magistrsko delo
Luka Vodišek (Author), Matej Vesenjak (Mentor), Miran Brezočnik (Co-mentor), Matej Borovinšek (Co-mentor)

Abstract

Zaradi uporabnih lastnosti kovinskih pen jih inženirji vse pogosteje vključujejo v nove izdelke. Njihove materialne lastnosti lahko navdano določijo z eksperimentalnimi preizkusi, saj so matematični modeli pen na podlagi enostavnih standardnih testov skoraj nedoločljivi. V tej nalogi se bomo seznanili z drugim načinom, kjer bomo materialne lastnosti določili tako, da bomo dva različna eksperimenta poskušali ponoviti s trdnostno analizo po metodi končnih elementov. Okvirne vrednosti so dane, vendar pa bi iskanje rešitev po vseh spremenljivkah trajalo predolgo, če bi jih spreminjali ročno, zato si bomo pomagali z genetskimi algoritmi in programom, ki to spreminja sam, oziroma se uči iz prejšnjih rešitev in išče boljše. Ker smo obravnavali dva eksperimenta, bomo tudi rešitve ocenjevali po dveh kriterijih, zato bomo potrebovali Paretto fronte, ki jih bomo dobili s pomočjo algoritma NSGA II. Metoda, po kateri smo v tej nalogi iskali najboljše rešitve, se je izkazala za dobro, predvsem obetavna je optimizacija po več parametrih. Ker ima še veliko možnosti za nadgradnjo, lahko vključimo večje število spremenljivk in večje število parametrov optimiziranja z dodajanjem različnih enostavnih preizkusov. Za komercialno uporabo bi bilo treba raziskati, kateri preizkusi dajejo najboljše rezultate. Obravnavani način določanja materialnih lastnosti s pomočjo eksperimentalnih poizkusov ter ponovitvijo v simulacijskem okolju ima velik potencial, ki se lahko razširi tudi na druge materiale, ne le na kovinske pene.

Keywords

aluminijeva zlitina;kovinska pena;materialni model;ABAQUS;genetski algoritem;NSGA II;magistrske naloge;

Data

Language: Slovenian
Year of publishing:
Typology: 2.09 - Master's Thesis
Organization: UM FS - Faculty of Mechanical Engineering
Publisher: [L. Vodišek]
UDC: [620.17:621.762]:004.89(043.2)
COBISS: 21292566 Link will open in a new window
Views: 1070
Downloads: 143
Average score: 0 (0 votes)
Metadata: JSON JSON-RDF JSON-LD TURTLE N-TRIPLES XML RDFA MICRODATA DC-XML DC-RDF RDF

Other data

Secondary language: English
Secondary title: Determination of closed-cell foam material parameters at compressive load
Secondary abstract: Due to the useful properties of metal foams, engineers increasingly involved them in new products. However, their material properties can often be defined only by experimental tests, since the mathematical models of the foams are almost indeterminate on the basis of simple standard tests. In this task, we will learn about another way in which we will determine the material properties by trying to repeat two different experiments with the strength analysis based on the finite element method. The reference values are given, but finding the solution across all variables would take too long to be manually modified. Therefore, genetic algorithms and a program that changes it simultaneously or learns from previous solutions and searches for better ones have been used. Since two experiments have been considered, we will also evaluate the solutions according to two criteria. Therefore, we will need the Paretto fronts contained in the NSGA II algorithm. The method with which we sought the best solutions in this task has proved to be good, especially optimization for several parameters. However, there is still a lot of room for improvement, since we can include a larger number of variables and a greater number of optimization parameters by adding a variety of simple tests. For commercial use, it would be necessary to investigate which tests give the best results. The discussed method of determining material properties through experiments and repetition in the simulation environment has a great potential, which can be extended to other materials and is not only exclusive for metal foams.
Secondary keywords: aluminium alloy;metal foam;material model;genetic algorithm;
URN: URN:SI:UM:
Type (COBISS): Master's thesis/paper
Thesis comment: Univ. v Mariboru, Fak. za strojništvo, Računalniško inženirsko modeliranje
Pages: XVI, 67 f.
ID: 10865260