Приказ основних података о дисертацији

Modelling, simulation and optimization of obtaining products based on different rubber mixtures

dc.contributor.advisorBera, Oskar
dc.contributor.otherIkonić, Bojana
dc.contributor.otherBera, Oskar
dc.contributor.otherPavličević, Jelena
dc.contributor.otherKojić, Predrag
dc.contributor.otherNikačević, Nikola
dc.creatorLubura, Jelena
dc.date.accessioned2023-03-20T21:41:48Z
dc.date.available2023-03-20T21:41:48Z
dc.date.issued2023-03-10
dc.identifier.urihttps://www.cris.uns.ac.rs/DownloadFileServlet/Disertacija167093221528533.pdf?controlNumber=(BISIS)127380&fileName=167093221528533.pdf&id=21042&source=NaRDuS&language=srsr
dc.identifier.urihttps://www.cris.uns.ac.rs/record.jsf?recordId=127380&source=NaRDuS&language=srsr
dc.identifier.urihttps://www.cris.uns.ac.rs/DownloadFileServlet/IzvestajKomisije167093222451099.pdf?controlNumber=(BISIS)127380&fileName=167093222451099.pdf&id=21043&source=NaRDuS&language=srsr
dc.identifier.urihttps://nardus.mpn.gov.rs/handle/123456789/21315
dc.description.abstractU disertaciji je razvijen novi pristup optimizacije vulkanizacije koji omogućava određivanje optimalnih radnih parametara, vremena i temperature kalupa, kako bi se omogućilo dobijanje kvalitetnih homogenih proizvoda. Optimizacioni model je razvijen na primeru četiri sfere, različitih prečnika (2,5; 5; 10 i 20 cm). Dobijeni reometarski podaci vulkanizacije, komercijalno dostupne kaučukove smeše, fitovani su novim pristupom koji deli vulkanizacionu krivu na dva skupa fitovanja, umrežavanje i reverziju. Tačnost razvijenog pristupa fitovanja je potvrđena drugim tipom kaučukove smeše na bazi prirodnog kaučuka, sa različitim sadržajem biočađi kao punila i dobijena je visoka tačnost, gde vrednosti R2 nisu niže od 0,84,a MAPE vrednosti nisu više od 2,4 %. Kaučukova smeša sa biočađi kao punilom je detaljno ispitana različitim metodama kako bi se utvrdila mogućnost delimične zamene čađi sa biopunilom. Ustanovljeno je da komercijalno dostupna čađ može da se zameni za10 phr sa biočađi dobijenom od otpadne drvene biomase, bez značajne promene mehaničkih osobina konačnog gumenog proizvoda. Uveden je novi temperaturno zavisni parametar, nazvan stepen reverzije, koji omogućava određivanje najniže radne temperature (Tmin = 132,36 °C) na kojoj ne dolazi do reverzije i pregrevanja proizvoda. Određena je zavisnost optimalne temperature i vremena vulkanizacije od dimenzija gumenih proizvoda, a predloženi model ostvaruje visoki stepen tačnosti, gde su vrednosti R2 veće od od 0,8328 i MAPE manje od 2,3099 %. Jednačine prenosa toplote za odgovarajuću geometriju su rešavane istovremeno sa razvijenim kinetičkim modelom kako bi se izvršila simulacija vulkanizacije tri sfere različitih prečnika i dva gumena točka, od kojih je jedan komercijalni proizvod gumarske industrije. Predloženi model simulacije uključuje i hlađenje nakon vađenja proizvoda iz prese, jer se reakcije umrežavanja i dalje odvijaju zbog tople unutrašnjosti proizvoda. Kriterijum za vađenje proizvoda je vrednost srednjeg stepena vulkanizacije od 0,9, kako bi se obezbedilo dobijanje gumenih proizvoda odgovarajućih oblika. Optimizacija radnih parametara je izvršena na osnovu minimalne razlike maksimalnog i minimalnog stepena vulkanizacije, što omogućava efikasno izvođenje procesa za kraće vreme i dobijanje kvalitetnih homogenih proizvoda. Određeni su optimalni radni uslovi za sve gumene proizvode. Model veštačke neuronske mreže je razvijen za predviđanje reoloških podataka vulkanizacije komercijalno dostupne kaučukove smeše. Ispitivan je uticaj tri aktivacione funkcije i različit broj skrivenih slojeva i neurona u njma na tačnost predviđanja mreže, pri čemu neuronska mreža sa Softplus aktivacionom funkcijom i 20 neurona u dva skrivena sloja predviđa zavisnost obrtnog momenta od vremena sa visokom tačnošću, tako da su vrednosti MAPE i MSE niže od 2 % i 0,032 dN2m2, redom, a R2 više od 0,98. Razvijeni model veštačke neuronske mreže je potvrđen kaučukovom smešom sa biočađi kao punilom.sr
dc.description.abstractA novel approach for the optimization of vulcanization process was developed, which enables the determination of optimal operating parameters, time exposure and mold temperature, in order to obtain a high-quality products. The spheres of different diameters (2.5, 5, 10 and 20 cm) were selected for development of optimization model. Obtained vulcanization rheometer data of commercially available rubber blend was fitted using a novel approach, which devide the vulcanization curve into two fitting sets, curing and reversion. The accuracy of the developed fitting approach was examined with another type of rubber mixture based on natural rubber, with different content of biochar as a filler. High accuracy was obtained, where R2 values were not lower than 0.84 and MAPE values were not higher of 2.4%. The rubber compound with biochar as a filler was thoroughly tested by different methods in order to determine the possibility of carbon black partial replacement with biofiller. Commercially available carbon black can be replaced by 10 phr of biochar obtained from hardwood waste biomass, without significantly changing the mechanical properties of the final rubber product. It was introduced a new temperature-dependent parameter, named the reversion degree, enabling the determination of the lowest vulcanization operating temperature(Tmin = 132.36 °C), ensuring that reversion and product overheating do not occur. The dependence of the optimal temperature and vulcanization time on the dimensions of rubber products was determined, and the proposed model achieves a high degree of accuracy, where R2 values were greater than 0.8328 and MAPE was less than 2.3099 %. The heat transfer equations for the appropriate geometry were solved simultaneously with the developed kinetic model for simulation the vulcanization process of three spheres of different diameters and two rubber wheels, where one is a commercial product of the rubber industry. The proposed simulation model includes cooling after removing the product from the mold, as the crosslinking reactions continue to take place due to the warm interior of the product. The criteria for extracting the product from mold was set to value 0,9 of the average vulcanization degree, ensuring obtaining the appropriate shapes of rubber products. The optimization of vulcanization was carried out with obtained the minimum difference between the maximum and minimum vulcanization degree, enabling efficient process and obtaining high-quality homogeneous products. The optimal process parameters for all rubber products have been determined. An artificial neural network model was developed to predict the vulcanization rheological data of a commercially available rubber compound. The influence of three activation functions and different number of hidden layers and neurons, on the neural network prediction accuracy was examined, where by the neural network with the Softplus activation function and 20 neurons in two hidden layers predicts the torque dependence on time with high accuracy, where the MAPE and MSE values were lower than 2% and 0.032 dN2m2, respectively, and R2 values were higher than 0.98. The developed artificial neural network model was validated with a rubber mixture with biochar as a filler.en
dc.languagesr (latin script)
dc.publisherУниверзитет у Новом Саду, Технолошки факултетsr
dc.rightsopenAccessen
dc.rights.urihttps://creativecommons.org/licenses/by-nc-nd/4.0/
dc.sourceУниверзитет у Новом Садуsr
dc.subjectModelovanjesr
dc.subjectModelingen
dc.subjectkineticsen
dc.subjectsimulationen
dc.subjectoptimizationen
dc.subjectrubber productsen
dc.subjectbiocharen
dc.subjectartificial neural networksen
dc.subjectkinetikasr
dc.subjectsimulacijasr
dc.subjectoptimizacijasr
dc.subjectproizvodiod gumesr
dc.subjectbiočađsr
dc.subjectveštačke neuronske mrežesr
dc.titleModelovanje, simulacija i optimizacija dobijanja gumenih proizvoda na osnovu različitih kaučukovih smešasr
dc.title.alternativeModelling, simulation and optimization of obtaining products based on different rubber mixturesen
dc.typedoctoralThesissr
dc.rights.licenseBY-NC-ND
dc.identifier.fulltexthttp://nardus.mpn.gov.rs/bitstream/id/150195/Izvestaj_komisije_13413.pdf
dc.identifier.fulltexthttp://nardus.mpn.gov.rs/bitstream/id/150194/Disertacija_13413.pdf
dc.identifier.rcubhttps://hdl.handle.net/21.15107/rcub_nardus_21315


Документи за докторску дисертацију

Thumbnail
Thumbnail

Ова дисертација се појављује у следећим колекцијама

Приказ основних података о дисертацији