Tied Interfaces
Tied interfaces are composed of primary and constrained surfaces.
Internally these are broken down to nodes on the constrained side and
element (faces) on the primary side. The nodes on the constrained side
are connected to the adjacent primary face via a set of constraint
equations.
The (r,s) coordinates of the nodes relative to the primary face are
established and then the shape functions are used to construct a set of
constraint equations
us(r,s)=i∑hiup,i In the case of a quad-4 face this expands to
us(r,s)=41(1−r)(1−s)up,1+41(1+r)(1−s)up,2+41(1+r)(1+s)up,3+41(1−r)(1+s)up,4 which forms the constraint equation. This is repeated for all the
displacement directions.
The special case is the drilling degree of freedom. As the 2D elements
have either no drilling freedom or one which can work quite locally. For
this degree of freedom the rotation is linked to the translations of the
2D element. If the node is internal to the element base the rotation of
the element as a whole. If the node is on the edge use the rotation of
just that edge.
For each element node define a vector vi from the
constrained node position to the node in the plane of the element. Let
li=∣vi∣ tanαi=vi,xvi,y The displacement at the centre of the 2D element is
uc=i∑hiui Then the rotation of the node at a distance li is and angle
αi
θi=li−(ui,x−uc,x)sinαi+(ui,y−uc,y)cosαi So rotation at (r,s) is
θ(r,s)=i∑hili−(ui,x−uc,x)sinαi+(ui,y−uc,y)cosαi Or expanding
θ(r,s)θ(r,s)=i∑hili−ui,xsinαi+ui,ycosαi−i∑hili−uc,xsinαi+uc,ycosαi=i∑hili−ui,xsinαi+ui,ycosαi−i∑hili−∑j(hjuj,x)sinαi+∑j(hjuj,y)cosαi