Title :
The stochastic design of force-minimized compact magnets for high-field magnetic resonance imaging applications
Author :
Crozier, Stuart ; Jenkinson, Christopher J Snape ; Forbes, Larry K.
Author_Institution :
Centre for Magnetic Res., Queensland Univ., Brisbane, Qld., Australia
fDate :
6/1/2001 12:00:00 AM
Abstract :
New designs for force-minimized compact high-field clinical MRI magnets are described. The design method is a modified simulated annealing (SA) procedure which includes Maxwell forces in the error function to be minimized. This permits an automated force reduction in the magnet designs while controlling the overall dimensions of the system. As SA optimization requires many iterations to achieve a final design, it is important that each iteration in the procedure is rapid. We have therefore developed a rapid force calculation algorithm. Novel designs for short 3- and 4-T clinical MRI systems are presented in which force reduction has been invoked. The final designs provide large homogeneous regions and reduced stray fields in remarkable short magnets. A shielded 4-T design that is approximately 30% shorter than current designs is presented. This novel magnet generates a full 50-cm diameter homogeneous region
Keywords :
biomedical MRI; iterative methods; simulated annealing; superconducting magnets; 3 T; 4 T; 50 cm; Maxwell forces; automated force reduction; clinical MRI magnets; error function; force reduction; force-minimized compact magnets; high-field magnetic resonance imaging applications; homogeneous regions; iterations; modified simulated annealing procedure; overall dimensions; rapid force calculation algorithm; stochastic design; stray fields; Design methodology; Design optimization; High-resolution imaging; Image resolution; Magnetic resonance imaging; Process design; Simulated annealing; Spatial resolution; Stochastic resonance; Superconducting magnets;
Journal_Title :
Applied Superconductivity, IEEE Transactions on