Javascript must be enabled for the correct page display

Dipole Source Localization using the Continuous Wavelet Transform Method in a Dynamic Fluid Simulation

van Keulen, Ingeborg (2019) Dipole Source Localization using the Continuous Wavelet Transform Method in a Dynamic Fluid Simulation. Bachelor's Thesis, Artificial Intelligence.

[img]
Preview
Text
AI_BA_2019_IngeborgVanKeulen.pdf

Download (778kB) | Preview
[img] Text
toestemming.pdf
Restricted to Registered users only

Download (120kB)

Abstract

There are different methods available to localize a vibrating source in water using Artificial Lateral Lines that measure velocity profiles. There are several methods to reconstruct the location from a velocity profile. One of these methods is the continuous wavelet transform (CWT). The CWT method is a template matching method, which uses wavelets as templates. The CWT method has been tested on dipoles producing potential flow, which describes an ideal fluid. A more realistic fluid may have turbulence or vorticity, which is caused by viscosity. In this research, the performance of the CWT method has been evaluated on different fluid flow models: non-viscous potential flow (with and without added noise), simulated laminar flow and simulated turbulent flow. The results show that the CWT method is resilient to noise up to a moderate degree. The performance of the CWT method in simulated laminar and turbulent flow does not decrease much with respect to potential flow, although the error landscape becomes slightly more complex. The performance of the CWT is a promising method for source localization applications with data in the real world.

Item Type: Thesis (Bachelor's Thesis)
Supervisor name: Netten, S.M. van and Wolf, B.J.
Degree programme: Artificial Intelligence
Thesis type: Bachelor's Thesis
Language: English
Date Deposited: 11 Feb 2019
Last Modified: 15 Feb 2019 15:37
URI: https://fse.studenttheses.ub.rug.nl/id/eprint/19143

Actions (login required)

View Item View Item