Options
Entwicklung von Methoden zum KI-gestützten Propellerentwurf im frühen Entwurfsstadium
Citation Link: https://doi.org/10.15480/882.4653
Publikationstyp
Master Thesis
Date Issued
2022-08-10
Sprache
German
Author(s)
Advisor
Title Granting Institution
Technische Universität Hamburg
Place of Title Granting Institution
Hamburg
Examination Date
2022-08-22
Institut
TORE-DOI
Citation
Technische Universität Hamburg (2022)
The aim of the work is to develop an AI-based design method for the creation of a propeller geometry that serves as a starting point for the design process. It should be possible to consider boundary conditions such as the maximum diameter, the required thrust or the available power. Based on a literature review regarding possible AI approaches or methods, these are evaluated in terms of their suitability for supporting propeller design. They range from simple regressions to artificial neural networks. Subsequently, the AI approaches suitable for the task are to be selected and corresponding models are to be implemented from them under consideration of the mentioned boundary conditions. The models are to be trained on the basis of available data from propeller series. In order to assess their quality, the calculated and available free-running characteristics are to be compared for an example propeller. Based on the comparison, neural networks are best suited for this application.
Subjects
Propeller
Entwurf
Maschinelles Lernen
DDC Class
600: Technik
620: Ingenieurwissenschaften
Loading...
Name
MA_Strecker.pdf
Size
3.86 MB
Format
Adobe PDF