APPLYING RANDOM FOREST CLASSIFICATION ALGORITHM TO PREDICT FAUL OF THE MARINE DIESEL ENGINES DUE TO WEAR OF PISTON RINGS

Abstract

The wear of piston rings in diesel engines is a critical factor affecting engine performance and longevity. This study proposes a method utilizing the Random Forest (RF) algorithm to predict piston ring wear based on engine operating parameters, such as combustion chamber pressure, exhaust gas temperature, fuel injection rate, and torsional vibration. The dataset was collected from engines operating under various conditions and analyzed to identify the relationship between operating parameters and wear. The results demonstrate that the RF model achieves high accuracy, effectively supporting predictive maintenance and optimizing diesel engine performance.