On Aspects of Magnetic Signature Source Modelling using Near-Field Measurements

Using Multipolar Bases and Bayesian Inference on Complex Ship Geometries

Master Thesis (2026)
Author(s)

O.A. Schokker (TU Delft - Electrical Engineering, Mathematics and Computer Science)

Contributor(s)

A.R.P.J. Vijn – Mentor (TU Delft - Mathematical Physics)

J.L.A. Dubbeldam – Mentor (TU Delft - Mathematical Physics)

N.V. Budko – Graduation committee member (TU Delft - Numerical Analysis)

Faculty
Electrical Engineering, Mathematics and Computer Science
More Info
expand_more
Publication Year
2026
Language
English
Graduation Date
27-03-2026
Awarding Institution
Delft University of Technology
Programme
['Applied Mathematics']
Faculty
Electrical Engineering, Mathematics and Computer Science
Reuse Rights

Other than for strictly personal use, it is not permitted to download, forward or distribute the text or part of it, without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license such as Creative Commons.

Abstract

his thesis investigates the modelling of a ship’s magnetic field, by means of an equivalent source model defined on an arbitrarily shaped surface enclosing the source. This setup allows for the use of measurements close to the enclosing surface, such as on the seabed in shallow waters. A multipolar basis is introduced to reduce the dimensionality of the problem, which allows for a convenient mapping to the Decreasing Spherical Harmonic Expansion to describe the far field with low computational costs. It is found that the method performs well in determining a source model and predicting the magnetic field in new locations, provided that enough well-distributed measurements are available. The method remains accurate in the presence of noise, although measurements very close to the enclosing surface reduce performance, a reduction attributed to an approximation that may be improved. Bayesian inference is used to stabilise the method when a low number, badly distributed and/or noisy measurements are available. We find that this regularisation indeed enables the use of the method in such situations. In addition, it is investigated whether orthogonality of the basis is required. It is found that it is not, and that a non-orthogonal basis can yield slightly better results. This suggests that the discretisation of the surface and the choice of solution space of the equivalent source are the main limiting factors in improving the solutions.

Files

License info not available