Wrong external force on shell with contact

[victorkemp]

Nodes 23 and 32 in the solution should each have external force X =-2.5e7 but they're ~1e-7 instead.

This bug has been there since 2.17 but it was OK in 2.16 and earlier.

The contact doesn't have to be on the elements with the wrong external forces, just anywhere in the same model. So I've added some C3D8I to have contact.

*NODE
1,0,0,0
2,0,2,2
3,0,2,1
4,1,0,0
5,1,1,0
6,1,2,2
7,1,2,1
8,1,3,2
9,0,1,0
10,1,3,1
11,0,3,1
12,0,3,2
13,0,2,0
14,1,2,0
15,1,3,0
16,0,3,0
17,0,2,1
18,1,2,1
19,1,3,1
20,0,3,1
*ELEMENT,TYPE=C3D8I,ELSET=1
1,3,7,10,11,2,6,8,12
2,13,14,15,16,17,18,19,20
*ELEMENT,TYPE=S4,ELSET=2
3,4,5,9,1
*SURFACE,NAME=MASTER
2,S2
*SURFACE,NAME=SLAVE
1,S1
*MATERIAL,NAME=MATERIAL
*ELASTIC,TYPE=ISOTROPIC
210000000000,0
*SOLID SECTION,ELSET=1,MATERIAL=MATERIAL
*SHELL SECTION,ELSET=2,MATERIAL=MATERIAL
0.1
*BOUNDARY
1,1,6,0
9,1,6,0
*CONTACT PAIR,INTERACTION=SI_3,TYPE=SURFACE TO SURFACE
SLAVE,MASTER
*SURFACE INTERACTION,NAME=SI_3
*SURFACE BEHAVIOR,PRESSURE-OVERCLOSURE=LINEAR
2.1E+13,1
*STEP,NLGEOM=YES
*STATIC
1,1,1E-06,0
*CLOAD
4,1,50000000
5,1,50000000
*NODE FILE
U,RF
*END STEP