An Operational-Relationship-Based Diagnostic Model for Evaluating Fuel Injector System Performance in a Marine Diesel Engine Within a Predictive Maintenance Framework

Abstract

The fuel injector system is one of the most critical systems in marine diesel engines, as it directly controls the amount of fuel injected into the combustion chamber, thereby affecting combustion stability and engine performance efficiency. This study aims to develop a diagnostic and operational simulation model based on a condition-based predictive maintenance methodology to analyze injector-system performance and detect operational faults at an early stage before complete equipment failure occurs. The study was conducted using a diagnostic model based on the fuel injector system of an MTU 20V538 TB91 marine diesel engine. The proposed methodology can be generalized to marine diesel engines with similar injection systems. The analysis focuses on the effects of key operational variables, including inlet fuel pressure (OP), fuel pressure factor (FP), and internal injector leakage (LEAK), on the effective injected fuel quantity (DQ_eff). A simulation model was developed using the Arduino UNO platform to measure operational variables and acquire data through the Serial Monitor interface. The data were then analyzed to derive the physical relationships governing injection-system behavior. The simulation results showed a clear positive relationship between inlet pressure and effective fuel quantity, whereas increasing internal injector leakage reduces the injected fuel quantity because part of the fuel is lost through leakage paths. Based on the reference effective fuel quantity, a diagnostic indicator, the fuel quantity deviation (ΔDQ), was defined as the difference between the actual effective injected quantity and the reference value used for diagnosis. This indicator was used to classify the operational condition of the injector system into four main operating zones: Normal, Warning, Critical, and Shutdown. The results indicate that injector-performance degradation occurs gradually rather than abruptly, allowing operational deviations to be detected at an early stage before complete failure is reached. The study also demonstrates that integrating physical-relationship analysis of the injection system with low-cost measurement technologies, such as Arduino-based embedded systems, can provide an effective tool for implementing predictive maintenance in marine engines, improving operational reliability, reducing unexpected failures, and lowering maintenance costs.