Machine learning for wear forecasting of naval assets for condition-based maintenance applications

Andrea Coraddu, Luca Oneto, Alessandro Ghio, Stefano Savio, Massimo Figari, Davide Anguita

Research output: Chapter in Book/Report/Conference proceedingConference contribution book

5 Citations (Scopus)

Abstract

Economic sustainability of running Naval Propulsion Plants is a key element to cope with, and maintenance costs represent a large slice of total operational expenses: last decades' approaches, based on a repairing-replacing methodology, are being trespassed by more effective approaches, relying on effective continuous monitoring of assets wear. In this framework, Condition-Based Maintenance (CBM) is becoming key thanks to the enhancing capabilities of monitoring the propulsion equipment by exploiting heterogeneous sensors: this enables diagnosis and prognosis of the propulsion system's components and of their potential future failures. The success of CBM is based on the capability of developing effective predictive models, for which purpose state-of-the-art Machine Learning (ML) methods must be developed. Nevertheless, testing the performance of ML models for CBM purposes is not straightforward, mostly due to the lack of publicly available datasets for benchmarking purposes: thus, we present in this work a new dataset, that will be freely distributed to the community working on ML models for CBM, generated from an accurate simulator of a naval vessel Gas Turbine propulsion plant. The latter is then used for benchmarking the effectiveness of two state-of-the-art ML techniques in the considered maritime domain.

LanguageEnglish
Title of host publication2015 International Conference on Electrical Systems for Aircraft, Railway, Ship Propulsion and Road Vehicles, ESARS 2015
Place of PublicationPiscataway, NJ
PublisherIEEE
Number of pages5
ISBN (Electronic)9781479974009
DOIs
Publication statusPublished - 4 May 2015
EventInternational Conference on Electrical Systems for Aircraft, Railway, Ship Propulsion and Road Vehicles, ESARS 2015 - Aachen, Germany
Duration: 3 Mar 20155 Mar 2015

Conference

ConferenceInternational Conference on Electrical Systems for Aircraft, Railway, Ship Propulsion and Road Vehicles, ESARS 2015
Abbreviated titleESARS2015
CountryGermany
CityAachen
Period3/03/155/03/15

Fingerprint

Learning systems
Propulsion
Wear of materials
Benchmarking
Naval vessels
Monitoring
Gas turbines
Sustainable development
Simulators
Economics
Sensors
Testing
Costs

Keywords

  • condition-based maintenance
  • machine Learning
  • naval propulsion plant
  • publicly distributed dataset

Cite this

Coraddu, A., Oneto, L., Ghio, A., Savio, S., Figari, M., & Anguita, D. (2015). Machine learning for wear forecasting of naval assets for condition-based maintenance applications. In 2015 International Conference on Electrical Systems for Aircraft, Railway, Ship Propulsion and Road Vehicles, ESARS 2015 [7101499] Piscataway, NJ: IEEE. https://doi.org/10.1109/ESARS.2015.7101499
Coraddu, Andrea ; Oneto, Luca ; Ghio, Alessandro ; Savio, Stefano ; Figari, Massimo ; Anguita, Davide. / Machine learning for wear forecasting of naval assets for condition-based maintenance applications. 2015 International Conference on Electrical Systems for Aircraft, Railway, Ship Propulsion and Road Vehicles, ESARS 2015. Piscataway, NJ : IEEE, 2015.
@inproceedings{25ff586744514230ba4342464a935503,
title = "Machine learning for wear forecasting of naval assets for condition-based maintenance applications",
abstract = "Economic sustainability of running Naval Propulsion Plants is a key element to cope with, and maintenance costs represent a large slice of total operational expenses: last decades' approaches, based on a repairing-replacing methodology, are being trespassed by more effective approaches, relying on effective continuous monitoring of assets wear. In this framework, Condition-Based Maintenance (CBM) is becoming key thanks to the enhancing capabilities of monitoring the propulsion equipment by exploiting heterogeneous sensors: this enables diagnosis and prognosis of the propulsion system's components and of their potential future failures. The success of CBM is based on the capability of developing effective predictive models, for which purpose state-of-the-art Machine Learning (ML) methods must be developed. Nevertheless, testing the performance of ML models for CBM purposes is not straightforward, mostly due to the lack of publicly available datasets for benchmarking purposes: thus, we present in this work a new dataset, that will be freely distributed to the community working on ML models for CBM, generated from an accurate simulator of a naval vessel Gas Turbine propulsion plant. The latter is then used for benchmarking the effectiveness of two state-of-the-art ML techniques in the considered maritime domain.",
keywords = "condition-based maintenance, machine Learning, naval propulsion plant, publicly distributed dataset",
author = "Andrea Coraddu and Luca Oneto and Alessandro Ghio and Stefano Savio and Massimo Figari and Davide Anguita",
year = "2015",
month = "5",
day = "4",
doi = "10.1109/ESARS.2015.7101499",
language = "English",
booktitle = "2015 International Conference on Electrical Systems for Aircraft, Railway, Ship Propulsion and Road Vehicles, ESARS 2015",
publisher = "IEEE",

}

Coraddu, A, Oneto, L, Ghio, A, Savio, S, Figari, M & Anguita, D 2015, Machine learning for wear forecasting of naval assets for condition-based maintenance applications. in 2015 International Conference on Electrical Systems for Aircraft, Railway, Ship Propulsion and Road Vehicles, ESARS 2015., 7101499, IEEE, Piscataway, NJ, International Conference on Electrical Systems for Aircraft, Railway, Ship Propulsion and Road Vehicles, ESARS 2015, Aachen, Germany, 3/03/15. https://doi.org/10.1109/ESARS.2015.7101499

Machine learning for wear forecasting of naval assets for condition-based maintenance applications. / Coraddu, Andrea; Oneto, Luca; Ghio, Alessandro; Savio, Stefano; Figari, Massimo; Anguita, Davide.

2015 International Conference on Electrical Systems for Aircraft, Railway, Ship Propulsion and Road Vehicles, ESARS 2015. Piscataway, NJ : IEEE, 2015. 7101499.

Research output: Chapter in Book/Report/Conference proceedingConference contribution book

TY - GEN

T1 - Machine learning for wear forecasting of naval assets for condition-based maintenance applications

AU - Coraddu, Andrea

AU - Oneto, Luca

AU - Ghio, Alessandro

AU - Savio, Stefano

AU - Figari, Massimo

AU - Anguita, Davide

PY - 2015/5/4

Y1 - 2015/5/4

N2 - Economic sustainability of running Naval Propulsion Plants is a key element to cope with, and maintenance costs represent a large slice of total operational expenses: last decades' approaches, based on a repairing-replacing methodology, are being trespassed by more effective approaches, relying on effective continuous monitoring of assets wear. In this framework, Condition-Based Maintenance (CBM) is becoming key thanks to the enhancing capabilities of monitoring the propulsion equipment by exploiting heterogeneous sensors: this enables diagnosis and prognosis of the propulsion system's components and of their potential future failures. The success of CBM is based on the capability of developing effective predictive models, for which purpose state-of-the-art Machine Learning (ML) methods must be developed. Nevertheless, testing the performance of ML models for CBM purposes is not straightforward, mostly due to the lack of publicly available datasets for benchmarking purposes: thus, we present in this work a new dataset, that will be freely distributed to the community working on ML models for CBM, generated from an accurate simulator of a naval vessel Gas Turbine propulsion plant. The latter is then used for benchmarking the effectiveness of two state-of-the-art ML techniques in the considered maritime domain.

AB - Economic sustainability of running Naval Propulsion Plants is a key element to cope with, and maintenance costs represent a large slice of total operational expenses: last decades' approaches, based on a repairing-replacing methodology, are being trespassed by more effective approaches, relying on effective continuous monitoring of assets wear. In this framework, Condition-Based Maintenance (CBM) is becoming key thanks to the enhancing capabilities of monitoring the propulsion equipment by exploiting heterogeneous sensors: this enables diagnosis and prognosis of the propulsion system's components and of their potential future failures. The success of CBM is based on the capability of developing effective predictive models, for which purpose state-of-the-art Machine Learning (ML) methods must be developed. Nevertheless, testing the performance of ML models for CBM purposes is not straightforward, mostly due to the lack of publicly available datasets for benchmarking purposes: thus, we present in this work a new dataset, that will be freely distributed to the community working on ML models for CBM, generated from an accurate simulator of a naval vessel Gas Turbine propulsion plant. The latter is then used for benchmarking the effectiveness of two state-of-the-art ML techniques in the considered maritime domain.

KW - condition-based maintenance

KW - machine Learning

KW - naval propulsion plant

KW - publicly distributed dataset

UR - http://www.scopus.com/inward/record.url?scp=84946196162&partnerID=8YFLogxK

U2 - 10.1109/ESARS.2015.7101499

DO - 10.1109/ESARS.2015.7101499

M3 - Conference contribution book

BT - 2015 International Conference on Electrical Systems for Aircraft, Railway, Ship Propulsion and Road Vehicles, ESARS 2015

PB - IEEE

CY - Piscataway, NJ

ER -

Coraddu A, Oneto L, Ghio A, Savio S, Figari M, Anguita D. Machine learning for wear forecasting of naval assets for condition-based maintenance applications. In 2015 International Conference on Electrical Systems for Aircraft, Railway, Ship Propulsion and Road Vehicles, ESARS 2015. Piscataway, NJ: IEEE. 2015. 7101499 https://doi.org/10.1109/ESARS.2015.7101499