Coal-fired boiler fault prediction using artificial neural networks

Nong Nurnie Mohd Nistah, King Hann Lim, Lenin Gopal, Firas Basim Ismail

Research output: Contribution to journalArticle

Abstract

Boiler fault is a critical issue in a coal-fired power plant due to its high temperature and high pressure characteristics. The complexity of boiler design increases the difficulty of fault investigation in a quick moment to avoid long duration shut-down. In this paper, a boiler fault prediction model is proposed using artificial neural network. The key influential parameters analysis is carried out to identify its correlation with the performance of the boiler. The prediction model is developed to achieve the least misclassification rate and mean squared error. Artificial neural network is trained using a set of boiler operational parameters. Subsequenlty, the trained model is used to validate its prediction accuracy against actual fault value from a collected real plant data. With reference to the study and test results, two set of initial weights have been tested to verify the repeatability of the correct prediction. The results show that the artificial neural network implemented is able to provide an average of above 92% prediction rate of accuracy.

Original languageEnglish
Pages (from-to)2486-2493
Number of pages8
JournalInternational Journal of Electrical and Computer Engineering
Volume8
Issue number4
DOIs
Publication statusPublished - 01 Aug 2018

Fingerprint

Coal fired boilers
Boilers
Neural networks
Power plants
Coal

All Science Journal Classification (ASJC) codes

  • Computer Science(all)
  • Electrical and Electronic Engineering

Cite this

Nistah, Nong Nurnie Mohd ; Lim, King Hann ; Gopal, Lenin ; Ismail, Firas Basim. / Coal-fired boiler fault prediction using artificial neural networks. In: International Journal of Electrical and Computer Engineering. 2018 ; Vol. 8, No. 4. pp. 2486-2493.
@article{a3f5c7b9af9f45c9a65ebbf97c44cc5c,
title = "Coal-fired boiler fault prediction using artificial neural networks",
abstract = "Boiler fault is a critical issue in a coal-fired power plant due to its high temperature and high pressure characteristics. The complexity of boiler design increases the difficulty of fault investigation in a quick moment to avoid long duration shut-down. In this paper, a boiler fault prediction model is proposed using artificial neural network. The key influential parameters analysis is carried out to identify its correlation with the performance of the boiler. The prediction model is developed to achieve the least misclassification rate and mean squared error. Artificial neural network is trained using a set of boiler operational parameters. Subsequenlty, the trained model is used to validate its prediction accuracy against actual fault value from a collected real plant data. With reference to the study and test results, two set of initial weights have been tested to verify the repeatability of the correct prediction. The results show that the artificial neural network implemented is able to provide an average of above 92{\%} prediction rate of accuracy.",
author = "Nistah, {Nong Nurnie Mohd} and Lim, {King Hann} and Lenin Gopal and Ismail, {Firas Basim}",
year = "2018",
month = "8",
day = "1",
doi = "10.11591/ijece.v8i4.pp2486-2493",
language = "English",
volume = "8",
pages = "2486--2493",
journal = "International Journal of Electrical and Computer Engineering",
issn = "2088-8708",
publisher = "Institute of Advanced Engineering and Science (IAES)",
number = "4",

}

Coal-fired boiler fault prediction using artificial neural networks. / Nistah, Nong Nurnie Mohd; Lim, King Hann; Gopal, Lenin; Ismail, Firas Basim.

In: International Journal of Electrical and Computer Engineering, Vol. 8, No. 4, 01.08.2018, p. 2486-2493.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Coal-fired boiler fault prediction using artificial neural networks

AU - Nistah, Nong Nurnie Mohd

AU - Lim, King Hann

AU - Gopal, Lenin

AU - Ismail, Firas Basim

PY - 2018/8/1

Y1 - 2018/8/1

N2 - Boiler fault is a critical issue in a coal-fired power plant due to its high temperature and high pressure characteristics. The complexity of boiler design increases the difficulty of fault investigation in a quick moment to avoid long duration shut-down. In this paper, a boiler fault prediction model is proposed using artificial neural network. The key influential parameters analysis is carried out to identify its correlation with the performance of the boiler. The prediction model is developed to achieve the least misclassification rate and mean squared error. Artificial neural network is trained using a set of boiler operational parameters. Subsequenlty, the trained model is used to validate its prediction accuracy against actual fault value from a collected real plant data. With reference to the study and test results, two set of initial weights have been tested to verify the repeatability of the correct prediction. The results show that the artificial neural network implemented is able to provide an average of above 92% prediction rate of accuracy.

AB - Boiler fault is a critical issue in a coal-fired power plant due to its high temperature and high pressure characteristics. The complexity of boiler design increases the difficulty of fault investigation in a quick moment to avoid long duration shut-down. In this paper, a boiler fault prediction model is proposed using artificial neural network. The key influential parameters analysis is carried out to identify its correlation with the performance of the boiler. The prediction model is developed to achieve the least misclassification rate and mean squared error. Artificial neural network is trained using a set of boiler operational parameters. Subsequenlty, the trained model is used to validate its prediction accuracy against actual fault value from a collected real plant data. With reference to the study and test results, two set of initial weights have been tested to verify the repeatability of the correct prediction. The results show that the artificial neural network implemented is able to provide an average of above 92% prediction rate of accuracy.

UR - http://www.scopus.com/inward/record.url?scp=85049557496&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85049557496&partnerID=8YFLogxK

U2 - 10.11591/ijece.v8i4.pp2486-2493

DO - 10.11591/ijece.v8i4.pp2486-2493

M3 - Article

VL - 8

SP - 2486

EP - 2493

JO - International Journal of Electrical and Computer Engineering

JF - International Journal of Electrical and Computer Engineering

SN - 2088-8708

IS - 4

ER -