Remaining Useful Life Estimation on Turbofan Engines Using Joint Autoencoder-Regression

Abstract

Maintenance costs of industrial machinery often exceed the initial investment cost. Predictive maintenance, which analyzes the health of the machine and suggests maintenance planning, is one of the strategies implemented to reduce maintenance costs. The health status and life estimation of the machines are the most researched topics in this context. In this paper we analyzed the NASA Turbofan Engine Degredation Dataset for useful life predictions. We used joint autocoder-regression models based on deep learning architecture to estimate the useful life of the turbofan jet engines. We compared the results to current studies on the same dataset. Although our results are on the FD001 subset are not very successful, we expect the architecture to be useful for other more complex subsets.