System Modeling of Predictive Maintenance for Engine Health Monitoring on Ship Auxiliary Engines Using Vibration Variables

Abstract

This research presents a dynamic system modelling approach for predictive maintenance of ship auxiliary engines using vibration variables. The model integrates key mechanical components—piston, crankshaft, camshaft, valve train, and timing gear—based on the specifications of a Yanmar TF85 diesel engine. Each subsystem is modelled using a multi-degree-of-freedom (MDOF) state-space framework to represent vibrational and structural dynamics. Simulations are carried out in MATLAB/Simulink under various engine operating conditions, including normal operation at high, medium, and low RPMs, as well as fault scenarios such as damping degradation in the piston and crankshaft. A fuzzy logic system is employed to interpret the vibration data and determine the impact level for each condition. The results indicate that under normal conditions, the engine maintains stable vibration levels, while faults lead to significant increases in velocity RMS values and impact severity. Disturbances in the piston result in dominant amplitude changes, while crankshaft faults affect the frequency propagation throughout the system. These findings confirm that the proposed model can effectively detect early mechanical deviations and support the implementation of predictive maintenance strategies for marine diesel engines.