Multi-scale modelling and simulation in electromagnetics

Principal Investigator: Professor C Christopoulos

Other Investigators: Dr P D Sewell

Project Partner: B A E Systems Plc

Starts: 14 October 2002
Ends: 13 October 2005

Value: £206,738

The proposal addresses the problem of modelling fine features (such as wires) in electromagnetic numerical problems. The finite spatial resolution inherent in such models makes the description of fine wire; and other similar features and their exact placement problematic, except when an unrealistically fine mesh is used. Even with the most optimistic projections for the future speed and memory of computers it is essential that efficient models for fine features are developed to reduce computational and problem definition costs. The proposed models are based on analytical modal expansions for the field around wires. These modal components are then mapped onto the surrounding numerical mesh to accomplish large scale computation. The proposed models are capable of describing emission from and coupling to wires. The wire diameter can be set independently of the prevailing spatial resolution and the wire(s) can be placed anywhere in space and not just on specific mesh points. This is a major advantage over existing techniques which suffer from many restrictions and inaccuracies. Other objects, in addition to wires, may be described by the proposed models and this possibility will be explored in connection with other developments at Nottingham based on scattering algorithms and digital filter techniques.