회귀분석을 이용한 선박용 디젤 기관 고장진단에 관한 연구

Abstract

The engine machinery, especially main engine's fault and unplanned stopping cause operational loss and decline the shipping company's reliability in merchant fleet. Therefore, the shipping company invest the facilities for getting qualification for UMA(Unattended Machinery Area) vessel and ship's crew maintain duty system always in order to prevent and minimize loss. The AMS(Alarm Monitoring System) is one of the facility that can grasp operating condition in ECR(Engine Control Room) and getting qualification for UMA Vessel. Sometimes the AMS can combine control function. When the AMS detect abnormal condition, it makes visual-aural alerts and stopping also. Visual-aural alerts are designed for distinguishing abnormal condition and type easily. The method of distinguishing abnormal condition is designated monitoring knowledge of expert without additional sensors.

Also I am supposed to demonstrate the effectiveness. The subject and data of study are damaged piston ring and cylinder liner of main engine data that is collected from may 2015 year to february 2016 that was recorded 1 minute interval. The main engine with low speed(Crosshead type) is made by "W" company.

Contents of collected data are 133 parameter including fuel pressure filter outlet, main bearing pressure, thrust pad bearing lubricating oil pressure, crosshead lubricating oil pressure, turbocharger speed, scavenge air receiver pressure, each cylinder exhaust gas temperature, engine speed, etc. The result of time series analysis is earlier at least 3 day or at most 1 week than reporting time by vessel's crew. These method of fault diagnosis have to need many case study and accumulated data. Also optimized position of sensor is required for confirming exact engine condition and necessary sensor is verified for fault diagnosis.