Hidden Markov model-based predictive maintenance in semiconductor manufacturing: A genetic algorithm approach

Jakob Kinghorst; Omid Geramifard; Ming Luo; Hian Leng Chan; Khoo Yong; Jens Folmer; Minjie Zou; Birgit Vogel-Heuser

doi:10.1109/COASE.2017.8256274

Hidden Markov model-based predictive maintenance in semiconductor manufacturing: A genetic algorithm approach

Jakob Kinghorst
, Omid Geramifard
, Ming Luo
, Hian Leng Chan
, Khoo Yong
, Jens Folmer
, Minjie Zou
, Birgit Vogel-Heuser

Chair of Automation and Information Systems

Technical University of Munich
A STAR Singapore Institute of Manufacturing Technology
GLOBALFOUNDRIES Singapore Pte. Ltd

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

11 Scopus citations

Abstract

The accuracy of data-mining based predictive maintenance often relies on extensive process and machine knowledge to enable appropriate feature selection and data preprocessing. Measurement data obtained may be asynchronous and result in inaccurate features, affecting the accuracy of maintenance prediction. To overcome this drawback, this paper introduces an approach to automatically select a feature subset through a genetic algorithm. The full feature set is created based on different sliding windows characterizing different time shifts on adopted statistical metrics of the measurement data. The fitness function of the genetic algorithm is then developed based on the preliminary fitting of a hidden Markov model (HMM) on the selected subset of features and assumed machines' condition in the training data. Ultimately the fittest subset of features is used to enable HMM-based predictive maintenance. The proposed approach is evaluated using data from semi-conductor wafer production equipment, recorded over a period of one year.

Original language	English
Title of host publication	2017 13th IEEE Conference on Automation Science and Engineering, CASE 2017
Publisher	IEEE Computer Society
Pages	1260-1267
Number of pages	8
ISBN (Electronic)	9781509067800
DOIs	https://doi.org/10.1109/COASE.2017.8256274
State	Published - 1 Jul 2017
Event	13th IEEE Conference on Automation Science and Engineering, CASE 2017 - Xi'an, China Duration: 20 Aug 2017 → 23 Aug 2017

Publication series

Name	IEEE International Conference on Automation Science and Engineering
Volume	2017-August
ISSN (Print)	2161-8070
ISSN (Electronic)	2161-8089

Conference

Conference	13th IEEE Conference on Automation Science and Engineering, CASE 2017
Country/Territory	China
City	Xi'an
Period	20/08/17 → 23/08/17

UN SDGs

This output contributes to the following UN Sustainable Development Goals (SDGs)

SDG 9 Industry, Innovation, and Infrastructure

Keywords

Condition monitoring
feature extraction
genetic algorithms
hidden Markov models
predictive maintenance
semiconductor device reliability

Access to Document

10.1109/COASE.2017.8256274

Cite this

Kinghorst, J., Geramifard, O., Luo, M., Chan, H. L., Yong, K., Folmer, J., Zou, M., & Vogel-Heuser, B. (2017). Hidden Markov model-based predictive maintenance in semiconductor manufacturing: A genetic algorithm approach. In 2017 13th IEEE Conference on Automation Science and Engineering, CASE 2017 (pp. 1260-1267). (IEEE International Conference on Automation Science and Engineering; Vol. 2017-August). IEEE Computer Society. https://doi.org/10.1109/COASE.2017.8256274

@inproceedings{1ddbd6d8e67743809de1bfdcf4d395ec,

title = "Hidden Markov model-based predictive maintenance in semiconductor manufacturing: A genetic algorithm approach",

abstract = "The accuracy of data-mining based predictive maintenance often relies on extensive process and machine knowledge to enable appropriate feature selection and data preprocessing. Measurement data obtained may be asynchronous and result in inaccurate features, affecting the accuracy of maintenance prediction. To overcome this drawback, this paper introduces an approach to automatically select a feature subset through a genetic algorithm. The full feature set is created based on different sliding windows characterizing different time shifts on adopted statistical metrics of the measurement data. The fitness function of the genetic algorithm is then developed based on the preliminary fitting of a hidden Markov model (HMM) on the selected subset of features and assumed machines' condition in the training data. Ultimately the fittest subset of features is used to enable HMM-based predictive maintenance. The proposed approach is evaluated using data from semi-conductor wafer production equipment, recorded over a period of one year.",

keywords = "Condition monitoring, feature extraction, genetic algorithms, hidden Markov models, predictive maintenance, semiconductor device reliability",

author = "Jakob Kinghorst and Omid Geramifard and Ming Luo and Chan, \{Hian Leng\} and Khoo Yong and Jens Folmer and Minjie Zou and Birgit Vogel-Heuser",

note = "Publisher Copyright: {\textcopyright} 2017 IEEE.; 13th IEEE Conference on Automation Science and Engineering, CASE 2017 ; Conference date: 20-08-2017 Through 23-08-2017",

year = "2017",

month = jul,

day = "1",

doi = "10.1109/COASE.2017.8256274",

language = "English",

series = "IEEE International Conference on Automation Science and Engineering",

publisher = "IEEE Computer Society",

pages = "1260--1267",

booktitle = "2017 13th IEEE Conference on Automation Science and Engineering, CASE 2017",

}

Kinghorst, J, Geramifard, O, Luo, M, Chan, HL, Yong, K, Folmer, J, Zou, M & Vogel-Heuser, B 2017, Hidden Markov model-based predictive maintenance in semiconductor manufacturing: A genetic algorithm approach. in 2017 13th IEEE Conference on Automation Science and Engineering, CASE 2017. IEEE International Conference on Automation Science and Engineering, vol. 2017-August, IEEE Computer Society, pp. 1260-1267, 13th IEEE Conference on Automation Science and Engineering, CASE 2017, Xi'an, China, 20/08/17. https://doi.org/10.1109/COASE.2017.8256274

Hidden Markov model-based predictive maintenance in semiconductor manufacturing: A genetic algorithm approach. / Kinghorst, Jakob; Geramifard, Omid; Luo, Ming et al.
2017 13th IEEE Conference on Automation Science and Engineering, CASE 2017. IEEE Computer Society, 2017. p. 1260-1267 (IEEE International Conference on Automation Science and Engineering; Vol. 2017-August).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

TY - GEN

T1 - Hidden Markov model-based predictive maintenance in semiconductor manufacturing

T2 - 13th IEEE Conference on Automation Science and Engineering, CASE 2017

AU - Kinghorst, Jakob

AU - Geramifard, Omid

AU - Luo, Ming

AU - Chan, Hian Leng

AU - Yong, Khoo

AU - Folmer, Jens

AU - Zou, Minjie

AU - Vogel-Heuser, Birgit

PY - 2017/7/1

Y1 - 2017/7/1

N2 - The accuracy of data-mining based predictive maintenance often relies on extensive process and machine knowledge to enable appropriate feature selection and data preprocessing. Measurement data obtained may be asynchronous and result in inaccurate features, affecting the accuracy of maintenance prediction. To overcome this drawback, this paper introduces an approach to automatically select a feature subset through a genetic algorithm. The full feature set is created based on different sliding windows characterizing different time shifts on adopted statistical metrics of the measurement data. The fitness function of the genetic algorithm is then developed based on the preliminary fitting of a hidden Markov model (HMM) on the selected subset of features and assumed machines' condition in the training data. Ultimately the fittest subset of features is used to enable HMM-based predictive maintenance. The proposed approach is evaluated using data from semi-conductor wafer production equipment, recorded over a period of one year.

AB - The accuracy of data-mining based predictive maintenance often relies on extensive process and machine knowledge to enable appropriate feature selection and data preprocessing. Measurement data obtained may be asynchronous and result in inaccurate features, affecting the accuracy of maintenance prediction. To overcome this drawback, this paper introduces an approach to automatically select a feature subset through a genetic algorithm. The full feature set is created based on different sliding windows characterizing different time shifts on adopted statistical metrics of the measurement data. The fitness function of the genetic algorithm is then developed based on the preliminary fitting of a hidden Markov model (HMM) on the selected subset of features and assumed machines' condition in the training data. Ultimately the fittest subset of features is used to enable HMM-based predictive maintenance. The proposed approach is evaluated using data from semi-conductor wafer production equipment, recorded over a period of one year.

KW - Condition monitoring

KW - feature extraction

KW - genetic algorithms

KW - hidden Markov models

KW - predictive maintenance

KW - semiconductor device reliability

UR - https://www.scopus.com/pages/publications/85044946205

U2 - 10.1109/COASE.2017.8256274

DO - 10.1109/COASE.2017.8256274

M3 - Conference contribution

AN - SCOPUS:85044946205

T3 - IEEE International Conference on Automation Science and Engineering

SP - 1260

EP - 1267

BT - 2017 13th IEEE Conference on Automation Science and Engineering, CASE 2017

PB - IEEE Computer Society

Y2 - 20 August 2017 through 23 August 2017

ER -

Kinghorst J, Geramifard O, Luo M, Chan HL, Yong K, Folmer J et al. Hidden Markov model-based predictive maintenance in semiconductor manufacturing: A genetic algorithm approach. In 2017 13th IEEE Conference on Automation Science and Engineering, CASE 2017. IEEE Computer Society. 2017. p. 1260-1267. (IEEE International Conference on Automation Science and Engineering). doi: 10.1109/COASE.2017.8256274

Hidden Markov model-based predictive maintenance in semiconductor manufacturing: A genetic algorithm approach

Abstract

Publication series

Conference

UN SDGs

Keywords

Access to Document

Other files and links

Fingerprint

Cite this