Multivariate Time Series Forecasting for Remaining Useful Life of Turbofan Engine Using Deep-Stacked Neural Network and Correlation Analysis

This paper proposes a deep model structure to improve the prediction accuracy of remaining useful life (RUL) of the turbofan engine by using the correlation analysis that reduces the model complexity. The proposed model maximizes the performance by appropriately stacking one dimensional convolutiona...

Full description

Saved in:
Bibliographic Details
Main Authors: Hong, Chang Woo, Lee, Kwangsuk, Ko, Min-Seung, Kim, Jae-Kyeong, Oh, Kyungwon, Hur, Kyeon
Format: Conference Proceeding
Language:English
Subjects:
Correlation
Engines
Logic gates
Maintenance engineering
Neural Network
Neural networks
Predictive models
Prognostics and Health Monitoring
Remaining Useful Life
Time series analysis
Turbofan Engine
Online Access:Request full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:This paper proposes a deep model structure to improve the prediction accuracy of remaining useful life (RUL) of the turbofan engine by using the correlation analysis that reduces the model complexity. The proposed model maximizes the performance by appropriately stacking one dimensional convolutional neural network (1D-CNN), long short-term memory (LSTM), and bidirectional LSTM algorithm. It also includes residual network and dropout technique to improve the learning ability of the proposed model. The result of RUL forecasting using the proposed model is compared with that of various conventional methods, which demonstrates better efficiency yet high performance thanks to excluding the low correlated data.
ISSN:2375-9356
DOI:10.1109/BigComp48618.2020.00-98