Vibroacoustic matrices from Andre Cote, Univ. Sherbrooke, Quebec. Andre Cote GAUS, genie mecanique Universite de Sherbrooke | E-mail : Andre.Cote :at the domain: gme.usherb.ca 2500, boul. Universite | Tel : (819) 821-8000 ext. 3166 Sherbrooke ( Quebec ) | Fax : (819) 821-7163 J1K 2R1 | Web : http://www-gaus.gme.usherb.ca/ Andre Cote writes: Depending on the type of problems we solve ( structure only, fluid only, fluid-structure, etc.), and the type of solutions we are interested in (eigenvalue or forced problems), we use different calculations schemes. But, basically, the main bottleneck for every path comes down to two situations: 1. Solving Ax = b for A symmetric and real (double precision) file name: vibrobox.rsa How large ? 12328 DOFS What is the structure like ? symmetric with large bandwidth What structural problem ? flexible box (structure only) 2. Solving Ax = b for A symmetric and complex (double precision) file name: mplate.csa How large ? 5962 DOFS What is the structure like ? symmetric What structural problem ? multi-layer plate (that is a plate-air-poroelastic-air-plate system) Many solves are required (a few to 5000).