to forecast failures of sea water pump’s components through the changes of relevant state parameters.Then, the paper also introduces several methods of state forecasting are and compares with neural network algorithm and gray theory,the result shows that time series method is suitable for this paper’s research with the reason that the data collected in this paper is stable and arranged in chronological order.Meanwhile, steps of establishing ARMA time series model are introduced.And then this paper takes the historical parameters of outlet temperature of low temperature fresh water in central cooler and the outlet pressure of cylinder jacket cooling pump as inputs in order to obtain the future changing trend of these two state parameters, verifying the accuracy of model through the comparison of forecasting value and measured value.Finally, combined with the content of the front, forecasting the changing trend of outlet pressure of sea water pump,the prediction of residual life of sea water pump can be accomplished.To sum up,This paper establishes the ARMA time series model and verifies the accuracy of model. With the development of sensor technology, the collection of state parameters is becoming more convenient. The application of time series method can achieve that using historical data forecasts future data.