|Tipo di tesi||Tesi di dottorato di ricerca|
|Titolo||Tecniche di Ottimizzazione applicate alla Progettazione in Campo Automotive|
|Titolo in inglese||Applications of Optimization to Automotive Engineering Design|
|Settore scientifico disciplinare||ING-IND/14 - PROGETTAZIONE MECCANICA E COSTRUZIONE DI MACCHINE|
|Corso di studi||Scuola di D.R. in HIGH MECHANICS AND AUTOMOTIVE DESIGN & TECHNOLOGY / MECCANICA AVANZATA E TECNICA DEL VEICOLO|
|Data inizio appello||2014-03-26|
|Disponibilità||Accessibile via web (tutti i file della tesi sono accessibili)|
Il progresso scientifico consiste nell'idea che la scienza aumenti la sua capacità di problem-solving attraverso l'applicazione del metodo scientifico. Questo concetto trova la sua massima espressione nell'ormai diffusa tendenza a cercare sempre la migliore soluzione in precise condizioni, ovvero nell'ottimizzazione.
Scientific progress is the idea that science increases its problem-solving ability through the application of the scientific method. This concept finds its most intense expression in the diffused tendency of searching the best solution under definite circumstances: this is called optimization. Optimization finds one of its most important source of inspiration in the world of engineering. The reasons can be found in the increasingly elaborate design challenges and more in general, in the decision making process. Moreover, decisions very often need to be made in a short time in order to control a running process. In the last few years, the availability of powerful computers has made it possible to obtain quick detailed analyses of complex systems. Consequently, numerical simulation has become an essential tool in the industrial design process. A simulation allows the system behaviour to be described by solving a series of algebraic or differential equations. Otherwise, from the optimization point of view, when the equations are solved by an external software, the system becomes a black-box: the functional relationship between the independent variables and the performance indicators is known only implicitly. When high computational cost, non-linear behaviour, stochasticity and multiple conflicting objectives are added, the seek for the the best solution in compliance with all the constraints can require too much time. This leads to incompatibility between problem-solving methodology and the short time generally available. In this work, computational challenges arising from complex simulation-based optimization problems are addressed. All the considered test cases are automotive applications, with special emphasis on improving the predictability of finite element simulations in view of cars light-weighting. In fact this mission implies the reduction of the design safety margins and it becomes affordable only by increasing the confidence in simulations results. Particular effort has been spent in order to obtain methodologies compatible with the time scheduling of the specific project phase.