Machine learning and AI for long-term fault prognosis in complex manufacturing systems

Satish T.S. Bukkapatnam, Kahkashan Afrin, Darpit Dave, Soundar R.T. Kumara

Research output: Contribution to journalArticle

1 Citation (Scopus)

Abstract

Recent advances in sensors and other streaming data sources of plant floor automation and information systems open an exciting possibility to predict the risks of faults and breakdowns across a manufacturing plant over much longer time horizons than what is conceivable today. This paper introduces a Manufacturing System-wide Balanced Random Survival Forest (MBRSF), a nonparametric machine learning approach that can fuse complex dynamic dependencies underlying these data streams to provide a long-term prognosis of machine breakdowns. Experimental investigations with a 20 machine automotive manufacturing line suggest that MBRSF reduces prediction errors (Brier scores) by over 90% compared to other methods tested.

Original languageEnglish (US)
Pages (from-to)459-462
Number of pages4
JournalCIRP Annals
Volume68
Issue number1
DOIs
StatePublished - 2019

Fingerprint

Learning systems
Electric fuses
Information systems
Automation
Sensors

All Science Journal Classification (ASJC) codes

  • Mechanical Engineering
  • Industrial and Manufacturing Engineering

Cite this

Bukkapatnam, Satish T.S. ; Afrin, Kahkashan ; Dave, Darpit ; Kumara, Soundar R.T. / Machine learning and AI for long-term fault prognosis in complex manufacturing systems. In: CIRP Annals. 2019 ; Vol. 68, No. 1. pp. 459-462.
@article{9ee920dc08334fe09eeb982b5d8fc84d,
title = "Machine learning and AI for long-term fault prognosis in complex manufacturing systems",
abstract = "Recent advances in sensors and other streaming data sources of plant floor automation and information systems open an exciting possibility to predict the risks of faults and breakdowns across a manufacturing plant over much longer time horizons than what is conceivable today. This paper introduces a Manufacturing System-wide Balanced Random Survival Forest (MBRSF), a nonparametric machine learning approach that can fuse complex dynamic dependencies underlying these data streams to provide a long-term prognosis of machine breakdowns. Experimental investigations with a 20 machine automotive manufacturing line suggest that MBRSF reduces prediction errors (Brier scores) by over 90{\%} compared to other methods tested.",
author = "Bukkapatnam, {Satish T.S.} and Kahkashan Afrin and Darpit Dave and Kumara, {Soundar R.T.}",
year = "2019",
doi = "10.1016/j.cirp.2019.04.104",
language = "English (US)",
volume = "68",
pages = "459--462",
journal = "CIRP Annals - Manufacturing Technology",
issn = "0007-8506",
publisher = "Elsevier USA",
number = "1",

}

Machine learning and AI for long-term fault prognosis in complex manufacturing systems. / Bukkapatnam, Satish T.S.; Afrin, Kahkashan; Dave, Darpit; Kumara, Soundar R.T.

In: CIRP Annals, Vol. 68, No. 1, 2019, p. 459-462.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Machine learning and AI for long-term fault prognosis in complex manufacturing systems

AU - Bukkapatnam, Satish T.S.

AU - Afrin, Kahkashan

AU - Dave, Darpit

AU - Kumara, Soundar R.T.

PY - 2019

Y1 - 2019

N2 - Recent advances in sensors and other streaming data sources of plant floor automation and information systems open an exciting possibility to predict the risks of faults and breakdowns across a manufacturing plant over much longer time horizons than what is conceivable today. This paper introduces a Manufacturing System-wide Balanced Random Survival Forest (MBRSF), a nonparametric machine learning approach that can fuse complex dynamic dependencies underlying these data streams to provide a long-term prognosis of machine breakdowns. Experimental investigations with a 20 machine automotive manufacturing line suggest that MBRSF reduces prediction errors (Brier scores) by over 90% compared to other methods tested.

AB - Recent advances in sensors and other streaming data sources of plant floor automation and information systems open an exciting possibility to predict the risks of faults and breakdowns across a manufacturing plant over much longer time horizons than what is conceivable today. This paper introduces a Manufacturing System-wide Balanced Random Survival Forest (MBRSF), a nonparametric machine learning approach that can fuse complex dynamic dependencies underlying these data streams to provide a long-term prognosis of machine breakdowns. Experimental investigations with a 20 machine automotive manufacturing line suggest that MBRSF reduces prediction errors (Brier scores) by over 90% compared to other methods tested.

UR - http://www.scopus.com/inward/record.url?scp=85065018681&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85065018681&partnerID=8YFLogxK

U2 - 10.1016/j.cirp.2019.04.104

DO - 10.1016/j.cirp.2019.04.104

M3 - Article

AN - SCOPUS:85065018681

VL - 68

SP - 459

EP - 462

JO - CIRP Annals - Manufacturing Technology

JF - CIRP Annals - Manufacturing Technology

SN - 0007-8506

IS - 1

ER -