A stacked deep convolutional neural network to predict the remaining useful life of a turbofan engine

TL;DR

This paper introduces a novel stacked deep convolutional neural network approach for predicting the remaining useful life of aircraft engines, leveraging data-driven techniques and Bayesian optimization for model selection.

Contribution

It proposes a two-level DCNN architecture for RUL prediction and demonstrates its effectiveness through ranking in a major data challenge.

Findings

Achieved third place in the 2021 PHM Conference Data Challenge.

Developed a two-level DCNN model for RUL estimation.

Utilized Bayesian optimization for model selection.

Abstract

This paper presents the data-driven techniques and methodologies used to predict the remaining useful life (RUL) of a fleet of aircraft engines that can suffer failures of diverse nature. The solution presented is based on two Deep Convolutional Neural Networks (DCNN) stacked in two levels. The first DCNN is used to extract a low-dimensional feature vector using the normalized raw data as input. The second DCNN ingests a list of vectors taken from the former DCNN and estimates the RUL. Model selection was carried out by means of Bayesian optimization using a repeated random subsampling validation approach. The proposed methodology was ranked in the third place of the 2021 PHM Conference Data Challenge.