Can you provide a test that shows that CMS superelements, rotated with SETRAN, give a correct solution?
Yes. Here is a test that creates a 20x20 plate with a CMS superelement (free method) and three rotated duplicates using rotational SETRAN commands. The superelement matches a non-CMS modal solution. /prep7 et,1,181 mp,ex,1,1e7 mp,nuxy,1,.3 mp,dens,1,.0003 r,1,.1 rect,,10,,10 esize,2.5 amesh,1 fini /solu antyp,substr cmsop,free,12 seopt,sup1,2 nsel,s,loc,x,0 nsel,a,loc,y,0 nsel,u,loc,x,10 nsel,u,loc,y,10 m,all,all nsel,all nsel,s,loc,x,10 nsel,a,loc,y,10 d,all,all nsel,all solve fini save,substr,db /clear /filn,use /prep7 et,1,50 se,sup1 local,11,1,,,,90 csys,1 setran,sup1,11,200,sup2,sub,,,,,1 local,12,1,,,,180 csys,1 setran,sup1,12,400,sup3,sub,,,,,1 local,13,1,,,,270 csys,1 setran,sup1,13,600,sup4,sub,,,,,1 se,sup2 se,sup3 se,sup4 cpintf,all fini /solu antype,modal modop,lanb,12 mxpand,12 solve fini resu,substr,db /filn,file /solu expass,on seexp,sup1,use expsol,1,1 solve fini /clear ! compare with non-superelement solution /prep7 et,1,181 mp,ex,1,1e7 mp,nuxy,1,.3 mp,dens,1,.0003 r,1,.1 rect,,10,,10 rect,-10,0,,10 rect,-10,0,-10,0 rect,,10,-10,0 numm,kp esize,2.5 amesh,all nsel,s,ext d,all,all nsel,all fini /solu antyp,modal modop,lanb,12 solve fini /post1 set,list |
||
|