Using LSTM networks to predict engine condition on large scale data processing framework
Authors:
- 1 April 2017
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 281-285
Abstract
As the Internet of Things technology is developing rapidly, companies have an ability to observe the health of engine components and constructed systems through collecting signals from sensors. According to output of IoT sensors, companies can build systems to predict the conditions of components. Practically the components are required to be maintained or replaced before the end of life in performing their assigned task. Predicting the life condition of a component is so crucial for industries that have intent to grow in a fast paced technological environment. Recent studies on predictive maintenance help industries to create an alert before the components are corrupted. Thanks to prediction of component failures, companies have a chance to sustain their operations efficiently while reducing their maintenance cost by repairing components in advance. Since maintenance affects production capacity and the service quality directly, optimized maintenance is the key factor for organizations to have more revenue and stay competitive in developing industrialized world. With the aid of well-designed prediction system for understanding current situation of an engine, components could be taken out of active service before malfunction occurs. With the help of inspection, effective maintenance extends component life, improves equipment availability and keeps components in a proper condition while reducing costs. Real time data collected from sensors is a great source to model component deteriorations. Markov Chain models, Survival Analysis, Optimization algorithms and several machine learning approaches have been implemented in order to model predictive maintenance. In this paper Long Short Term Memory (LSTM) networks has been performed to predict the current situation of an engine. LSTM model deals with a sequential input data. Training process of LSTM networks has been performed on large-scale data processing engine with high performance. Since huge amount of data is flowing into the predictive model, Apache Spark which is offering a distributed clustering environment has been used. The output of the LSTM network is deciding the current life condition of components and offering the alerts for components before the end of their life. The proposed model also trained and tested on an open source data that is about an engine degradation simulation provided by the Prognostics CoE at NASA Ames.Keywords
This publication has 8 references indexed in Scilit:
- Meta-cognitive Regression Neural Network for function approximation: Application to Remaining Useful Life estimation2016
- Remaining useful life prediction of individual units subject to hard failureIIE Transactions, 2014
- Time-frequency complexity based remaining useful life (RUL) estimation for bearing faults2013
- Remaining useful life estimation for systems with non-trendability behaviour2012
- Predicting remaining useful life of rotating machinery based artificial neural networkComputers & Mathematics with Applications, 2010
- Gas storage valuation: Comparison of recombining trees and Least Squares Monte-Carlo simulation2008
- A review on machinery diagnostics and prognostics implementing condition-based maintenanceMechanical Systems and Signal Processing, 2006
- Using literature-based discovery to identify disease candidate genesInternational Journal of Medical Informatics, 2005