Development of a Magnetic Particle Tracking Technique using a Hybrid Numerical Optimization Algorithm

Abstract

This thesis focuses on the development of a 3D Magnetic Particle Tracking (MPT) technique using a hybrid numerical optimization algorithm which is able to reconstruct not only the particle’s location but also its orientation in the given measurement domain. The MPT is an inexpensive particle tracking technique which can be used in both transparent and opaque working conditions and requires no specific safety precautions for its use. In this work, initially, the algorithm is tested for accuracy numerically for simulation cases with continuous velocity/acceleration, sudden acceleration, sudden rotation, and the Brownian motion simulation. The accuracy of the algorithm is found to be comparable with the state-of-the-art optimization approach, but the reconstruction is orders of magnitude faster. The numerical simulation for the Brownian motion showed that the MPT position uncertainty can reach 0.86 % and the angular error is 1.5o for a measurement domain with a size of about 10 cm. On validation from the above simulations, the MPT is tested in an experimental setup to study the dense granular shear flow. The MPT algorithm is used to track a cylindrical tracer particle of aspect ratio 1 encapsulated with a neodymium magnetic bead in a cylinder filled with plastic balls, which acts as the bulk material, of similar density. It is observed that the balls show a layered structure in both the X-Y position distribution and in the vertical direction. Because of the above advantages and its high accuracy, MPT is a powerful tool for studying dense granular flows and provide insight into this physical phenomenon.