An artificial intelligence approach for biomass devolatilisation in an industrial CFD model with advanced turbulence-chemistry interaction

Simple item page
dc.contributor.advisorMeyer, Chris J.en_ZA
dc.contributor.advisorLaubscher, Rynoen_ZA
dc.contributor.authorDu Toit, Philip C.en_ZA
dc.contributor.otherStellenbosch University. Faculty of Engineering. Dept. of Mechanical and Mechatronic Engineering.en_ZA
dc.date.accessioned2018-02-28T13:04:42Z
dc.date.accessioned2018-04-09T07:10:25Z
dc.date.available2018-02-28T13:04:42Z
dc.date.available2018-04-09T07:10:25Z
dc.date.issued2018-03
dc.descriptionThesis (PhD)--Stellenbosch University, 2018.en_ZA
dc.description.abstractENGLISH SUMMARY: The ground work to include more detailed chemistry than global approaches in a combustion simulation was completed. A reduced-order model of the Biomass Chemical Percolation Devolatilisation model, ANN-Bio-CPD, was developed and implemented with artifcial neural networks in order to achieve ease of execution and computational cost reduction with regard to an industrial computational fluids dynamics application. ANN-Bio-CPD was validated with wire-mesh reactor and drop-tube furnace experiments from literature. Subsequently, the Eddy Dissipation Concept (EDC) turbulence-chemistry interaction model was implemented and validated with ANN-Bio-CPD in a bagasse- fired boiler simulation. The EDC model constants were adjusted to achieve the correct temperature and intermediate species results in combination with a two-step global reaction mechanism.en_ZA
dc.description.abstractAFRIKAANSE OPSOMMING: 'n Kunsmatige intelligensiebenadering vir biomassa-devolatilisering in 'n industriële CFD-model met gevorderde turbulensie-chemie-interaksie. Die basis om meer gedetailleerde chemie as globale benaderings in 'n verbrandingsimulasie in te sluit, is voltooi. 'n Verminderde-orde model van die Biomassa Chemiese Perkolasie Devolatilisering model, ANN-Bio-CPD, is ontwikkel en met kunsmatige neurale netwerke geïmplementeer om uitvoering te vergemaklik en berekeningskostes te verminder rakende die toepassing van numeriese vloeidinamika op 'n industriële skaal. ANN-Bio-CPD is gevalideer met die eksperimente van draad-maas reaktors- en valbuis-oonde uit die literatuur. Vervolgens is die "Eddy Dissipation Concept"(EDC) turbulensie-chemie interaksie model geïmplementeer en gevalideer met ANN-Bio-CPD in 'n bagasse-gestookte ketelsimulasie. Die EDCmodelkonstantes is aangepas om die korrekte temperatuur en intermediêre spesies resultate te bereik in kombinasie met 'n tweestap globale reaksie meganisme.af_ZA
dc.format.extentxix, 141 pages ; illustrationsen_ZA
dc.identifier.urihttp://hdl.handle.net/10019.1/103817
dc.language.isoen_ZAen_ZA
dc.publisherStellenbosch : Stellenbosch Universityen_ZA
dc.rights.holderStellenbosch Universityen_ZA
dc.subjectEddy Dissipation Concepten_ZA
dc.subjectCombustionen_ZA
dc.subjectArtificial intelligence -- Engineering applicationsen_ZA
dc.subjectBiomass devolatilisationen_ZA
dc.subjectNeural networks (Computer science)en_ZA
dc.subjectUCTD
dc.titleAn artificial intelligence approach for biomass devolatilisation in an industrial CFD model with advanced turbulence-chemistry interactionen_ZA
dc.typeThesisen_ZA
Files
Original bundle
Now showing 1 - 1 of 1
Thumbnail Image
Name:
dutoit_artificial_2018-5.pdf
Size:
11.71 MB
Format:
Adobe Portable Document Format
Description:
Download thesis
Download
License bundle
Now showing 1 - 1 of 1
Thumbnail Image
Name:
license.txt
Size:
1.71 KB
Format:
Plain Text
Description:
Download
Collections
Doctoral Degrees (Mechanical and Mechatronic Engineering)