國立虎尾科技大學 |

Combustion Kinetics Study and Machine Learning for the Simulation and Control of Industry Flares.

Record Type:	Language materials, manuscript : Monograph/item
Title/Author:	Combustion Kinetics Study and Machine Learning for the Simulation and Control of Industry Flares./
Author:	Wang, Anan.
Description:	1 online resource (95 pages)
Notes:	Source: Dissertations Abstracts International, Volume: 81-12, Section: B.
Contained By:	Dissertations Abstracts International81-12B.
Subject:	Chemical engineering. -
Online resource:	click for full text (PQDT)
ISBN:	9798645479169

Combustion Kinetics Study and Machine Learning for the Simulation and Control of Industry Flares.
Wang, Anan.

Combustion Kinetics Study and Machine Learning for the Simulation and Control of Industry Flares. - 1 online resource (95 pages)

Source: Dissertations Abstracts International, Volume: 81-12, Section: B.

Thesis (Ph.D.)--Lamar University - Beaumont, 2019.

Includes bibliographical references

Industrial flare is used to burn off the unwanted gas during production. Global gas flaring increased by 3% in 2018 to 145 billion cubic meters (bcm) based on satellite data of The World Bank. If not combusted completely, intermediate products or incomplete combustion products are formed. The EPA Refinery Sector Rule (RSR) emphasize smokeless flaring with combustion efficiency (CE) >= 96.5% and destruction and remove efficiency (DRE) >= 98% for all types of flares. RSR also specified operating conditions to achieve those objectives including combustion zone net heating value (NHVcz) >= 270 BTU/scf, dilution operating limits for flares with perimeter assist air (NHVdil)>= 22 BTU/ft2. Flare performance is influenced by many operating parameters. Yet, simple and clear guidelines for smokeless flaring under compliance still remains elusive. The objective of this study is to identify the optimal operation strategies for industrial flares. The detailed combustion mechanisms are too complicated for the speciation study of flaring events with Computational Fluid Dynamic (CFD). To alleviate this computational difficulty, a reduced mechanism with 50-species, namely Vsoot, which can predict the generation of soot accurately during the combustion of C1-C4 hydrocarbons, was generated. Comparison results showed that the concentration of soot precursor species predicted by the Vsoot combustion mechanism was comparable to the results predicted by full USC-II combustion mechanism. The Vsoot combustion mechanism was further validated successfully against experimental performance indicators including ignition delay and adiabatic flame temperature. Computational Fluid Dynamic (CFD) models combined with The Vsoot combustion mechanism were developed to simulate the flaring events. It was observed that simulation results agree well with experimental data.Random Forest (RF) predictive models for flare CE, Opacity and S were built based on literature data collected from 1983 to 2014 for steam assist flare. The outliers were deleted. Flare gas compositions include natural gas, methane, propylene, ethylene, propane, and typical refinery fuel. RF prediction results agreed well with the experiment results with high value of correlation coefficient R; In addition, RF models have good time and space complexity.

Electronic reproduction.
Ann Arbor, Mich. :
ProQuest,
2024

Mode of access: World Wide Web

ISBN: 9798645479169Subjects--Topical Terms:

555952
Chemical engineering.
Subjects--Index Terms:

Combustion kineticsIndex Terms--Genre/Form:

554714
Electronic books.

Combustion Kinetics Study and Machine Learning for the Simulation and Control of Industry Flares.
LDR:03612ntm a22003497 4500 001 1148106
005 20240916070000.5
006 m o d
007 cr bn ---uuuuu
008 250605s2019 xx obm 000 0 eng d
020 $a 9798645479169
035 $a (MiAaPQ)AAI27666379
035 $a AAI27666379
040 $a MiAaPQ $b eng $c MiAaPQ $d NTU
100 1 $a Wang, Anan. $3 1473999
245 1 0 $a Combustion Kinetics Study and Machine Learning for the Simulation and Control of Industry Flares.
264 0 $c 2019
300 $a 1 online resource (95 pages)
336 $a text $b txt $2 rdacontent
337 $a computer $b c $2 rdamedia
338 $a online resource $b cr $2 rdacarrier
500 $a Source: Dissertations Abstracts International, Volume: 81-12, Section: B.
500 $a Advisor: Lou, Helen H.
502 $a Thesis (Ph.D.)--Lamar University - Beaumont, 2019.
504 $a Includes bibliographical references
520 $a Industrial flare is used to burn off the unwanted gas during production. Global gas flaring increased by 3% in 2018 to 145 billion cubic meters (bcm) based on satellite data of The World Bank. If not combusted completely, intermediate products or incomplete combustion products are formed. The EPA Refinery Sector Rule (RSR) emphasize smokeless flaring with combustion efficiency (CE) >= 96.5% and destruction and remove efficiency (DRE) >= 98% for all types of flares. RSR also specified operating conditions to achieve those objectives including combustion zone net heating value (NHVcz) >= 270 BTU/scf, dilution operating limits for flares with perimeter assist air (NHVdil)>= 22 BTU/ft2. Flare performance is influenced by many operating parameters. Yet, simple and clear guidelines for smokeless flaring under compliance still remains elusive. The objective of this study is to identify the optimal operation strategies for industrial flares. The detailed combustion mechanisms are too complicated for the speciation study of flaring events with Computational Fluid Dynamic (CFD). To alleviate this computational difficulty, a reduced mechanism with 50-species, namely Vsoot, which can predict the generation of soot accurately during the combustion of C1-C4 hydrocarbons, was generated. Comparison results showed that the concentration of soot precursor species predicted by the Vsoot combustion mechanism was comparable to the results predicted by full USC-II combustion mechanism. The Vsoot combustion mechanism was further validated successfully against experimental performance indicators including ignition delay and adiabatic flame temperature. Computational Fluid Dynamic (CFD) models combined with The Vsoot combustion mechanism were developed to simulate the flaring events. It was observed that simulation results agree well with experimental data.Random Forest (RF) predictive models for flare CE, Opacity and S were built based on literature data collected from 1983 to 2014 for steam assist flare. The outliers were deleted. Flare gas compositions include natural gas, methane, propylene, ethylene, propane, and typical refinery fuel. RF prediction results agreed well with the experiment results with high value of correlation coefficient R; In addition, RF models have good time and space complexity.
533 $a Electronic reproduction. $b Ann Arbor, Mich. : $c ProQuest, $d 2024
538 $a Mode of access: World Wide Web
650 4 $a Chemical engineering. $3 555952
653 $a Combustion kinetics
653 $a Industry flares
653 $a Machine learning
655 7 $a Electronic books. $2 local $3 554714
690 $a 0542
710 2 $a ProQuest Information and Learning Co. $3 1178819
710 2 $a Lamar University - Beaumont. $b Chemical Engineering. $3 1181367
773 0 $t Dissertations Abstracts International $g 81-12B.
856 4 0 $u http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=27666379 $z click for full text (PQDT)