Development Of Ship Engine Room Equipment Health Management System

As the core of a ship,the health status of its internal equipment plays a crucial role in the normal operation of the entire ship.The mechanical and electrical equipment inside the cabin is very complex,and the harsh working environment and short-term high-power operating conditions can easily cause irreversible damage to the equipment.Vibration is the earliest manifestation of equipment failure,and some minor mechanical defects or damage can cause abnormal vibration of engine room equipment.Currently,most intelligent ship engine rooms only meet the standard requirements of Auto-0 for vibration detection,achieving simple control or status monitoring.The exploration and utilization of hardware resources and software facilities are not yet mature.Therefore,developing a health management system that can automatically identify and alarm engine room equipment faults is of great significance for the safe operation of ships.This article will be based on practical engineering applications,and the research content will be divided into the following four aspects:1.Starting from the principle of ship engine room equipment PHM,it is determined that the system should be implemented from two levels: vibration signal acquisition and processing algorithms;Analyze the current international and national standards related to vibration measurement and health assessment,simulate and compare the velocity and displacement errors of time-frequency domain integration in MATLAB software,and determine the calculation process of evaluation indicators;Elaborate the algorithm principles of BP neural network and D-S evidence theory,laying the foundation for decision-making level fusion fault diagnosis of subsequent systems.2.Construct a hardware unit for vibration monitoring system based on MEMS sensors,and determine a hardware architecture for separating signal acquisition and data processing.The collection and processing modules all use STM32 series microprocessors,fully utilizing the hardware resources of the chip for distributed development of the system,completing the hardware circuit schematic and PCB board design of the minimum system,main control unit,communication module,and storage unit of the module.3.Design the software part of the system using Keil and Visual Studio as platforms.Firstly,transplant the UCOS-II embedded operating system,and write initialization,peripheral driver,and data transmission programs for the signal acquisition unit and vibration processing module.Secondly,develop upper computer software based on MVVM architecture,using C # language to implement ETH network interface instructions and data transmission,waveform display,storage,and historical query functions.Finally,algorithms such as My SQL data reading,signal analysis,and feature extraction were completed in MATLAB,and a diagnostic model combining BP neural network with D-S evidence theory was built to improve the joint debugging mechanism of the health management system.4.Conduct joint debugging of the software and hardware functions of the vibration monitoring system.In order to further test the reliability of the health assessment algorithm and fault diagnosis model,a centrifugal pump experimental platform was built.Experiments were conducted on the loosening of motor anchor bolts and different degrees of cavitation of the centrifugal pump,completing the validation of the health assessment algorithm.After extracting time-domain feature parameters and calculating wavelet packet frequency band energy input into the BP neural network,the fault recognition errors of different parameter inputs and fusion of feature level and decision level were compared.