Monitoring of MGB Lubrication and Cooling System based on Big Data Normality Models and Fuzzy Expert Rules

Abstract

This paper presents an original method for the monitoring of the helicopter main gearbox lubrication and cooling system. The method relies on sensor data of oil pressures and temperature, as well as on domain expert knowledge. It combines oil pressures and temperature normality models built from a significant amount of training data collected from customers' helicopters, with fuzzy expert rules which allow precise root cause identification in case of lubrication or cooling sub-systems anomalies. The results have been validated based on known maintenance findings, showing that the proposed method allows to accurately detect anomalies related to the lubrication and cooling sub-systems as soon as they appear.