Master Thesis Projects - in ModSimCompMech

From UmIT
Revision as of 07:38, 6 October 2010 by MartinServin (talk | contribs)
Jump to navigation Jump to search

We offer projects mainly in computational science and engineering. Your background could be applied mathematics, computer science, computational physics, control, mechanical engineering, robotics - but this is not exclusive. Please contact us with your interest and CV. Feel free also to propose thesis project by yourself.

Announced projects

Merge/split in rigid multibody simulation

Adaptive simplification of rigid multibody systems by merging bodies into agglomerate superbodies can accelerate large-scale simulation by many orders in magnitude. The actions should preserves physical invariants, minimizes computational time, produce consistent pressure force distributions and allow for impulse propagation. Application to vehicle or robot simulations with large piles of contacting bodies.

Comparative study between different rigid body simulation software

Study quantitative results in dynamics, computational performance, memory managment and parallelization opportunities between different rigid body simulation software. Both theory and numerical experiments with a set of test scenes relevant for using the software for virtual prototyping of complex mechanical systems. The softwares are (tentative list) AgX Multiphysics, Bullet Physics, EDEM.

Quantitative analysis of frictional contact models and solvers

Implement different formulations of dry friction for rigid bodies - some old and some new; performance analysis including: testing pipeline, design of statistical measures, scientific visualization of results.

Parallel numerical solutions of sparse linear systems

  • hierarchical data formats for sparse matrices
  • factorization update and downdate
  • load balancing
  • quantitative comparisons to existing libraries

Geometric contact reduction for 3D objects

It is possible to identify contact points that are redundant to the dynamics contacting rigid body system simulations. Eliminating redundant contact can have dramatic effect on the computational performance. The project include: implementing existing models, development of new techniques (advanced), quantitative evaluation of contact set quality.

Multiphysics: coupling of fluid with rigid multibody dynamics on GPGPU

  • evaluation of solvers
  • evaluation of models for boundary conditions
  • management of shared memory between CPU and GPGPU

Analytical system dynamics

Testing theoretical models and numerical simulations of coupled systems including: rigid multibodies, hydraulics (fluid power), electrical networks, heat transfer.

Adaptive resolution in particle fluid simulation

Extension of existing constraint based particle fluids to adaptive level of detail. Alternatives to particle representation of coarse grained regions. Application to industrial systems or scientific studies of ocean freak waves.

Real-time simulation models of ground-tire interaction and transmission lines

New contact model for ground-tire interaction suitable for real-time simulation of constrained multibody systems. Constraint based modeling of transmission lines. Applications to new design of terrain vehicles.

New contact and friction models

Contact reduction (theory, implementation, analysis); evaluation of a number of existing contact models (implementation, analysis) of Anitescu/Potra, Anitescu/Hart, Kaufamn/Pai, pairwise models, Kane/Pandolfi; development of a new model including reduction, delocalization and aspirity (theory, implementation, analysis); splitting technique for friction LCP solver.

Sparse parallel solvers with applications to QP

Sparse parallel factorization of LDL update and downdate with applications to quadratic programming (QP); block pivot methods for QPs (smoothing, splitting, application to frictional contacts, direct iterative hybrid solver); splitting techniques for frictional LCP solver; GPGPU techniques (sparse direct LDL solver, CG preconditioning, application to QP contact problems).

Time-integration of non-smooth dynamical systems

Linear and nonlinear constraint stabilization, impacts and nonsmooth mechanics, geometric integration and invariants, stable integration of rotational degrees of frodeom, robotics O(N) solver for closed loop systems.

Continuous collision detection

New approach to continuous collision detection between fast moving thin cylinder geometries.

Interactive visco-elastic-plastic simulation (with Stanford University)

Haptic fluid simulation (with Stanford University)

Running projects

  • Projected conjugate-gradient solver for contacting rigid bodies on GPGPU
  • Parallel factorization of symmetric indefinite linear systems
  • Viscoplastic constraint fluids
  • Parallelization of conjugate gradient solver for mixed linear complementarity problem

Previous projects

  • A Parallel Blocked Multifrontal Implementation of Colesky Factorization for Sparse Matrices (Olof Sabelström)
  • Shared Control of Mechanical Systems in Virtual Environments (Anders Hansson)
  • Constraint Fluids on GPU* (Martin Nilsson)
  • Realtime Simulation of Wires (Fredrik Nordfelth)
  • Phun* (Emil Ernefeldt)
  • Rigid Body Simulation of Macro Molecules (Christian Svebilius)
  • Simulation of off-road Vehicle (Erik Linder)
  • Real-Time Simulation of Deformable Objects (Niklas Melin)
  • Smoothed Particle Hydrodynamics on the Cell Broadband Engine* (Nils Hjelte)
  • Parallel Simulation of Particle Fluids* (Mattias Linde)
  • ... and several more

'*' was run at VRLab

External projects

DSCN1804.JPG