Research On Data-driven Ship Power Health Management Methods

Prognostics and Health Management(PHM)technology can help us better understand the operation of equipment,identify faults more effectively,and take effective maintenance measures,thereby achieving effective health assessment,fault diagnosis,and maintenance decision-making.The correct use and maintenance of power equipment are crucial as they not only help ships detect potential hazards,but also provide reliable transportation services to ensure the safe operation of ships.In terms of engineering practice and application,the intelligent technology of ship power systems has been widely applied in fields such as ship manufacturing,ship operation,and maintenance.For example,the intelligent engine control system has been applied in the operation control of diesel engines,which can adjust the engine parameters in real time according to the ship’s operating conditions,improving fuel efficiency and reducing pollution emissions.The intelligent monitoring system can monitor and diagnose ship systems in real-time,timely detect potential faults,and take corresponding measures to improve the reliability and safety of the system.In terms of quantitative situation,the application of intelligent technology for ship power systems is still in its infancy and has not yet formed a unified quantitative indicator.However,through the evaluation and analysis of actual application effects,it can be seen that its positive impact on fuel conservation,pollution reduction,and improving ship operation efficiency can be seen.The thesis focuses on a certain type of 40 DF dual fuel engine and conducts research on its key PHM technologies.The main tasks are as follows:(1)The main composition and structure of the dual fuel engine were introduced,and an optimized Grey Wolf algorithm was used to improve the traditional sensor configuration based on key components,in order to optimize the data source of the data-driven PHM system.(2)A fault diagnosis model of multi-sensor information fusion technology based on 1DCNN is proposed.This model uses one-dimensional convolutional neural network to extract and classify the characteristics of bearing fault vibration data of dual fuel engine.With waveform signals collected by different sensors as input,real-time fault diagnosis and monitoring are realized,so as to better meet the current needs.(3)A health assessment method for dual fuel engines based on state monitoring,constructing a dual fuel engine state assessment index system,and establishing a data-driven health condition assessment fusion model.The health assessment index is used as a constraint to solve the optimal intelligent maintenance decision.Realize early identification of characteristic status and equipment defect status of key components such as combustion chamber related components,Common rail system,crankcase and supercharger,and improve the early warning capability of key equipment and component failures.(4)Exploring how to make maintenance decisions based on the reliability of dual fuel engines.We use the Weber proportional risk model WPHM(WPHM)to evaluate the reliability of dual fuel engines.Based on this,a series of maintenance decision rules have been formulated for dual fuel engines,aiming to minimize maintenance costs,maximize the potential of dual fuel engines,and provide strong support for the effective operation of dual fuel engines.