Analysis of a turbofan engine via predictive maintenance approach

The predictive maintenance is a maintenance method that is performed immediately before the malfunction and sufficient amount of maintenance as a result of the evaluation of physical data collected from the sensors in the system through a software. Thanks to the referenced predictive maintenance approach in order to increase efficiency and shorten the unplanned stop times caused by unexpected malfunctions, the losses in production and the duration of the inactivity of the machine decreases. This method prevents the implementation of unnecessary maintenance costs by implementing only in cases where necessary because it is a condition-based method not time-based method. This study aims to estimate the remaining useful life of aircraft engines until they fail, in order to ensure that aircraft engines are taken into maintenance before the failure occurs, to prevent great financial losses and especially loss of life caused by maintenance / repair / replacement of parts due to failures. In this study, the competition data organized in the International Prognostic and Health Management Conference in 2008 were used. Using the data from the turbofan engines in the training set, machine learning models are trained, using these models, the remaining useful life of the motors in the validation set were estimated and the performance of the predictive models were compared.