Multi-objective Optimization in Automotive Design
Independent thesis Advanced level (professional degree), 20 credits / 30 HE creditsStudent thesis
When developing complex systems like an automobile it is common that one of the many departments involved gets the lead in the development process. The overall design involves many highly coupled subsystems, requiring the making of design decisions by other departments. The traditional sequential, spiral design process tends to reduce the other departments (subsystems) influences on the overall design and optimization. To avoid a development process taking this path, numerical design and optimization models can be used initially during the development to optimize the automobile system with all department objectives and requirements weighted simultaneously and equally in the overall design. Thereby shortening development cycle time and improving the final product and reducing the development cost. During this Master Thesis a Multidisciplinary Design Optimization (MDO) model has been developed that expresses marketability of a standard sedan on the U.S market based on fuel efficiency, curb weight and market price. The model was developed by following Product Design and Development methodology developed by Ulrich & Eppinger to optimize an automobile in the early stages of the development process. The model consists of several coupled subsystems which can be optimized using Collaborative Optimization (CO) theory and it is developed in MATLAB. By combining an Engine performance analysis program, Regression analysis’s, Physical relations and Market trends. The model was also used in PHX ModelCenterTM, a commercially available Process Integration and Design Optimization (PIDO) software to perform a Design Space Exploration (DSE) and multi-objective optimization. The optimization was made with Darwin- and Boeing’s Design Explorer algorithms. An optimal solution with each algorithm was found for the automobile system. The results of this work show that it’s possible to develop multidisciplinary optimization methods for use in the early stage synthesis design of automobiles based on advanced MDO and CO theories together with appropriate software. It also shows that developing this kind of models is a comprehensive task suited for a group of people with as many different proficiency as there are departments involved in an automobile development process. The fidelity of the optimized design produced by this method depends on the complexity of the models used to define the overall system as well as the subsystems.
Place, publisher, year, edition, pages
2014. , 128 p.
Technology, Automobile, Numerical Optimization, Multidisciplinary Design Optimization (MDO), Collaborative Optimization (CO), Ulrich & Eppinger, MATLAB, Engine Performance Analysis program, Regression Analysis, PHX ModelCenterTM, Process Integration and Design Optimization (PIDO), Design Space Exploration (DSE), Darwin Algorithm, Boeing’s Design Explorer Algorithm
IdentifiersURN: urn:nbn:se:ltu:diva-57571Local ID: e39b6765-39be-4c99-b7d3-5b9a01388a29OAI: oai:DiVA.org:ltu-57571DiVA: diva2:1030959
Subject / course
Student thesis, at least 30 credits
Mechanical Engineering, master's level
Hefazi, HamidJohn Micklow, Gerald
Validerat; 20140924 (global_studentproject_submitter)2016-10-042016-10-04Bibliographically approved