Full Text Available
Note: Clicking the button above will open the full text document at the original institutional repository in a new window.
Intelligent optimisation of deep drawing process parameters in the production of aluminium non-rotational symmetric monolithic components
Sarema, B. 2025. Intelligent optimisation of deep drawing process parameters in the production of aluminium non-rotational symmetric monolithic components. Unpublished doctoral dissertation. Stellenbosch: Stellenbosch University [online]. Available: https://scholar.sun.ac.za/items/fe12875c-4550-41b2...
Saved in:
| Main Author: | |
|---|---|
| Other Authors: | |
| Format: | Thesis |
| Language: | English |
| Published: |
Stellenbosch : Stellenbosch University
2025
|
| Subjects: | |
| Tags: |
No Tags, Be the first to tag this record!
|
| _version_ | 1867614139725119488 |
|---|---|
| access_status_str | Open Access |
| author | Sarema, Blessed |
| author2 | Matope, Stephen |
| author_browse | Matope, Stephen Sarema, Blessed |
| author_facet | Matope, Stephen Sarema, Blessed |
| author_sort | Sarema, Blessed |
| collection | Thesis |
| dc_rights_str_mv | Stellenbosch University |
| description | Sarema, B. 2025. Intelligent optimisation of deep drawing process parameters in the production of aluminium non-rotational symmetric monolithic components. Unpublished doctoral dissertation. Stellenbosch: Stellenbosch University [online]. Available: https://scholar.sun.ac.za/items/fe12875c-4550-41b2-a3c7-65125f0a644f |
| format | Thesis |
| id | oai:scholar.sun.ac.za:10019.1/132269 |
| institution | Stellenbosch University (South Africa) |
| language | English |
| last_indexed | 2026-06-10T12:47:17.937Z |
| license_str | Other — see source repository |
| provenance_str_mv | Harvested via OAI-PMH from SUNScholar — Stellenbosch University Repository |
| publishDate | 2025 |
| publishDateRange | 2025 |
| publishDateSort | 2025 |
| publisher | Stellenbosch : Stellenbosch University |
| publisherStr | Stellenbosch : Stellenbosch University |
| record_format | dspace |
| source_str | SUNScholar — Stellenbosch University Repository |
| spelling | oai:scholar.sun.ac.za:10019.1/132269 Intelligent optimisation of deep drawing process parameters in the production of aluminium non-rotational symmetric monolithic components Sarema, Blessed Matope, Stephen Sterzing, Andreas Stellenbosch University. Faculty of Engineering. Dept. of Industrial Engineering. Deep drawing (Metal-work) Manufacturing processes -- Mathematical models Aluminum forming Finite element method Mathematical optimization UCTD Sarema, B. 2025. Intelligent optimisation of deep drawing process parameters in the production of aluminium non-rotational symmetric monolithic components. Unpublished doctoral dissertation. Stellenbosch: Stellenbosch University [online]. Available: https://scholar.sun.ac.za/items/fe12875c-4550-41b2-a3c7-65125f0a644f Thesis (PhD)--Stellenbosch University, 2025. ENGLISH ABSTRACT: Deep drawing is an important manufacturing process in the automotive and aerospace industries. This study employed the deep drawing process to manufacture a non-rotational, monolithic automotive accessory, a rooftop, used in overlanding. The monolithic component replaces a polylithic product manufactured by joining 23 parts, using riveting, bolting, adhesive bonding and tungsten-inert-gas welding. This part count reduction brings several advantages, like shorter cycle times, from 127.5 minutes to 3.95 minutes, and less weight, as the mass is reduced from 29.78kg to 18kg. These changes bring sustainability to the automotive and overlanding industries. To achieve this shift, there was a need to optimise the deep drawing process parameters to enhance the quality of the monolithic rooftop automotive accessory. The optimisation process started with establishing product quality parameters, followed by critical process parameters. Material characterisation tests were performed, and the deep drawing process was simulated using AutoForm Finite Element Analysis (FEA) software guided by the Taguchi design of experiment procedures. The FEA results were analysed using Minitab® Statistical Software. Mathematical models for the deep drawing process were intelligently developed using the Automatic Relevance Determination (ARD) exponential and squared exponential functions of the Gaussian Process Models in the Model-Based Calibration Toolbox™ application in MATLAB®. The same toolbox was used for optimisation using the multi-objective genetic algorithm. Optimum parameters were successfully validated using FEA in AutoForm, Computer- Aided-Design (CAD) model comparison using MeshLab software and physical prototyping using a scaled-down part. The impact of six process parameters, varied at three levels on thirteen output parameters, was investigated and established. The six process parameters under investigation were the drawing gap, the die radius, the initial blankholder force, the coefficient of friction, the blank size, and the draw bead height. The output parameters were thickening, compression, insufficient stretch, safe forming, excessive thinning, risk of splits, splits, springback, wrinkles, surface lows, plastic strain, final blankholder force, and punch force. The modelling and optimisation results were presented in the form of unique operational windows. The coefficient of friction significantly impacted all the quality parameters; its optimum value converged at 0.10. A medium-sized blank, corresponding to 0.3% to 0.4% area offset of the flattened part, yielded optimum quality results. A 9mm or 15mm die radius, and a 2.0mm or 2.5mm drawing gap resulted in good quality products. This meant that a total of four toolsets could potentially be used. Depending on the tool set chosen, a bead height between 5.0mm and 6.5mm produced optimum results. The same applied to the initial blankholder force that ranged between 120kN and 210kN. The validation proved that all the quality results were within the specifications of safe forming namely: above 98%, thickening and compression less than 1%, excessive thinning less than 2%, zero splits, wrinkles less than 0.05, surface lows lower than 0.0002, a plastic strain of 0.2 and springback less than 5mm for all the optimum toolsets. A press size of 800 tons with a bed size of 3600mm by 2500mm, is required to produce the full-size monolithic component. AFRIKAANSE OPSOMMING: Dieptekening is ’n belangrike vervaardigingsproses in die motor- en lugvaartbedryf. Hierdie studie gebruik die dieptetekenproses om ’n nie-roterende, monolitiese motor-bykomstigheid vir oorlandsereise, genaamd die “bo-dak”, te vervaardig. Dié monolitiese komponent verbeter op sy voorganger deurdat dit 23 komponente met verbindingsprosesse (met behulp van klinknaels, boute, kleefmiddele en sweiswerk) saamvoeg. Die vermindering in komponente bied verskeie voordele. Naamlik; korter siklustye (127.5 minute tot 3.95 minute) en ʼn vermindering van massa (29.78kg tot 18kg). Hierdie positiewe veranderinge ondersteun en bemoedig die volhoubaarheid van die motor en oorlandse-reis bedrywe. Om die voorafgenoemde te kon realiseer moes die kwaliteitaanwysers van die monolitiese bykomstigheid se dieptetekenproses, geoptimeer word. Die optimeringsproses het begin met die vasstelling van aanwysers vir produkgehalte, gevolg deur aanwysers vir die kritiese proses. Toetse om die materiaal se karakterseienskappe te bepaal, is ook uitgevoer. Die dieptetekenproses kon toe met behulp van die AutoForm Eindige Elemente Analise (EEA) sagteware gesimuleer word. Die EEA-simulasies is met die Taguchi ontwerp vir eksperimentele prosedures belyn en die resultate daarvan met behulp van die Minitab® statisties ontleed. Wiskundige modelle vir die dieptetekenproses is deur beide die eksponensiële Automatiese Relevansie Bepaling (ARB) en die kwadraties eksponensiële funksies van die Gaussiese Proses Model asook deur die Model-Based Calibration Toolbox™ toepassing in MATLAB® statisties ontleed. Deur gebruik te maak van ʼn multi-doelwitte genetiese algoritme, kon dieselfde gereedskapstel in die optimeringsproses gebruik word. Die optimale aanwysers is deur EEA in AutoForm en deur Computer-Aided-Design (CAD) se modelvergelyking met die MeshLab sagteware geverifieer. Die impak van ses proses-aanwysers, wat op drie vlakke geverifieer en op 13 uitset-aanwysers getoets is, is ondersoek. Die proses-aanwysers sluit in die: tekenopening, radius, aanvanklike openinghouerkrag, wrywingskoëffisiënt, openinggrootte, en die trek-krale hoogte. Die uitset-aanwysers is verdikking, kompressie, onvoldoende rek, veilige vorming, oormatige verdunning, risiko van skeure, skeure, terugspring, kreukels, oppervlaklaagtes, plastiese spanning, finale openinghouerkrag, en ponskrag. Die resultate van die optimeringsproses is in die vorm van verskillende operasionele toetsgevalle voorgestel. Die wrywingskoëffisiënt het ’n beduidende impak op al die kwaliteitaanwysers getoon; sy optimale waarde het by 0.10 gekonvergeer. ’n Medium-grootte oop area, wat ooreenstem met 0.3% tot 0.4% area-afwyking van die platgemaakte deel, het optimale kwaliteit resultate opgelewer. ’n 9mm of 15mm radius, en ’n 2.0mm of 2.5mm tekenopening, het produkte binne aanvaarbare kwaliteitsstandaarde opgelewer. gebaseer op die voorafgenoemde is daar vier gereedskapstelle wat die potensiaal gehad het om ʼn hoë kwaliteit produk te kan vervaardig. Afhangend van gelange gereedskapstel, het ’n krale hoogte tussen 5.0mm en 6.5mm optimale resultate gebied. Dieselfde geld vir die aanvanklike openinghouerkrag wat tussen 120kN en 210kN gewissel het. Met die verifiëring van die EEA en CAD vergelykings, is dit bewys dat al die kwaliteit verwante resultate binne die spesifikasies van veilige vorming was. Naamlik; bo 98% verdikking en kompressie minder as 1%, oormatige verdunning minder as 2%, geen skeure, kreukels minder as 0.05, oppervlaklaagtes laer as 0.0002, gemiddelde plastiese spanning van 0.2 en terugspring minder as 5mm. ’n Drukgrootte van 800 ton met ’n bedgrootte van 3600mm by 2500mm sou nodig wees om die vol-grootte monolitiese bykomstigheid te vervaardig. Doctoral 2025-06-02T09:04:05Z 2025-06-02T09:04:05Z 2025-03 Thesis https://scholar.sun.ac.za/handle/10019.1/132269 en Stellenbosch University xxv, 347 pages : illustrations application/pdf Stellenbosch : Stellenbosch University |
| spellingShingle | Deep drawing (Metal-work) Manufacturing processes -- Mathematical models Aluminum forming Finite element method Mathematical optimization UCTD Sarema, Blessed Intelligent optimisation of deep drawing process parameters in the production of aluminium non-rotational symmetric monolithic components |
| title | Intelligent optimisation of deep drawing process parameters in the production of aluminium non-rotational symmetric monolithic components |
| title_full | Intelligent optimisation of deep drawing process parameters in the production of aluminium non-rotational symmetric monolithic components |
| title_fullStr | Intelligent optimisation of deep drawing process parameters in the production of aluminium non-rotational symmetric monolithic components |
| title_full_unstemmed | Intelligent optimisation of deep drawing process parameters in the production of aluminium non-rotational symmetric monolithic components |
| title_short | Intelligent optimisation of deep drawing process parameters in the production of aluminium non-rotational symmetric monolithic components |
| title_sort | intelligent optimisation of deep drawing process parameters in the production of aluminium non rotational symmetric monolithic components |
| topic | Deep drawing (Metal-work) Manufacturing processes -- Mathematical models Aluminum forming Finite element method Mathematical optimization UCTD |
| url | https://scholar.sun.ac.za/handle/10019.1/132269 |
| work_keys_str_mv | AT saremablessed intelligentoptimisationofdeepdrawingprocessparametersintheproductionofaluminiumnonrotationalsymmetricmonolithiccomponents |