*heading
Verification of closed end loading; ELBOW31C

According to Archimedes' Principle, TOTAL of RF3 on element 1
should be -2203.04; total of RF1 should be 0.

For element 2 (curved elbow) the reaction forces should sum to 0.
Constant Pressure integrated over a closed volume = 0.

For element 3 (curved elbow) the total RF3 after step 2 should be
1420; RF1 should be zero. A curved elbow is straightened out, keeping
its length constant. This test shows that the loads are integrated 
correctly on deforming elbows.

** 
*node,nset=nall
1,0.,0.,-1.
2,.707,0.,-1.707
**
11, 1.,0.,-1.
12,1.707,0.,-1.707
**
21,0.,0.,-1.
23,.5833,0.,-1.242
*element,elset=test1,type=elbow31C
1,1,2
*element,elset=test2,type=elbow31C
2,11,12
*element,elset=test3,type=elbow31C
3,21,23
*beam section, elset=test1,material=elas,section=elbow
0.275,0.025,0.
0.,0.,-1.707
1,18,1
*beam section, elset=test2,material=elas,section=elbow
0.275,0.025,0.75
1.,0.,-1.707
1,18,1
*beam section, elset=test3,material=elas,section=elbow
0.275,0.025,0.825
0.58,0.,-1.0
1,18,1
*material,name=elas
*elastic
1.e10,0.33
*boundary
1,1,6,0
2,1,6,0
**
11,1,6,0
12,1,6,0
**
21,1,6,0
23,2,2,0
**
*step,nlgeom
Totals for Step 1  --- RF1 = 0.0 ; RF3 = -2203.04
*static
*boundary
23,1,1,0.0646
23,3,3,.242
23,5,5,-0.78539
*dload
1,hpe,1.96e4,0.,-2.,0.535
2,pi,10.,0.535
*node print,totals=yes
rf,
*elprint,f=0
*OUTPUT,history,freq=100
*NODE OUTPUT,nset=nall
rf1,rf3
*end step
**
*step,nlgeom
Totals for Step 2  --- RF1 = 0.0 ; RF3 = -3623.04
*static
*boundary,fixed
21,1,6
23,1,6
*dload
3,hpe,1.96e4,0.,-2.,0.535
*end step